The impact of psycholeptic reduction on
residential aged care facility residents
by
Daniel Jake Hoyle
BPharm (Hons) BCGP
School of Pharmacy and Pharmacology | University of Tasmania
Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy
University of Tasmania
November, 2020 Declaration of originality
This thesis contains no material which has been accepted for a degree or diploma by the
University or any other institution, except by way of background information and duly acknowledged in the thesis, and to the best of my knowledge and belief no material previously published or written by another person except where due acknowledgement is made in the text of the thesis, nor does the thesis contain any material that infringes copyright.
Name: Daniel Hoyle
Signed:
Dated: 17/02/2020
Authority of access
This thesis may be available for loan and limited copying and communication in accordance with the Copyright Act 1968.
Name: Daniel Hoyle
Signed:
Dated: 17/02/2020
ii | Page Daniel J. Hoyle Statement of ethical conduct
The research associated with this thesis abides by the international and Australian codes on human and animal experimentation, the guidelines by the Australian Government's Office of the Gene Technology Regulator and the rulings of the Safety, Ethics and Institutional Biosafety
Committees of the University. Ethics Approval Number H0013999.
Name: Daniel Hoyle
Signed:
Dated: 17/02/2020
Statement regarding published work contained in thesis
Chapter 4 has been modified from a published paper. The publisher of that paper holds the copyright for that content and access to the material should be sought from the respective journal. The remaining non-published content of the thesis may be available for loan and limited copying and communication in accordance with the Copyright Act 1968.
Name: Daniel Hoyle
Signed:
Dated: 17/02/2020
Statement of co-authorship
Given that this thesis contains work that has been published, a statement of co-authorship is provided for each chapter that includes published work.
The following people and institutions contributed to the publication of work undertaken as part of this thesis:
iii | Page Daniel J. Hoyle Candidate: Daniel J. Hoyle1
Author 1: Juanita L. Breen2
Author 2: Ivan K. Bindoff1
Author 3: Lisa M. Clinnick3
Author 4: Aidan D. Bindoff2
Author 5: Gregory M. Peterson1
1School of Pharmacy and Pharmacology, College of Health and Medicine, University of Tasmania,
Private Bag 26, Hobart, TAS 7001
2Wicking Dementia Research & Education Centre, College of Health and Medicine, University of
Tasmania, Private Bag 143, Hobart, TAS 7001
3Aged Operations, Ballarat Health Services
Contribution of work by co-authors for each paper:
PAPER 1: Incorporated into Chapter 4
Hoyle DJ, Bindoff IK, Clinnick LM, Peterson GM, Westbury JL. Clinical and economic outcomes of interventions to reduce antipsychotic and benzodiazepine use within nursing homes: a systematic review. Drugs & Aging. 2018;35(2):123-34.
Author contributions:
Candidate acted as the corresponding author and contributed towards conception and design of the systematic review; data extraction and synthesis; data analysis and interpretation; and writing the manuscript.
Authors 1, 2, 3 and 5 contributed towards the conception and design of the systematic review; data analysis and interpretation; and revising the manuscript critically for intellectual content.
iv | Page Daniel J. Hoyle PAPER 2: Incorporated into Chapter 6
Hoyle DJ, Peterson GM, Bindoff IK, Clinnick LM, Bindoff AD, Breen JL. Clinical impact of antipsychotic and benzodiazepine reduction: findings from a multicomponent psychotropic reduction program within long-term aged care. International Psychogeriatrics. 2020; Under review.
Candidate is acting as the corresponding author and contributed to the conception and design of the study; data collection; data analysis and interpretation; and writing the manuscript.
Authors 1, 2, 3 and 5 contributed towards the conception and design of the study; data analysis and interpretation; and revising the manuscript critically for intellectual content.
Author 4 provided statistical support and reviewed the manuscript critically for intellectual content.
We, the undersigned, endorse the above stated contribution of work undertaken for each of the published (or submitted) peer-reviewed manuscripts contributing to this thesis:
Signed:
Daniel Hoyle Dr Juanita Breen Associate Professor Candidate Primary Supervisor Glenn Jacobson School of Pharmacy and Wicking Dementia Research Head of School Pharmacology & Education Centre School of Pharmacy and University of Tasmania University of Tasmania Pharmacology University of Tasmania Date: 29/01/2020 29/01/2020 03/02/2020
v | Page Daniel J. Hoyle Abstract
Introduction
Antipsychotics and benzodiazepines, collectively known as psycholeptics, are widely used in residential aged care facilities (RACFs), leading to concern that residents are being inappropriately sedated. Whilst antipsychotics are clinically appropriate for severe psychiatric illnesses (e.g., bipolar disorder, schizophrenia), they are often used long-term for neuropsychiatric symptoms, particularly in residents with dementia. Benzodiazepines are typically used for insomnia, anxiety, and acute agitation. Considering evidence of only modest efficacy for these indications and the risk of severe adverse effects, guidelines generally recommend reserving these medications for scenarios where non-pharmacological measures are ineffective and/or where there is a risk of harm or severe distress to the resident or others.
Benzodiazepines, in particular, are not recommended for neuropsychiatric symptoms in people with dementia. If used, antipsychotics should be reviewed and reduced at a minimum of six to
12 weeks, depending on the country, and the use of benzodiazepines should be limited to a maximum of two to four weeks. Despite these recommendations, prolonged and widespread use suggests that these medications are not being reviewed and reduced regularly.
Many interventions have, to varying degrees of success, reduced the use of these medications in RACFs. However, clinical outcomes for the RACF residents and staff, and the economic impact of these interventions are rarely reported (see Chapter 4). Concern of potential deterioration in resident symptoms resulting in increased workloads is one of the reasons for a lack of willingness to reduce these medications.
Therefore, to address these gaps in the literature and minimise perceived barriers to sedative reduction, the focus of this thesis was to evaluate the impact that antipsychotic and benzodiazepine dose reduction had on resident- and RACF staff-related outcomes and health
vi | Page Daniel J. Hoyle care costs within RACFs involved in a multicomponent program to improve antipsychotic and benzodiazepine use.
Methods
The national expansion of the ‘Reducing Use of Sedatives’ (RedUSe) program involved 150
Australian RACFs. The main components of RedUSe comprised two quality improvement cycles incorporating auditing and benchmarking of prescribing, education, and multidisciplinary sedative reviews. The program was implemented between March 2014 and April 2016. A prospective, observational, longitudinal study design was utilised to investigate outcomes related to changes in antipsychotic and/or benzodiazepine doses between baseline and four months among residents of a convenience sample of 28 RACFs involved in RedUSe between
January 2015 and March 2016.
Resident-related outcomes
Permanent residents (n=206) taking antipsychotics and/or benzodiazepines on a daily basis, who did not have a severe psychiatric illness (where antipsychotic therapy should generally be maintained long-term e.g., bipolar disorder, schizophrenia) and were not considered end-stage palliative, were recruited. Neuropsychiatric symptoms (Neuropsychiatric Inventory-Nursing
Home version (NPI-NH) and Cohen-Mansfield Agitation Inventory (CMAI)), quality of life (QoL;
Assessment of Quality of Life-4D (AQoL-4D)), social withdrawal (Multidimensional Observation
Scale for Elderly Subjects (MOSES)-withdrawal subscale) and activities of daily living (Physical
Self-Maintenance Scale (PSMS)) were assessed at baseline and four months by nursing staff who completed psychometric tests as proxy raters.
Lastly, the number and severity of falls and behavioural events, number and reason for general practitioner (GP) consultations and number, length and reason for hospitalisations of
vii | Page Daniel J. Hoyle recruited residents were recorded by the nursing staff between baseline and four months.
Figure 1 illustrates all data collection associated with this project.
Staff-related outcomes
The occupational disruptiveness of the neuropsychiatric symptoms was assessed at baseline and four months during psychometric testing with the NPI-NH. Additionally, a survey of job satisfaction (adjusted-Measure of Job Satisfaction) was made available to all RACF nurses and carers for anonymous completion at baseline and four months (Figure 1).
Basic costing analysis
Health care utilisation costs were estimated using the cost of prescribed medications at baseline and four months, and the estimated cost of GP consultations and hospitalisations between baseline and four months.
Statistical methodology
Considering the longitudinal design, analyses were restricted to residents who were followed up at baseline and four months. Associations between changes in the NPI-NH (total, factor and occupational disruptiveness scores), CMAI (total and factor scores), AQoL-4D (utility and dimension scores), MOSES-withdrawal subscale and PSMS, and changes in the antipsychotic and benzodiazepine doses between baseline and four months were investigated using regression models.
The number of falls, behavioural events, GP consultations and hospitalisations were pro-rated to a monthly count. Associations between changes in the antipsychotic and/or benzodiazepine dose and these pro-rated counts were investigated using Spearman’s Rank correlations.
viii | Page Daniel J. Hoyle Differences in the adjusted-Measure of Job Satisfaction (total and factor) scores at baseline and four months were investigated using independent t-tests. Re-analysis was performed based on the occupation of the respondent.
Lastly, estimated medication and health care costs of residents who had their antipsychotic and/or benzodiazepine dose reduced were extrapolated to 12 months and were compared to (i) residents who did not have their dose reduced and (ii) the overall costs for all residents involved in the clinical outcomes study regardless of dose change.
Results
Follow-up data were available for 179 residents. Thirty of 83 residents (36%) taking an antipsychotic and 42 of 118 residents (36%) taking a benzodiazepine at baseline had reductions in their dose at four months. This included 18 out of 83 residents (22%) taking an antipsychotic and 25 out of 118 residents (21%) taking a benzodiazepine who had their medication ceased.
Resident-related outcomes
There was no evidence that resident-related outcomes were worsened by dose reductions.
Whilst there were no notable trends when measuring neuropsychiatric symptoms with the NPI-
NH, dose reduction was associated with small, albeit non-significant, improvements in behaviour, particularly less physically nonaggressive behaviour with both drug groups (-0.36 points per 10% reduction in antipsychotic dose, -0.17 per 10% reduction in benzodiazepine dose) and verbally agitated behaviour with benzodiazepine reduction (-0.16 per 10% dose reduction), as measured with the CMAI. Furthermore, antipsychotic reduction was associated with non- significant improvements in QoL and benzodiazepine dose reduction was correlated with a lower rate of falls in mobile residents.
ix | Page Daniel J. Hoyle Staff-related outcomes
Like resident-related outcomes, there was no evidence that occupational disruptiveness was worsened by dose reductions. Instead, there were trends toward improved occupational disruptiveness related to agitation with antipsychotic dose reduction and night-time behaviour with benzodiazepine dose reduction. No differences in job satisfaction were detected between baseline and four months.
Basic costing analysis
Savings between $938 and $1,486/resident/year were found among residents who had their antipsychotic and/or benzodiazepine dose reduced. Savings were driven by lower hospitalisation (savings between $897 and $1,421/resident/year) and medication costs (savings between $163 and $257/resident/year). On the other hand, residents who had their antipsychotic and/or benzodiazepine dose reduced had higher costs associated with GP consultations (costs increased by $121 to $192/resident/year).
Discussion
Overall, the research presented in this thesis found no evidence to suggest that antipsychotic and/or benzodiazepine dose reduction was associated with deteriorations in neuropsychiatric symptoms, QoL, social engagement, activities of daily living, job satisfaction or occupational disruptiveness. Moreover, trends towards improved agitation with both antipsychotic and benzodiazepine dose reduction were identified as potential benefits. Similarly, occupational disruptiveness related to agitation and night-time behaviour tended to improve with antipsychotic and benzodiazepine dose reduction, respectively. There were also savings with antipsychotic and benzodiazepine dose reduction, driven by lower costs related to hospitalisations and medications.
x | Page Daniel J. Hoyle Whilst this research suggests that antipsychotic and benzodiazepine dose reduction was safe and potentially cost saving for most residents, future larger, prospective, controlled studies are required to investigate these outcomes further.
xi | Page Daniel J. Hoyle
Figure 1: Data collection flowchart. *Resident consent sought if able to give informed consent otherwise consent sought from the ‘person responsible’. **Resident only interviewed if able to provide informed consent. PCA=Personal Care Assistant, EN=Enrolled Nurse, RN=Registered Nurse, MJS=Measure of Job Satisfaction, NPI-NH=Neuropsychiatric Inventory-Nursing Home version, CMAI=Cohen-Mansfield Agitation Inventory, MOSES=Multidimensional Observation Scale for Elderly Subjects-withdrawal subscale, PSMS=Physical Self-Maintenance Scale, AQoL- 4D=Assessment Quality of Life-4D, RACF=Residential aged care facility.
xii | Page Daniel J. Hoyle Acknowledgements
I would like to express my sincerest gratitude to the people who made this work possible.
First, to my brilliant supervisors, Associate Professor Juanita Breen, Dr Ivan Bindoff,
Associate Professor Lisa Clinnick and Distinguished Professor Gregory Peterson, your expert guidance from the design of this study to completion of the thesis is greatly appreciated. This thesis would not have been possible without your input.
To Mr Aidan Bindoff and Dr Jim Stankovich, thank you for lending your time and expertise during the development of the statistical methodology used in this thesis.
To the RedUSe project team, particularly the project pharmacists Sue Edwards and
Hilary Edwards, your assistance with recruitment and continued liaison with the residential aged care facilities helped significantly at a difficult stage of the project. I would also like to thank Donnamay Brown who provided a substantial amount of time assisting with data collection.
I would also like to widely acknowledge the support of the staff at the aged care facilities involved in this project. From resident recruitment to completion of questionnaires, your involvement made this research possible.
Research does not occur without financial support. I would like to acknowledge the following funding bodies for their support:
• The University of Tasmania for providing me with an Elite Research scholarship
and the opportunity to perform this research;
• The Australian Government for providing me with an Australian Postgraduate
Award and top-up scholarship, and funding the RedUSe project between 2013
and 2017 under the Dementia and Aged Care Service Fund.
xiii | Page Daniel J. Hoyle I must thank my friends and family who have provided me with the support needed to complete this research.
Lastly, to my biggest supporter, Dana, you have been there for me at every twist and turn. I cannot wait to come home and tell you that I have ‘PhinisheD’.
xiv | Page Daniel J. Hoyle List of publications
All publications listed resulted from the work described in this thesis.
Peer-reviewed journal publications
Hoyle DJ, Peterson GM, Bindoff IK, Clinnick LM, Bindoff AD, Breen JL. Clinical impact of antipsychotic and benzodiazepine reduction: findings from a multicomponent psychotropic reduction program within long-term aged care. International Psychogeriatrics. 2020; Under review.
Hoyle DJ, Bindoff IK, Clinnick LM, Peterson GM, Westbury JL. Clinical and economic outcomes of interventions to reduce antipsychotic and benzodiazepine use within nursing homes: a systematic review. Drugs & Aging. 2018;35(2):123-34.
Final report
Westbury J, Gee P, Ling T, Bindoff I, Brown D, Franks K, Hoyle D, Peterson G. Reducing the use of sedative medication in aged care facilities: implementation of the ‘RedUSe’ Project into everyday practice. Hobart (AU): Wicking Dementia Research and Education Centre; 2017.
Peer-reviewed full papers in conference proceedings
Hoyle D, Westbury J, Bindoff I, Clinnick L, Peterson G, 'The impact of sedative reduction on agitation and falls in aged care facilities: preliminary findings', Bringing Research to Life: 14th
National Conference of Emerging Researchers in Ageing Program and Proceedings, 7-8
December 2015, Melbourne, Australia, pp. 63-66.
Hoyle D, Westbury J, Bindoff I, Peterson G, ‘Why " RedUSe "? Rationale for studying the clinical outcomes of sedative reduction in the residential aged care setting.’ 13th National Conference of Emerging Researchers in Ageing Program and Proceedings, 24-25 November 2014, Adelaide,
Australia, pp. 84-87.
xv | Page Daniel J. Hoyle Conference abstracts (oral presentations)
Hoyle D, Bindoff I, Clinnick L, Peterson G, Westbury J, ‘Clinical impact of antipsychotic and benzodiazepine reduction: findings from a multicomponent psychotropic reduction program within long-term aged care’, International Psychogeriatric Association Congress, 1-3 September
2019, Santiago de Compostela, Spain.
Hoyle D, Westbury J, Bindoff I, Clinnick L, Peterson G, 'Clinical impact of antipsychotic and benzodiazepine reduction within residential aged care facilities: findings from a multicomponent sedative reduction program', 17th National Conference of Emerging
Researchers in Ageing, 19 - 20 October 2018, Melbourne, Australia (2018)
Hoyle D, Westbury J, Bindoff I, Clinnick L, Peterson G, 'Impact of sedative reduction, via a multifaceted intervention, on long-term care facility residents', 18th International
Psychogeriatric Association Annual Congress, 6-9 September 2016, San Francisco, USA, pp. 41-
42.
Hoyle D, Westbury J, Bindoff I, Clinnick L, Peterson G, 'The impact of sedative reduction on agitation and falls in aged care facilities: preliminary findings', Bringing Research to Life: 14th
National Conference of Emerging Researchers in Ageing Program and Proceedings, 7-8
December 2015, Melbourne, Australia, pp. 32.
Hoyle D, Westbury JL, Bindoff IK, Clinnick L, 'The impact of psychotropic reduction within aged care facilities', 48th Australian Association of Gerontology National Conference, 4-6 November
2015, Alice Springs, Australia
Hoyle D, Westbury JL, Bindoff IK, Peterson GM, 'The impact of reducing sedatives on residents: the rationale for translating sedative reduction into clinical outcomes', Australasian
Pharmaceutical Science Association Annual Conference, 5-7 December 2014, Brisbane,
Queensland, pp. 134
xvi | Page Daniel J. Hoyle Hoyle D, Westbury JL, Bindoff IK, Peterson GM, 'Why 'RedUSe'? Rationale for studying the clinical outcomes of sedative reduction in the residential aged care setting', 13th National
Conference of Emerging Researchers in Ageing, 24-25 September 2014, Adelaide, Australia, pp.
42.
Hoyle D, Westbury JL, Bindoff IK, Peterson GM, 'Clinical outcomes of reducing the use of sedatives in residents of aged care facilities', Tasmanian Health Science HDR Student
Conference, 2014, Hobart, Tasmania
Conference abstracts (posters)
Hoyle D, Westbury J, Bindoff I, Clinnick L, Peterson G, 'The impact of psychotropic reduction on nursing home residents: interim results', 17th International Psychogeriatric Association Annual
Congress, 13-16 October 2015, Berlin, Germany
Hoyle D, Westbury JL, Bindoff IK, Clinnick L, Peterson GM, 'The impact of psycholeptic reduction in residential aged care facilities: preliminary findings', 2015 Joint Australasian Pharmaceutical
Science Association-Australasian Society of Clinical and Experimental Pharmacologists and
Toxicologists (APSA-ASCEPT) Annual Conference, 29 November - 2 December 2015, Hobart,
Tasmania
List of awards
The following awards were received for work described in this thesis:
September 2019 First prize International Psychogeriatric Association Junior Research Awards in Psychogeriatrics Awarded at the 2019 International Psychogeriatric Association International Congress in Santiago de Compostela, Spain.
November 2014 Best Full Paper Awarded at the 13th National Conference of Emerging Researchers in Ageing in Adelaide, Australia
xvii | Page Daniel J. Hoyle Table of contents
Declaration of originality ...... ii
Authority of access ...... ii
Statement of ethical conduct ...... iii
Statement regarding published work contained in thesis ...... iii
Statement of co-authorship ...... iii
Abstract ...... vi
Acknowledgements ...... xiii
List of publications ...... xv
List of awards ...... xvii
Chapter 1 General introduction ...... 1
1.1 Scope of thesis...... 3
Chapter 2 Residential aged care facilities ...... 4
2.1 Ageing population ...... 4
2.2 Residential aged care ...... 5
2.3 Summary ...... 27
Chapter 3 Antipsychotic and benzodiazepine use in Australian residential aged care facilities 29
3.1 Introduction ...... 29
3.2 Antipsychotics ...... 31
3.3 Benzodiazepines ...... 50
3.4 Inappropriate use of antipsychotic and benzodiazepine in residential aged care
facilities ...... 75
xviii | Page Daniel J. Hoyle 3.5 Barriers to antipsychotic and benzodiazepine reduction ...... 80
Chapter 4 Clinical and economic outcomes of interventions to reduce antipsychotic and benzodiazepine use within residential aged care facilities ...... 84
4.1 Introduction ...... 84
4.2 Methods ...... 86
4.3 Results ...... 89
4.4 Discussion ...... 98
4.5 Conclusion ...... 100
4.6 Recent interventions ...... 100
Chapter 5 General methods ...... 106
5.1 Aim of the thesis...... 106
5.2 Introduction ...... 107
5.3 Intervention ...... 107
5.4 Study design ...... 115
5.5 Setting ...... 118
5.6 Participants ...... 119
5.7 Data collection methods ...... 123
5.8 Data validation ...... 138
5.9 Data storage ...... 139
5.10 Statistical methodology ...... 139
5.11 Cost analysis ...... 144
5.12 Feasibility of clinical evaluation ...... 144
xix | Page Daniel J. Hoyle 5.13 Ethics and study registration ...... 144
Chapter 6 Results and discussion ...... 145
6.1 Representativeness of residential aged care facilities within the clinical outcomes
study ...... 145
6.2 Identification of potentially confounding variables ...... 166
6.3 Neuropsychiatric symptoms ...... 184
6.4 Quality of life ...... 208
6.5 Social engagement ...... 224
6.6 Activities of daily living ...... 234
6.7 Risk assessment ...... 245
6.8 Staff-related outcomes...... 264
6.9 Basic costing analysis...... 297
Chapter 7 Summary and conclusion ...... 309
7.1 Future research directions ...... 311
7.2 Implications for policy and practice ...... 314
7.3 Conclusion ...... 318
References ...... 319
Appendices ...... 360
xx | Page Daniel J. Hoyle List of figures
Figure 1: Data collection flowchart...... xii
Figure 2: Proportion (%) of population aged 65 years and over and 80 years and over residing in
residential aged care facilities. Adapted from the Australian Institute of Health and Welfare,
2018.6 ...... 1
Figure 3: Predicted age composition of Australia in 2016 and 2101. Data from the Australian
Bureau of Statistics, 2013.2 ...... 4
Figure 4: People using aged care services, by care type, 30 June 2008-2018. Reproduced from
the Australian Institute of Health and Welfare, 2019.58 ...... 6
Figure 5: Proportion (%) of population aged 65 years and over and 80 years and over residing in
residential aged care facilities. Adapted from the Australian Institute of Health and Welfare,
2018.6 ...... 6
Figure 6: Number of residential aged care facilities and operational places in Australia in June
2008, 2011 and 2017. Data from several sources.62-64 ...... 7
Figure 7: Number of beds in residential aged care facilities per 1,000 people in the population
aged 65 years and over. Adapted from the Australian Institute of Health and Welfare,
2018.6 ...... 8
Figure 8: Proportion of permanent residential aged care residents, by age group, on 30 June
2000, 2007 and 2017. Data from referenced sources.59, 63 ...... 9
Figure 9: Care needs ratings of people in permanent residential care for cognition and behaviour,
30 June 2009-18. Reproduced from the Australian Institute of Health and Welfare, 2019.68
...... 10
Figure 10: Care needs ratings of people in permanent residential care for activities of daily living,
30 June 2009-18. Reproduced from the Australian Institute of Health and Welfare, 2019.68
...... 10
xxi | Page Daniel J. Hoyle Figure 11: Share of the occupations for the direct-care RACF employees (headcount and full-
time employment (FTE), %). Adapted from Mavromaras et al., 2017.65 ...... 19
Figure 12: Mechanism of benzodiazepines ...... 50
Figure 13: Prevalence of antipsychotic use in Australian residential aged care facilities by year
of publication. Adapted from Westaway et al., 2018.464 ...... 78
Figure 14: Proportion of residents prescribed antipsychotics, anxiolytics and hypnotics for an
inappropriate length of time. Data from van der Spek et al., 2016.138 ...... 79
Figure 15: Preferred reporting items for systematic reviews and meta-analyses (PRISMA)
flowchart of the study selection process...... 89
Figure 16: Implementation and clinical data collection timeline within residential aged care
facilities (RACFs). Adapted from Westbury et al., 2016.217 ...... 109
Figure 17: Data collection flowchart...... 117
Figure 18: Relationship between the tool or element used for each evaluation and the results
sub-chapter they are reported...... 118
Figure 19: Flow of the analysis for the clinical outcomes study. This sub-chapter is dedicated to
investigating the representativeness of the participants involved in the clinical outcomes
study...... 146
Figure 20: Diagram showing the flow of participants through the study...... 158
Figure 21: Beeswarm plots of the (A) chlorpromazine and (B) diazepam daily dose equivalents
(mg) at baseline and four months...... 159
Figure 22: Flow of the analysis for the clinical outcomes study. This sub-chapter is dedicated to
identifying characteristics related to changes in antipsychotic and benzodiazepine doses.
...... 167
Figure 23: Change in the chlorpromazine and diazepam daily dose equivalent between baseline
and four months...... 170
xxii | Page Daniel J. Hoyle Figure 24: Scatterplot of the chlorpromazine daily dose equivalent at baseline (green) and four
months (purple) versus the baseline Cohen-Mansfield Agitation Inventory (CMAI) factor 3
score...... 173
Figure 25: Scatterplot of the chlorpromazine daily dose equivalent at baseline (green) and four
months (purple) versus the baseline Cohen-Mansfield Agitation Inventory (CMAI) factor 2
score...... 174
Figure 26: Beeswarm plot of the diazepam daily dose (mg) at baseline and four months based
on whether the resident had a documented diagnosis of dementia (red) or not (black).
...... 176
Figure 27: Scatterplot of the diazepam daily dose equivalent at baseline (green) and four months
(purple) versus the baseline Cohen-Mansfield Agitation Inventory (CMAI) factor 2 score.
...... 177
Figure 28: Flowchart of the analysis for the clinical outcomes study. This sub-chapter is
dedicated to investigating the association between antipsychotic and benzodiazepine dose
reduction and change in neuropsychiatric symptoms...... 184
Figure 29: Beeswarm plot of the NPI-NH (A) total, (B) factor 1, (C) factor 2, (D) factor 3, (E) factor
4 and (F) factor 5 scores at baseline and four months for all residents taking an
antipsychotic at baseline...... 188
Figure 30: Scatterplots showing the relationship between change (%) in the chlorpromazine
daily dose equivalent and changes in the NPI-NH (A) total, (B) factor 1, (C) factor 2, (D)
factor 3, (E) factor 4 and (F) factor 5 scores...... 190
Figure 31: Beeswarm plot of the CMAI (A) total, (B) factor 1, (C) factor 2 and (D) factor 3 scores
at baseline and four months for all residents taking an antipsychotic at baseline ...... 192
Figure 32: Scatterplots showing the relationship between change (%) in the chlorpromazine
daily dose equivalent and changes in the CMAI (A) total, (B) factor 1, (C) factor 2 and (D)
factor 3 scores...... 194
xxiii | Page Daniel J. Hoyle Figure 33: Beeswarm plot of the NPI-NH (A) total, (B) factor 1, (C) factor 2, (D) factor 3, (E) factor
4 and (F) factor 5 scores at baseline and four months for all residents taking a
benzodiazepine at baseline...... 195
Figure 34: Scatterplots showing the relationship between change (%) in the diazepam daily dose
equivalent and changes in the NPI-NH (A) total, (B) factor 1, (C) factor 2, (D) factor 3, (E)
factor 4 and (F) factor 5 scores...... 197
Figure 35: Beeswarm plot of the CMAI (A) total, (B) factor 1, (C) factor 2 and (D) factor 3 scores
at baseline and four months for all residents taking a benzodiazepine at baseline...... 198
Figure 36: Scatterplots showing the relationship between the change (%) in the diazepam daily
dose equivalent and changes in the CMAI (A) total, (B) factor 1, (C) factor 2 and (D) factor
3 scores...... 200
Figure 37: Flow of the analysis for the clinical outcomes. This sub-chapter is dedicated to
investigating the association between antipsychotic and benzodiazepine dose reduction
and change in quality of life...... 209
Figure 38: Beeswarm plot of the AQoL-4D (A) utility and (B) independent living, (C) relationships,
(D) physical senses and (E) psychological wellbeing dimension scores at baseline and four
months for all residents taking an antipsychotic at baseline...... 213
Figure 39: Scatterplots showing the relationship between change (%) in the chlorpromazine
daily dose equivalent and changes in the AQoL-4D (A) utility and (B) independent living, (C)
relationships, (D) physical senses and (E) psychological wellbeing dimension scores. .... 215
Figure 40: Beeswarm plot of the AQoL-4D (A) utility and (B) independent living, (C) relationships,
(D) physical senses and (E) psychological wellbeing dimension scores at baseline and four
months for all residents taking a benzodiazepine at baseline...... 216
Figure 41: Scatterplots showing the relationship between change (%) in the diazepam daily dose
equivalent (DDE) and changes in the AQoL-4D (A) utility and (B) independent living, (C)
relationships, (D) physical senses and (E) psychological wellbeing dimension scores. .... 218
xxiv | Page Daniel J. Hoyle Figure 42: Flowchart of the analysis for the clinical outcomes study. This sub-chapter is
dedicated to investigating the association between antipsychotic and benzodiazepine dose
reduction and change in social engagement...... 225
Figure 43: Beeswarm plot of the MOSES-withdrawal subscale scores at baseline and four
months for all residents taking an antipsychotic at baseline...... 228
Figure 44: Scatterplot of the change in the MOSES-withdrawal subscale score versus change (%)
in the chlorpromazine daily dose equivalent...... 229
Figure 45: Beeswarm plot of the MOSES-withdrawal subscale scores at baseline and four
months for all residents taking a benzodiazepine at baseline...... 230
Figure 46: Scatterplot of the change in the MOSES-withdrawal subscale score versus change (%)
in the diazepam daily dose equivalent...... 231
Figure 47: Flowchart of the analysis for the clinical outcomes study. This sub-chapter is
dedicated to investigating the association between antipsychotic and benzodiazepine dose
reduction and change in activities of daily living...... 236
Figure 48: Beeswarm plot of the Physical Self-Maintenance Scale scores at baseline and four
months for residents taking an antipsychotic at baseline...... 238
Figure 49: Scatterplot of the change in the Physical Self-Maintenance Scale score versus change
(%) in the chlorpromazine daily dose equivalent...... 239
Figure 50: Beeswarm plot of the Physical Self-Maintenance Scale scores at baseline and four
months for residents taking a benzodiazepine at baseline...... 240
Figure 51: Scatterplots of the change in the Physical Self-Maintenance Scale score versus change
(%) in the diazepam daily dose equivalent...... 241
Figure 52: Flowchart of the analysis for the clinical outcomes study. This sub-chapter is
dedicated to investigating the association between antipsychotic and benzodiazepine dose
reduction and the rate of falls, behavioural events, General Practitioner (GP) visits and
hospitalisations...... 245
xxv | Page Daniel J. Hoyle Figure 53: Number of falls per 30 days, stratified by severity, for residents initially taking (A)
antipsychotics or (B) benzodiazepines...... 253
Figure 54: Number of behavioural events per 30 days, stratified by severity, for residents initially
taking (A) antipsychotics or (B) benzodiazepines...... 254
Figure 55: Number of general practitioner (GP) consultations per 30 days for residents initially
taking (A) antipsychotics or (B) benzodiazepines...... 255
Figure 56: Number of hospitalisations per 30 days for residents initially taking (A) antipsychotics
or (B) benzodiazepines...... 256
Figure 57: Proportion (%) of data collection period spent in hospital for residents initially taking
(A) antipsychotics or (B) benzodiazepines who had at least one hospital stay...... 256
Figure 58: Flow of the analysis for the clinical outcomes. This sub-chapter is dedicated to
investigating the association between antipsychotic and benzodiazepine dose reduction
and change in occupational disruptiveness, and differences in job satisfaction between
baseline and four months...... 266
Figure 59: Beeswarm plot of the NPI-NH total occupational disruptiveness scores at baseline
and four months for residents taking an antipsychotic at baseline...... 268
Figure 60: Scatterplot of the change in the NPI-NH total occupational disruptiveness score
versus change (%) in the chlorpromazine (CPZ) daily dose equivalent (DDE)...... 269
Figure 61: Beeswarm plot of the Neuropsychiatric Inventory-Nursing Home version domain-
specific occupational disruptiveness scores at baseline and four months for residents taking
an antipsychotic at baseline...... 271
Figure 62: Beeswarm plot of the Neuropsychiatric Inventory-Nursing Home version domain-
specific occupational disruptiveness scores at baseline and four months for residents taking
an antipsychotic at baseline...... 272
xxvi | Page Daniel J. Hoyle Figure 63: Scatterplots of the changes in the Neuropsychiatric Inventory-Nursing Home version
(NPI-NH) domain-specific occupational disruptiveness scores versus change (%) in the
chlorpromazine (CPZ) daily dose equivalent (DDE)...... 275
Figure 64: Scatterplots of the changes in the Neuropsychiatric Inventory-Nursing Home version
(NPI-NH) domain-specific occupational disruptiveness scores versus change (%) in the
chlorpromazine (CPZ) daily dose equivalent (DDE)...... 276
Figure 65: Beeswarm plot of the NPI-NH total occupational disruptiveness scores at baseline
and four months for residents taking a benzodiazepine at baseline...... 277
Figure 66: Scatterplot of the change in the NPI-NH total occupational disruptiveness score
versus change (%) in the diazepam daily dose equivalent (DDE)...... 278
Figure 67: Beeswarm plot of the Neuropsychiatric Inventory-Nursing Home version domain-
specific occupational disruptiveness scores at baseline and four months for residents taking
a benzodiazepine at baseline...... 280
Figure 68: Beeswarm plot of the Neuropsychiatric Inventory-Nursing Home version domain-
specific occupational disruptiveness scores at baseline and four months for residents taking
a benzodiazepine at baseline...... 281
Figure 69: Scatterplots of the changes in the Neuropsychiatric Inventory-Nursing Home version
(NPI-NH) domain-specific occupational disruptiveness scores versus change (%) in the
diazepam daily dose equivalent (DDE)...... 285
Figure 70: Scatterplots of the changes in the Neuropsychiatric Inventory-Nursing Home version
(NPI-NH) domain-specific occupational disruptiveness scores versus change (%) in the
diazepam daily dose equivalent (DDE)...... 286
Figure 71: Beeswarm plots of the adjusted-Measure of Job Satisfaction factor scores at baseline
and four months...... 290
Figure 72: Beeswarm plot of the adjusted-Measure of Job Satisfaction (MJS) total scores at
baseline and four months...... 291
xxvii | Page Daniel J. Hoyle Figure 73: Flowchart of the analysis for the clinical outcomes study. This sub-chapter is
dedicated to a basic costing analysis of antipsychotic and benzodiazepine dose reduction.
...... 298
xxviii | Page Daniel J. Hoyle List of tables
Table 1: The prevalence and pathophysiological features of common dementia subtypes...... 12
Table 2: Neuropsychiatric symptoms associated with dementia. Information from several
sources.31, 75, 106 ...... 14
Table 3: Duration of frequently observed neuropsychiatric symptoms. Data from Voyer et al.,
2014.128 ...... 17
Table 4: Psycholeptic drug groups. Data from the World Health Organisation Collaborating
Centre for Drug Statistics Methodology, 2016.190 ...... 30
Table 5: Antipsychotic medications used in Australia. Information from the Australian Medicines
Handbook, 2018.205 ...... 32
Table 6: Example of differences in relative receptor binding affinities for antipsychotics. Adapted
from Neil et al., 2003.207 ...... 33
Table 7: Comparison of antipsychotic adverse effects. Information from referenced sources.205,
206 ...... 34
Table 8: Benzodiazepines available in Australia according to the length of action. Adapted from
the Australian Medicines Handbook, 2018.296 ...... 51
Table 9: Symptoms of acute and chronic benzodiazepine withdrawal syndrome. Information
from several sources.196, 443, 444 ...... 74
Table 10: Use of antipsychotics and benzodiazepines by residents of residential aged care
facilities according to country. Adapted from Hasan et al., 2019.471...... 77
Table 11: Use of psychotropic medication in Sydney residential aged care facilities in 1993, 1998,
2003 and 2009. Adapted from Snowdon et al., 2011.13 ...... 78
Table 12: Summary of clinical outcomes reported in included publications...... 94
Table 13: Intervention descriptions (summary using Cochrane EPOC Data Collection
Checklist).510 ...... 96
Table 14: Indicators of study quality...... 97
xxix | Page Daniel J. Hoyle Table 15: Estimated effect of antipsychotic deprescription on clinical outcomes within the
Halting Antipsychotic use in Long Term care (HALT) study. Data from Brodaty et al., 2018.530
...... 102
Table 16: Percentage reduction in antipsychotic and anxiolytic doses based on the intervention
used in the study by Weeks et al., 2019.534 ...... 104
Table 17: Location and number of residential aged care facilities consenting to participation.
...... 119
Table 18: Factors of the Neuropsychiatric Inventory-Nursing Home version.558 ...... 127
Table 19: Factor structure of the Cohen-Mansfield Agitation Inventory.561 ...... 129
Table 20: Fall severity assessment...... 134
Table 21: Behavioural episode severity assessment...... 135
Table 22: Factor structure of the adjusted-Measure of Job Satisfaction.592 ...... 137
Table 23: Meaningful minimum change in factor and total scores for the adjusted-Measure of
Job Satisfaction, utilising the 0.5 of a standard deviation rule...... 138
Table 24: Comparison between the residential aged care facilities (RACFs) involved in the clinical
outcomes study and RACFs involved in the Reducing Use of Sedatives (RedUSe) project.
...... 151
Table 25: Comparison between the residential aged care facilities (RACFs) involved in the clinical
outcomes study and RACFs data from the Australian Institute of Health and Welfare.595
...... 154
Table 26: Demographic details for residents at baseline of the clinical outcomes study and
permanent resident information from the Australian Institute of Health and Welfare,
where available.595 ...... 156
Table 27: Variation in the use of antipsychotics and benzodiazepines within a sample of
residents involved in the clinical outcomes research between baseline and four months.
...... 159
xxx | Page Daniel J. Hoyle Table 28: Baseline resident characteristics that may impact antipsychotic and/or
benzodiazepine use.299, 492, 524, 607, 608 ...... 169
Table 29: Estimated standardised changes in prescribed antipsychotic dosages (in standard
deviation units) and 95% highest density intervals (HDI)...... 173
Table 30: Estimated standardised changes in prescribed benzodiazepine dosages (in standard
deviation units) and 95% highest density intervals (HDI)...... 176
Table 31: Mean Neuropsychiatric Inventory-Nursing Home version total and factor scores at
baseline and four months...... 187
Table 32: Estimated changes in the NPI-NH total and factor scores associated with changes (%)
in the chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and
four months...... 189
Table 33: Mean Cohen-Mansfield Agitation Inventory total and factor scores at baseline and
four months for residents using an antipsychotic at baseline...... 191
Table 34: Estimated changes in the CMAI total and factor scores associated with changes (%) in
the chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and four
months...... 193
Table 35: Mean Neuropsychiatric Inventory-Nursing Home version total and factor scores at
baseline and four months...... 195
Table 36: Estimated changes in the NPI-NH total and factor scores associated with changes (%)
in the diazepam daily dose equivalent (benzodiazepine dose) between baseline and four
months while controlling for a documented diagnosis of dementia...... 196
Table 37: Mean Cohen-Mansfield Agitation Inventory total and factor scores at baseline and
four months for residents using a benzodiazepine at baseline...... 198
Table 38: Estimated changes in the CMAI total and factor scores associated with changes (%) in
the diazepam daily dose equivalent (benzodiazepine dose) between baseline and four
months...... 199
xxxi | Page Daniel J. Hoyle Table 39: Utilisation of other medication classes that may have beneficial or negative effects on
neuropsychiatric symptoms, at baseline and four months...... 201
Table 40: Proportion of residents using other medication classes that may have beneficial or
negative effects on neuropsychiatric symptoms, at baseline and four months...... 201
Table 41: Count of residents with an Assessment of Quality of Life-4D score at baseline and four
months, stratified by the rater...... 210
Table 42: Mean Assessment of Quality of Life-4D utility and dimension scores at baseline and
four months...... 213
Table 43: Estimated changes in the AQoL-4D utility and dimension scores associated with
changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between
baseline and four months...... 214
Table 44: Mean Assessment of Quality of Life-4D utility and dimension scores at baseline and
four months...... 216
Table 45: Estimated changes in the AQoL-4D utility and dimension scores associated with
changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between
baseline and four months...... 217
Table 46: Estimated changes in the MOSES-withdrawal subscale score associated with changes
(%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and
four months...... 229
Table 47: Estimated changes in the MOSES-withdrawal subscale score associated with changes
(%) in the diazepam daily dose equivalent (benzodiazepine dose) between baseline and
four months...... 231
Table 48: Estimated changes in the Physical Self-Maintenance Scale score associated with
changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between
baseline and four months...... 239
xxxii | Page Daniel J. Hoyle Table 49: Estimated changes in the Physical Self-Maintenance Scale score associated with
changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between
baseline and four months...... 241
Table 50: Information collected on falls, behavioural events, general practitioner (GP)
consultations and hospitalisations...... 250
Table 51: Falls and behavioural events severity assessment...... 251
Table 52: Correlation (rs) between change (%) in antipsychotic and benzodiazepine dose and the
rate of falls, behavioural events, general practitioner (GP) consultations and
hospitalisations per 30 days...... 257
Table 53: Correlation (rs) between change (%) in antipsychotic and benzodiazepine dose and
proportion of time spent in hospital for residents who had at least one hospital admission.
...... 258
Table 54: Estimated changes in the NPI-NH total occupational disruptiveness score associated
with changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between
baseline and four months...... 269
Table 55: Mean Neuropsychiatric Inventory-Nursing Home version domain-specific
occupational disruptiveness scores at baseline and four months...... 270
Table 56: Estimated changes in the Neuropsychiatric Inventory-Nursing Home version domain-
specific occupational disruptiveness scores associated with changes (%) in the
chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and four
months...... 273
Table 57: Estimated changes in the NPI-NH total occupational disruptiveness score associated
with changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between
baseline and four months...... 278
Table 58: Mean Neuropsychiatric Inventory-Nursing Home version domain-specific
occupational disruptiveness scores at baseline and four months...... 279
xxxiii | Page Daniel J. Hoyle Table 59: Estimated changes in the Neuropsychiatric Inventory-Nursing Home version domain-
specific occupational disruptiveness scores associated with changes (%) in the diazepam
daily dose equivalent (benzodiazepine dose) between baseline and four months...... 282
Table 60: Demographic data for nursing and care staff participating in the job satisfaction survey
at baseline and four months...... 289
Table 61: Mean adjusted-Measure of Job Satisfaction total and factor scores at baseline and
four months...... 290
Table 62: Descriptions of Medicare Benefits Scheme item numbers. Information from the
Department of Health, 2015.699 ...... 301
Table 63: Average potential cost savings on a per resident basis...... 304
xxxiv | Page Daniel J. Hoyle List of abbreviations
ACFI Aged Care Funding Instrument
ADL Activities of Daily Living
AIHW Australian Institute of Health and Welfare
AMA Australian Medical Association
AQoL-4D Assessment of Quality of Life-4D
ATC Anatomical Therapeutic Classification
BPRS Brief Psychiatric Rating Scale
CMAI Cohen-Mansfield Agitation Inventory
CNS Central nervous system
COTA Council on the Ageing
DATIS Drug And Therapeutics Information Service
DEMQOL Dementia Quality of Life
DSM Diagnostic Statistical Manual of Mental Disorders
EN Enrolled Nurse
EPOC Effective Practice and Organisation of Care
EQ-5D EuroQol-5D
FDA Food and Drug Administration
FTLD Frontotemporal Lobar Degeneration
GABA Gamma-aminobutyric acid
GP General Practitioner
GRIP Grip on Challenging Behaviour care programme
HDI High Density Intervals
ICC Intra-class correlation
ICD International Classification of Diseases
KPI Key Performance Indicators
xxxv | Page Daniel J. Hoyle LSM least-squares mean
MCMC Markov Chain Monte Carlo
MOSES Multidimensional Observation Scale for Elderly Subjects
NEAF National Ethics Application Form
NNH Number needed to harm
NPI-NH Neuropsychiatric Inventory-Nursing Home version
NPS National Prescribing Service
OECD Organisation for Economic Co-operation and Development
OR Odds Ratio
PBS Pharmaceutical Benefits Scheme
PCA Person Care Assistant
PSA Pharmaceutical Society of Australia
PSMS Physical Self-Maintenance Scale
QALY Quality-adjusted life years
QoL Quality of life
QUM Quality Use of Medicines
RACF Residential aged care facility
RANZCP Royal Australian and New Zealand College of Psychiatrists
RCT Randomised controlled trial
RedUSe Reducing Use of Sedatives project
RMMR Residential Medication Management Review
RN Registered Nurse
RR Rate Ratio
SF-36 Short-Form 36
SPSS Statistical Package for Social Sciences
VALMER Value of Medication Reviews
xxxvi | Page Daniel J. Hoyle WHELD Well-Being and Health for People with Dementia
xxxvii | Page Daniel J. Hoyle Chapter 1 General introduction
Internationally, populations are ageing rapidly.1-3 This has resulted in increased strain on aged care services. As of 30th of June 2017, one in five Australians aged 85 years and over were permanent residents of residential aged care facilities (RACFs).4, 5 Compared to member countries from the Organisation for Economic Co-Operation and Development (OECD), Australia had the second-highest proportion of the population, aged ≥65 years (6.3%) and ≥80 years
(19.7%), in RACFs (Figure 2).6 Those entering RACFs are older and have more complex health care needs.7, 8 These health care needs are often associated with the management of neuropsychiatric symptoms, such as agitation, depression and sleep disturbances, which may or may not be associated with dementia.7, 8
30%
25%
20%
15%
10%
5%
Proportion (%) of population of (%) Proportion 0% Israel Spain Japan Korea Latvia Poland Ireland Iceland Finland Estonia Canada Norway Sweden France* Belgium Slovenia Slovenia Hungary Portgual Australia Denmark Germany Switzerland Netherlands Luxembourg New Zealand United States OECD average Czech Republic United Kingdom* United
Proportion of population aged 65 years and over Proportion of population aged 80 years and over
Figure 2: Proportion (%) of population aged 65 years and over and 80 years and over residing in residential aged care facilities. Adapted from the Australian Institute of Health and Welfare, 2018.6 *Data for proportion of population aged 80 years and over not recorded.
There is concern that RACF residents are inappropriately sedated with antipsychotic and benzodiazepine medication.9-16 These medications are collectively known as sedatives or psycholeptic medication.17 Psycholeptic medications are the cornerstone to the management of severe psychiatric illnesses (e.g., schizophrenia, bipolar disorder)18-21 and antipsychotics are
1 | Page Daniel J. Hoyle indicated for the short-term management of psychotic or aggressive symptoms in people with dementia.22-24 However, within the RACF setting, antipsychotics are regularly used long-term to treat general neuropsychiatric symptoms in people with dementia, and benzodiazepines are typically used to mitigate agitation, anxiety and sleep disturbances.25-27 Generally, when used for these indications, guidelines recommend reserving these treatments as second-line options should non-pharmacological strategies alone be ineffective or where there is a risk of harm due to severe distress, agitation/aggression or psychosis.28-33 If prescribed, antipsychotics and benzodiazepines should be regularly reviewed with the aim to use the lowest effective dose possible for the shortest period of time.13-17, 19
These guidelines reflect the modest efficacy that antipsychotics and benzodiazepines have for these indications and the risk of severe adverse effects.22, 34-37 However, despite the risks of antipsychotic and benzodiazepine use, and the potential benefits derived from the reduction of these medications, studies have continued to report high rates of antipsychotic and benzodiazepine usage in Australian RACFs.10, 11, 13, 15, 16, 38, 39 The high rates of antipsychotic and benzodiazepine prescribing have been attributed, in part, to a lack of regular review or dose reduction attempts.12, 25, 40-42 Perceived risk of worsened symptoms or quality of life (QoL) are also barriers to dose reduction among nursing staff and physicians.43-45
Interventions to minimise the prescribing of antipsychotics and/or benzodiazepines have been trialled and implemented.46-48 However, there has been relatively little investigation into the clinical outcomes of antipsychotic and/or benzodiazepine dose reduction for the RACF residents or nursing staff.47 Following a successful pilot study in Tasmania,49 an intervention, called the ‘Reducing Use of Sedatives’ (RedUSe) project, received funding for national expansion across Australia.50 This project provided the ideal opportunity to assess the clinical outcomes associated with antipsychotic and benzodiazepine dose changes.
2 | Page Daniel J. Hoyle 1.1 Scope of thesis
This thesis investigated associations between changes in antipsychotic and benzodiazepine use and clinical outcomes for RACF residents involved in the national expansion of RedUSe. In this context, clinical outcomes are defined as indicators of health status for the residents.
Specifically, changes in neuropsychiatric symptoms, QoL, social engagement, activities of daily living (ADLs), falls, behavioural events, general practitioner (GP) consultations and hospitalisations were investigated over four months.
This thesis also investigated changes in the disruptiveness of neuropsychiatric symptoms on RACF nursing and care staff associated with antipsychotic and/or benzodiazepine dose reduction and differences in job satisfaction at baseline and four months.
Lastly, a basic costing analysis was performed to estimate changes in health care- related costs associated with antipsychotic and/or benzodiazepine dose reduction.
3 | Page Daniel J. Hoyle Chapter 2 Residential aged care facilities
2.1 Ageing population
Increasing life expectancy has led to an ageing population worldwide.51 Internationally, the proportion of people aged 65 years and over is projected to increase from 6% in 1990 to 16% in
2050.51 In Australia, the number of people aged 65 years and over has grown from 319,000 (5% of the population) in 1926 to 3.7 million (15% of the population) in 2016.1 By 2101, a projected
13.4 million Australians (25% of the projected population) will be aged 65 years or greater.2
Figure 3 illustrates the projected change in the age composition for the Australian population between 2016 and 2101 assuming that fertility, life expectancy and overseas migration remains constant.2
16% 14% 12% 10% 8% 6% 4% 2% % of the total population total of the % 0% 0-9 10-19 20-29 30-39 40-49 50-59 60-69 70-79 80 and older Age groups (years)
2016 2101
Figure 3: Predicted age composition of Australia in 2016 and 2101. Data from the Australian Bureau of Statistics, 2013.2
There is concern that the ageing population will increase health care demands, require a larger and better-trained health care workforce, and increase the need for age-friendly environments.52 These factors also detract from economic growth. The increase in the proportion of Australians at retirement age (65-years-old) relative to the proportion of people participating in the labour force results in increased expenditure on retirement pensions and reduced tax revenue.53 Additionally, there is a demand for increased government spending on
4 | Page Daniel J. Hoyle programs to support older Australians.53 Over the next decade, annual growth in revenue is projected to reduce by 0.4 points and annual growth in spending is projected to increase by 0.3 points. This equates to an annual cost to the Australian Government budget of approximately
$36 billion by 2028-29.53
2.2 Residential aged care
2.2.1 Definition
RACFs, also known as “nursing homes”, “long-term care facilities”, “aged care homes” and “care homes” internationally,54, 55 provide care for people who can no longer live at home and need ongoing help with everyday tasks or health care (e.g., nursing care, continence aids, basic medical, pharmaceutical supplies).56, 57 For the purposes of this thesis, we defined a RACF as any residential institution in which permanent residents (residents who receive long-term care in a
RACF) receive around-the-clock care from staff at the facility.
2.2.2 Use of residential aged care
The number of people requiring aged care, particularly home care services, has grown with
Australia’s ageing population (Figure 4). Although growth has not been as substantial, there is still a strong demand for residential care. On the 30th of June 2018, there were 180,923 permanent residents of RACFs.58 Most residents (n=174,875, 97%) were aged 65 years and over and 59% (106,445) were aged 85 years and over.59 This accounted for 21% of Australians aged
85 years and over.60
5 | Page Daniel J. Hoyle
Figure 4: People using aged care services, by care type, 30 June 2008-2018. Reproduced from the Australian Institute of Health and Welfare, 2019.58
Internationally, between 0.4% (Latvia) and 8.8% (Belgium) of populations aged 65 years and over are residents of a RACF (Figure 5).6 On the other hand, international data indicate that between 0.3% (Latvia) and 24% (Belgium) of people aged 80 years and over reside in a RACF
(Figure 5).6
30%
25%
20%
15%
10%
5%
Proportion (%) of population of (%) Proportion 0% Israel Spain Japan Korea Latvia Poland Ireland Iceland Finland Estonia Canada Norway Sweden France* Belgium Slovenia Slovenia Hungary Portgual Australia Denmark Germany Switzerland Netherlands Luxembourg New Zealand United States OECD average Czech Republic United Kingdom* United
Proportion of population aged 65 years and over Proportion of population aged 80 years and over
Figure 5: Proportion (%) of population aged 65 years and over and 80 years and over residing in residential aged care facilities. Adapted from the Australian Institute of Health and Welfare, 2018.6 *Data for proportion of population aged 80 years and over not recorded.
6 | Page Daniel J. Hoyle 2.2.3 Number of residential aged care facilities
Despite a decrease in the number of Australian RACFs over the last decade, there is a trend of existing facilities increasing bed numbers (Figure 6).61-64 Between 2007 and 2016, the proportion of Australian RACFs providing ≥61 operational places increased from 35 to 52%.65
3000 250000 Number of operational places 2500 200000
2000 150000 1500 100000 1000
50000 500
0 0 Number of residential aged care facilities 2008 2011 2018
Number of residential aged care facilities Number of operational places
Figure 6: Number of residential aged care facilities and operational places in Australia in June 2008, 2011 and 2017. Data from several sources.62-64
As of 30 June 2018, there were 2,695 RACFs in Australia run by 886 organisations.64
Most (62%) RACFs were based in major cities.64 These RACFs provided 207,142 places with an average occupancy rate of 90% across 2018.64 In 2017-18, residential aged care services cost
$12.4 billion accounting for 67.3% of the Australian aged care service funding.66
7 | Page Daniel J. Hoyle In 2017, there were approximately 51.1 beds per 1,000 Australians aged 65 years and over.6 Internationally, the number of beds per 1,000 of the population aged 65 years ranges from 1.3 (Greece) to 82.8 (Luxembourg)(Figure 7).6
90
80
70
60
50
40
30
20
10
0 Italy Israel Spain Beds Beds per 1,000 people aged 65 yeras and over Japan Korea Latvia France Turkey Poland Ireland Austria Greece Iceland Finland Estonia Norway Sweden Belgium Slovenia Hungary Canada* Denmark Germany Lithuania Australia* Switzerland Netherlands Luxembourg New Zealand United States Czech Repulic Slovak Republic United Kingdom United
Figure 7: Number of beds in residential aged care facilities per 1,000 people in the population aged 65 years and over. Adapted from the Australian Institute of Health and Welfare, 2018.6 *Estimated. 2.2.4 Resident characteristics
2.2.4.1 Age
Figure 8 illustrates the change in the proportion of permanent residents of RACFs in Australia.59,
63 Between 2000 and 2017 there was a 16% increase in the proportion of residents aged 85 years and over.59, 63 This is related to an increase in the admission of older residents into RACFs.
In 2008-09, 49% of admissions into permanent residential care were people aged 85 years and over.67 In 2017-18, 53% of people entering residential aged care were people aged 85 years and over.67
8 | Page Daniel J. Hoyle 70 60 50 40 30 20 10
Proportion(%) of residents 0 <65 65-69 70-74 75-79 80-84 85+ Age group (years)
2000 2007 2017
Figure 8: Proportion of permanent residential aged care residents, by age group, on 30 June 2000, 2007 and 2017. Data from referenced sources.59, 63
2.2.4.2 Care level
The needs of RACF residents are assessed through the Aged Care Funding Instrument (ACFI), which reviews three domains of care needs: ADLs, cognition and behaviour, and complex health care.68 Since 2009, care need ratings for cognition and behaviour (Figure 9), complex health care and ADLs (Figure 10) have increased. The cognition and behavioural domain has consistently had the largest proportion of people with needs rated as high.7, 68 Between 2009 and 2018, high care need ratings in this domain increased from 36.9% to 64.1%.68 This coincides with an increase in the proportion of residents with at least one mental health or behavioural condition, which has increased from 68% of residents in 2009 to 86% of residents in 2018.8
The large proportion of residents with high needs related to the cognition and behavioural domain is not surprising considering that the level of funding is based on the resident’s ACFI score (greater care need equals more funding).69-71 Whilst ACFI data provides an overall snapshot of the level of care required across different domains of care, it is likely that the level of care required has been over-estimated to attract more funding. Additionally, it should be noted that there has been limited research on the validity of the ACFI tool.
9 | Page Daniel J. Hoyle
Figure 9: Care needs ratings of people in permanent residential care for cognition and behaviour, 30 June 2009-18. Reproduced from the Australian Institute of Health and Welfare, 2019.68
Figure 10: Care needs ratings of people in permanent residential care for activities of daily living, 30 June 2009-18. Reproduced from the Australian Institute of Health and Welfare, 2019.68
2.2.4.3 Mental health among residential aged care residents
High rates of mental health and behavioural conditions in residential aged care can be explained by several factors. First, many residents have conditions, such as dementia, with associated neuropsychiatric symptoms.7, 68 Second, the closure of long-stay psychiatric hospital beds as a result of deinstitutionalisation of mental health has increased the need to manage psychological issues within the RACF setting.72 Third, reduced mortality in younger mentally ill people has led to an increasing number of RACF residents with mental health issues.73 Fourth, the number of
10 | Page Daniel J. Hoyle age-related physical conditions and disabilities, which predispose older people to psychological issues, are increasing with the ageing population.74
2.2.4.3.1 Dementia
Dementia, subsumed under the entity “major neurocognitive disorder” in the Diagnostic
Statistical Manual of Mental Disorders (DSM-5),75 is defined as “a syndrome due to disease of the brain, usually of chronic or progressive nature, in which there is disturbance of multiple higher cortical functions, including memory, thinking, orientation, comprehension, calculation, learning capacity, language, and judgement. Consciousness is not clouded. Impairments of cognitive function are commonly accompanied, and occasionally preceded, by deterioration in emotional control, social behaviour, or motivation”.76 This definition highlights three distinct presentations of dementia, including cognitive and behavioural components, and difficulties performing ADLs.
The risk of dementia in people aged 70 to 74 years is 3.4%. This risk is increased to 20% in 85 to 89-year-olds and 40% in people aged over 95 years.77 In 2018, an estimated 376,000 people in Australia had dementia.78 By 2030, the number of people with dementia is projected to increase to approximately 550,000 people.78 This will likely have flow-on effects for residential aged care. Currently, 52% of RACF residents have a documented diagnosis of dementia.7 This proportion, however, may be an under-estimation as the dementia is often under-reported due to poor documentation or diagnosis.79 Illustrating this, a study used the
Neuropsychiatric Inventory (NPI) to assess the incidence of dementia in 1,163 Norwegian RACF residents.80 Whilst only 55% of residents had a diagnosis of dementia recorded in their medical records, 80.5% of residents were found to have dementia using the NPI as a diagnostic tool.80
This is consistent with prevalence rates of dementia reported in Australian (78.1%)81 and Finnish
(81%) studies.82
11 | Page Daniel J. Hoyle 2.2.4.3.1.1 Subtypes of dementia
Dementia subtypes are classified according to the known or presumed aetiology.75 More than
100 diseases can cause dementia.83 The four most common subtypes of dementia include
Alzheimer’s disease, vascular dementia, dementia with Lewy bodies, and frontotemporal lobar degeneration (FTLD).84 The prevalence and pathophysiological features of the individual dementia subtypes are briefly described in Table 1. It needs to be acknowledged though, dementia due to mixed aetiologies is the most common form of dementia.85 A study of 179 people with probable Alzheimer’s disease found evidence of mixed pathologies in 46% of patients; the most common pathology was Alzheimer’s disease and macroscopic infarcts (i.e., vascular dementia).86 Considering the high prevalence of mixed dementias, rigid separation of dementia subtypes is not always helpful.
Table 1: The prevalence and pathophysiological features of common dementia subtypes.
Dementia subtype Prevalence Changes to the central nervous system Alzheimer’s disease 50-80%87-89 Hallmark feature is the formation of amyloid protein plaques and tau-protein neurofibrillary tangles.90 This leads to damage to neurons, particularly cholinergic neurons.91 Vascular dementia 10-20%79, 84 Macrovascular events (e.g., stroke) or microvascular disease leads to central nervous system damage.75 Dementia with Lewy 15-25%92 Abnormal aggregations of protein alpha-synuclein, bodies called Lewy bodies, accumulate in neurons throughout the cortex and cause cognitive impairment.89 Frontotemporal lobar 10%28, 89 There is atrophy of the frontal and temporal lobes. degeneration Additionally, the upper layers of the cortex become soft and spongy and have protein inclusions (usually tau proteins).89
Pathophysiological differences among the dementia subtypes provide explanations for differences in the cognitive symptoms. These differences are mainly seen between Alzheimer’s disease/vascular dementia and dementia with Lewy bodies/Parkinson’s disease dementia/FTLD.
In Alzheimer’s disease, initial damage is usually located in the hippocampus resulting in memory and learning impairments.90 Cognitive function then progressively worsens over months to years as other areas of the brain are damaged.85, 88, 93, 94 Whilst some patients with
12 | Page Daniel J. Hoyle vascular dementia, due to multiple infarctions, present with an acute stepwise or fluctuating decline in cognition alongside periods of stability,75 many cases are due to microvascular disease leading to lesions in the white matter, basal ganglia and/or thalamus, resulting in symptoms that present and progress similarly to Alzheimer’s disease.75 Additionally, differential diagnosis is further complicated by the presence of mixed Alzheimer’s disease and vascular dementia.
Vascular disease is the sole cause of dementia in only five per cent of cases.85
In dementia with Lewy bodies, the progression of dementia is more rapid than
Alzheimer’s disease.95 Furthermore, initial symptoms, including sleep disturbances, visual hallucinations, visuospatial impairment, gait imbalance and parkinsonian movements,89 often occur in the absence of significant memory impairment and fluctuate regularly.85, 89, 95 Although similar to dementia with Lewy bodies, Parkinson’s disease dementia differs in that cognitive decline occurs at least 12 months after the diagnosis of Parkinson’s disease is established.75
Compared to Alzheimer’s disease, Parkinson’s disease dementia can be distinguished by the motor features associated with Parkinson’s disease.75 In FTLD, memory is generally spared in the early stages. Instead, initial symptoms usually include personality and behavioural changes, and difficulty producing and comprehending language.89 The pathophysiological differences between dementia subtypes also explain, in part, variability in the presentation of neuropsychiatric symptoms associated with the different dementia subtypes. Generally, differences are most distinct between Alzheimer’s disease/vascular dementia and dementia with Lewy bodies/FTLD.96
2.2.4.3.1.2 Neuropsychiatric symptoms in people dementia
Neuropsychiatric symptoms in people with dementia, also known as behavioural and psychological symptoms of dementia (BPSD), changed behaviours, challenging behaviours, behaviours that challenge, and responsive behaviours, is a term that describes symptoms of
13 | Page Daniel J. Hoyle disturbed perception, thought content, mood and behaviour that frequently occur in patients with dementia.97-99
Neuropsychiatric symptoms, described in Table 2, have been observed in 60 to 98% of people with dementia.80, 96, 98, 100-104 Generally, neuropsychiatric symptoms are more frequent in people in hospitals and RACFs compared to those dwelling in the community.96 Brodaty et al. found that 92% of residents from an Australian RACF, with a documented diagnosis of dementia or significant cognitive impairment, had at least one neuropsychiatric symptom.81 Zuidema et al. reported clinically significant symptoms among 81% of Dutch residents with dementia.105
Behavioural symptoms are usually identified based on observation of the patient and psychological symptoms are mainly assessed through an interview with patient and relatives/carers.31
Table 2: Neuropsychiatric symptoms associated with dementia. Information from several sources.31, 75, 106
Behavioural symptoms Psychological symptoms Physical aggression Anxiety Screaming Depressive mood Restlessness Hallucinations Agitation Delusions Wandering Psychosis Culturally inappropriate behaviour Misidentifications Sexual disinhibition Sleeplessness Hoarding Cursing Shadowing
The manifestation of neuropsychiatric symptoms is influenced by a complex interplay of different factors, such as unmet needs, environmental factors and pathophysiological changes to the brain.96 As such, there is substantial variability in the incidence and severity of neuropsychiatric symptoms within and between each dementia subtype.96
The development of neuropsychiatric symptoms in dementia is primarily the result of pathophysiological changes to the brain.31, 96, 107-110 Consequently, it could be assumed that particular neuropsychiatric symptoms may be more common in certain dementia subtypes.
14 | Page Daniel J. Hoyle However, studies comparing the profiles of neuropsychiatric symptoms across dementia subtypes have not reported consistent results.96 For example, Ballard et al. reported that depression and anxiety are more common among people with vascular dementia compared to
Alzheimer’s disease.111 Conversely, Srikanth et al. found similar symptom profiles in people with vascular dementia and Alzheimer’s disease.112 The difficulty differentiating the neuropsychiatric symptoms in Alzheimer’s disease and vascular dementia may, in part, be attributed to the overlap of Alzheimer’s disease and cerebrovascular pathology.86 On the other hand, dementia with Lewy bodies can be differentiated from Alzheimer’s disease by a consistently higher prevalence of delusions and hallucinations.113 Lastly, FTLD is distinguished by a “loss of basic emotions, food cramming, pacing a fixed route, preserved capacity of locating objects, and impairment insight”.96
People with dementia have heightened sensitivity to environmental stressors and cues.114 As such, the development of neuropsychiatric symptoms in dementia is strongly linked to the psychosocial and physical environment.114-116 In 2018, a review of 103 empirical studies
(n=94) and reviews (n=9) reported substantial evidence that the unit size, spatial layout, homelike character, sensory stimulation, and characteristics of social spaces were related to the behaviour and wellbeing of RACF residents.114 For example, larger units117 and RACFs81 have been linked to higher levels of neuropsychiatric symptoms. In 2001, Brodaty et al. showed that the rate and severity of neuropsychiatric symptoms, such as delusions, hallucinations and psychosis, tend to be greater in RACFs with larger populations.81 Additionally, Zuidema et al. have found lower levels of apathy among Dutch Dementia Special Care Units (n=56 units) with greater staff-to-resident ratios and in units where nurses spent more time on resident care.115
More recently, Jao et al. found that environments that are tailored to stimulate individual residents have been linked to lower levels of apathy.118 Many environmental factors, such as noise level, bright room light, temperature, unit size, and homelike décor are relatively easy to adjust.118 Whilst there is evidence to support the influence of psychosocial and physical
15 | Page Daniel J. Hoyle environments on neuropsychiatric symptoms of dementia, research in this area is generally limited to cross-sectional, small-scale studies in homogenous samples and further investigation is needed.114
Neuropsychiatric symptoms can also manifest as an expression of unmet needs, where individuals are unable to meet or communicate their needs (e.g., discomfort, boredom, sensory deprivation, under-treated pain).119, 120 A recent study of 140 residents, with and without dementia, from three Portuguese RACFs, found that the presence of neuropsychiatric symptoms, particularly psychotic symptoms, are related to the number of unmet needs the resident has.120 These results align with an earlier study of 238 RACF residents, which reported increased incidence and frequency of neuropsychiatric symptoms associated with greater unmet needs.121 In the RACF setting, anxiety and depression have been associated with loneliness, pain, visual impairment and hearing loss.121, 122 Furthermore, agitation has been linked with discomfort, pain and boredom.123 Consequently, identifying and addressing these unmet needs with a patient-centred care approach may improve some neuropsychiatric symptoms.120 In particular, improvement in agitation has been demonstrated with improved communication, sensory and activities-based interventions, and treatment of pain.124, 125
Whilst the prevalence and severity of these symptoms vary across the stages of dementia, most neuropsychiatric symptoms occur in people with moderate to moderately- severe dementia.96, 100, 109, 126 In particular, agitation, psychosis and apathy, often worsen with the progression of dementia.127 However, it should be noted that there is substantial heterogeneity among studies investigating the longitudinal course of neuropsychiatric symptoms in people with dementia.126
A 2014 prospective, observational study involving 146 residents from seven RACFs reported that the average duration of a neuropsychiatric symptom was 2.3 months.128 However, the duration of these symptoms varied widely. For example, the behaviour ‘awakens during the
16 | Page Daniel J. Hoyle night’ persisted for the longest period of time (mean=4.6 months) and ‘Asks or complains about his or her health although it is unjustified’, ‘accuses others of things that are not true’ and ‘tries to do things that are dangerous’ were exhibited for the shortest period (mean=0.5 months)
(Table 3).128 Conversely, a 2016 systematic review of 59 studies found a relatively low persistence for sleep disturbances and elation.126 The systematic review also reported moderate persistence of affective symptoms (e.g., depression, anxiety, apathy), low to moderate persistence of psychosis, and high persistence of hyperactivity and apathy. The authors reported increased psychosis, hyperactivity, agitation and physical aggression with greater cognitive impairment.126
Table 3: Duration of frequently observed neuropsychiatric symptoms. Data from Voyer et al., 2014.128
Neuropsychiatric symptoms Mean duration of symptoms (months) Resists care 3.3 Becomes upset or loses temper easily 2.0 Enter others’ rooms inappropriately 3.3 Awakens during the night 4.6 Talks, mutters, or mumbles to self 2.7 Fights or physically aggressive 3.0 Fidgets, is unable to sit still, restless 2.1 Has difficulty falling asleep 2.7 Says things that do not make sense 2.9 Screams, yells, or moans loudly 2.9 Argues, threatens, or curses 3.4 Tries to get in or out of wheelchair, bed, or chair unsafely 2.6 Asks or complains about his or her health although it is unjustified 0.5 Sees or hear things that are not there 1.7 Disturbs others during the night 2.1 Wanders, tries to escape or go to off-limits places 1.9 Accuses others of things that are not true 0.5 Asks for attention or help, even though it is not needed 2.7 Is uncooperative 2.8 Paces, walks or moves in wheelchair aimlessly back and forth 2.8 Tries to escape physical restraints 1.8 Complains or whines 1.4 Does something over and over although it makes no sense 1.3 Tries to do things that are dangerous 0.5
The variability in prevalence and severity of these symptoms across the stages of dementia has implications for their management. In some scenarios, an antipsychotic may initially be considered appropriate (e.g., the person may be exhibiting aggressive symptoms and
17 | Page Daniel J. Hoyle harming other residents). However, as their condition progresses, they may no longer exhibit these symptoms making the continuation of the antipsychotic inappropriate.126, 128
2.2.4.3.1.3 Neuropsychiatric symptoms in people without dementia
Neuropsychiatric symptoms also present among residents of RACFs who do not have a documented diagnosis of dementia, although the incidence of most symptoms is lower in this cohort.81, 100, 129-132 A 2009 study investigating the prevalence of neuropsychiatric symptoms among people aged ≥65 years with (n=587) and without dementia (n=2,050) reported a higher prevalence of neuropsychiatric symptoms among those with dementia except for sleep disturbances and anxiety, which were experienced at similar rates among both groups.100
A study of 252 residents without dementia residing in 95 Italian RACFs reported depression and anxiety as the most severe and common psychiatric issues.131 Similarly, Lyketsos et al. reported depression and anxiety as the most prevalent symptoms among older people
(n=5092) without dementia in Cache County, Utah.129 The presence of neuropsychiatric symptoms in RACF residents without dementia can be explained, in part, by the aforementioned reduction in the mortality of younger mentally ill people leading to an increased number of residents mental health issues,73 the greater risk of depression anxiety among the Baby Boomer generation,73, 133 and age-related physical conditions and disabilities predisposing older people to psychological issues.74 Additionally, residents may have other conditions which can result in the manifestation of neuropsychiatric symptoms, such as Parkinson’s disease, multiple sclerosis and hypothyroidism.134, 135
Considering the presentation of behavioural and psychological symptoms among people with and without dementia, and the widespread utilisation of antipsychotics and benzodiazepines for these symptoms regardless of an underlying diagnosis of dementia
(discussed in Chapter 3),136-138 the term “neuropsychiatric symptoms” will be used throughout
18 | Page Daniel J. Hoyle this thesis to describe behavioural and psychological symptoms in people with and without dementia, unless otherwise specified.
In summary, neuropsychiatric symptoms are common among residents of RACFs. These symptoms are expected to increase further with an increase in the prevalence of dementia with
Australia’s ageing population, deinstitutionalisation of mental health, reduced mortality of younger mentally ill people, and an increasing number psychological issues associated with age- related physical conditions and disabilities.
2.2.5 Residential aged care facility staff
The number of people employed in direct-care roles in RACFs is increasing. Between 2003 and
2016, the total number of employees in direct-care roles increased from 115,660 (76,006 full- time equivalent positions) to 153,854 (97,920 full-time equivalent positions).65 Despite an increase in the total number of direct-care workers, the proportion of registered nurses (RNs), and enrolled nurses (ENs) decreased and the proportion of the personal care assistant (PCA; also known as a ‘carer’) workforce increased (Figure 11), leaving a skill shortage.65 Other factors, such as staff turnover and time available for direct care, may also be influencing the care of residents in RACFs.65
100% 7.4% 7.6% 7.4% 6.6% 5.2% 5.3% 4.7% 4.0% 90% 80% 70% 56.5% 60% 58.5% 63.6% 64.1% 68.2% 68.2% 70.3% 71.5% 50% 40% 30% 13.1% 14.4% 20% 12.2% 12.5% 11.5% 11.6% 10.2% 9.3% 21.4% 10% 21.0% 16.8% 16.8% 14.9% 14.7% 14.8% 15.2% 0% Headcount FTE 2003 Headcount FTE 2007 Headcount FTE 2012 Headcount FTE 2016 2003 2007 2012 2016
Registered nurses Enrolled nurses Personal care assistants Allied health
Figure 11: Share of the occupations for the direct-care RACF employees (headcount and full-time employment (FTE), %). Adapted from Mavromaras et al., 2017.65
19 | Page Daniel J. Hoyle 2.2.5.1 Skill shortage
Skill shortages were reported by approximately half of Australia’s RACFs in the 2016 Aged Care
Workforce survey.65 A shortage in RNs was the most commonly identified issue (reported by 41% of RACFs).65 Qualitative interviews with 100 participants of the same survey highlighted concern that RNs are being replaced by less qualified (and lower cost) staff (e.g., PCAs), leading to potentially deleterious effects on the quality of resident care.65 This concern has been noted for several years.139-141 In addition to reduced employment of RNs, most of the direct care of residents is falling to the PCA.141 In 2016, 43.5% of RNs spent less than one-third of their time providing direct care, whereas 76.5% of PCAs spent more than two-thirds of their time providing direct care to residents.65 This reflects the increasing managerial role that RNs are performing and the growing role that PCAs have in the direct care of residents.141
As the primary providers of direct care, PCAs play a key role in the detection and management of mental health conditions (e.g., residents are more likely to describe hallucinations to PCAs given their increased contact time with residents).79, 139 However, there are concerns that PCAs do not have sufficient training to detect and manage these conditions.139-141 Failure to effectively detect and manage mental health conditions has deleterious effects for the resident and caregiver.141 For example, a caregiver with limited training may not recognise that a resident may be expressing neuropsychiatric symptoms, such as constant requests for attention, to communicate unmet needs. Instead of investigating the cause of the neuropsychiatric symptoms, as recommended by international guidelines,28, 31, 142 the caregiver may request that the prescriber provides unnecessary and potentially harmful pharmacological therapy.143 On the other hand, conditions with less overt or distressing symptoms (e.g., apathy in depression) may be more likely to go unidentified by RACF staff compared to disruptive behaviour, such as aggression.79, 144 Many GPs rely on RACF staff reports of the residents due to time restrictions that are not conducive towards a comprehensive assessment of the resident (discussed further in Chapter 2.2.6.2).43 As a result, conditions (e.g.,
20 | Page Daniel J. Hoyle depression) with less challenging symptoms (e.g., apathy) are more likely to go underdiagnosed and undertreated.145
Whilst there are calls to regulate the minimum number of nurses in RACFs,141, 146 there is also concern that RNs and ENs have little training in old age mental health and dementia.65 In the 2016 Aged Care Workforce survey, the majority of RNs (51.3%) and ENs (54.9%) identified dementia training as the area of training most needed over the next 12 months.65 Low levels of knowledge related to dementia care has also been reported among nurses in RACFs.147, 148 This is alarming considering that more than half of residents in Australian RACFs have a diagnosis of dementia.7 Research has also identified deficiencies in training and knowledge in late-life depression, suicide and other mental health issues commonly presenting in RACFs.149, 150
2.2.5.2 Turnover
Residential aged care providers face difficulties in recruiting and retaining direct-care staff.65, 141,
151 A 2016 survey of RACF staff found that a substantial proportion of RNs (17.1%), ENs (20.3%) and PCAs (15.7%) were actively seeking alternative work despite working in their current job for
≤12 months.65 In 2008, Martin and King reported 12-month turnover rates of 25% for PCAs and
20% for nurses.140 They commented that turnover rates have not changed significantly since
2003.140
High staff turnover has implications in the RACF setting. First, training of direct-care workers in an environment where a substantial proportion is likely to leave within a year may not be considered cost-effective by RACF providers.152 Second, studies have identified worse outcomes for residents of RACFs with high nurse turnover rates.153, 154 For example, Zimmerman et al. found that increased RN turnover was associated with an increased risk of infections and hospitalisations.154 Staff turnover has also been found to negatively affect the ability of staff to manage the behavioural issues of residents.155 Additionally, high staff turnover is problematic for ensuring sustained effects following educational interventions. Despite initial improvements
21 | Page Daniel J. Hoyle in staff knowledge and skills following training, outcomes do not seem to be sustained over time.156, 157
A major factor affecting the recruitment and retention of nurses in the RACF setting is the wage gap.65, 141, 158 Registered nurses in RACFs earn, on average, approximately $200 less per week (calculated on the base rate) than their colleagues in other sectors.65, 140, 141, 159, 160 The disparity in pay has been blamed on the Federal Government who set the level of funding that aged care service providers receive.160 Additionally, PCAs receive between $715 to $868 as a minimum per week, which is approximately half the average full-time adult weekly wage in
Australia ($1516).160 Alzheimer’s Australia has called for improved remuneration to mitigate the high staff turnover.141
2.2.5.3 Time available for residents
Currently, there are no legislated resident-to-staff ratios in the Australian RACF setting.152
Qualitative interviews following the 2016 Aged Care Survey, indicated that high workloads due to inadequate staff numbers and excessive amounts of administrative tasks and paperwork reduced the time available to care for the residents.65 The lack of time to provide quality care was also noted in the 2012 Aged Care Survey.161 Staffing issues were a major concern among all direct-care workers.65 Interviewees felt that funding constraints were contributing to low staff numbers within RACFs, including a lack in covering regular staff member absences and staff not being replaced when they retired.65 Consequently, this has led to increased resident-to-staff ratios, perceived increased workload pressures and poorer quality of care for the residents,65,
139, 141, 162 particularly at night when the care needs of people with neuropsychiatric symptoms are at their highest.152 A 2018 study of 40 staff from eight Sydney RACFs found that inappropriate psychotropic medication use was often a result of staff feeling helpless to provide person-centred care for neuropsychiatric symptoms within the restrictive timeframes imposed by the RACF.43
22 | Page Daniel J. Hoyle 2.2.6 General practitioners
GPs are primarily responsible for the medical management of Australian RACF residents.163-165
According to the latest Australian Medical Association (AMA) Aged Care Survey, the majority of respondents (63.8%) provided care to RACF residents.146 However, there is concern about the reduction in numbers of GPs providing these services,146, 166 reliance on information provided by nurses/carers,167 and the lack of specialist services utilised, despite the high prevalence of mental health issues in this group of older people.163 These concerns are explored below.
2.2.6.1 General practitioner residential aged care services
In 2017, the AMA Aged Care Survey was distributed to 5,599 AMA members, 608 responded.146
Whilst the majority (63.8%) of respondents indicated that they currently provide services to
RACFs, 10.5% of respondents had stopped visiting RACFs in the past five years and 35.7% of GPs who currently had patients in RACFs intended to decrease the number of visits (20.2%), not take on new patients (7.0%), or stop visiting RACFs altogether in the next two years (8.5%).146 There is also concern about the impending shortage of GPs providing services to RACFs when those currently providing these services retire.146, 164, 165, 168 Only a small proportion (3.6%) of GPs aged
<40 years provide RACF services.146 In comparison, GPs aged >60 years currently provide almost half (47.1%) of the monthly RACF visits.146
Almost half (48.5%) of the respondents who had decreased the number of visits in the past five years agreed that the reason they had decreased their visits was because consultations increasingly involved performing unpaid, non-face-to-face tasks (e.g., locating residents, writing prescriptions and paperwork, talking to relatives, discussing issues with RACF staff while in the surgery).146 The average reported length of non-face-to-face time per consultation was 13.58 minutes.146 Furthermore, a third of the respondents strongly agreed that patient rebates did not adequately compensate for lost time in the surgery.146 These issues have been consistently
23 | Page Daniel J. Hoyle identified in research investigating the reasons that influence GPs to reduce their RACF services.165, 168, 169
2.2.6.2 Reliance on nurse and carer information
As discussed in Chapter 2.2.4, the complex health needs of RACF residents present a major challenge to their GPs. Resident issues, particularly of a psychogeriatric nature, require a thorough assessment prior to a GP being able to intervene with confidence.79 Unfortunately, large patient loads and time restrictions create scenarios that are not conducive towards the comprehensive performance of these assessments.43 Instead, many GPs rely on nursing staff reports of the residents.43, 167 In Chapter 2.2.5, it was mentioned that PCAs spend the most time providing direct care to the residents and, hence, are in the best position to detect these issues.
However, with minimal training, many issues go unidentified.139-141, 167 In particular, symptoms that may not be considered ‘challenging’, such as apathy, tend to be underreported by RACF staff.79 Staff unfamiliar to the residents may also unintentionally trigger behavioural and/or psychological symptoms.152 With little training or time to implement non-pharmacological management strategies for these symptoms (e.g., massage therapy, exercise therapy),79, 170, 171 a psychotropic medication, such as an antipsychotic, may be initiated as a ‘quick-fix’ or
‘workaround’.43, 171
With a reliance on information from the nursing staff, the initiative for prescribing psychotropic medication, like antipsychotics, often comes from the nursing staff of RACFs.148,
172, 173 Studies investigating non-resident related factors influencing the prescribing of psychotropic medications have reported cases of pressure from the nursing staff to prescribe.162,
171, 172 Of particular concern, Cornege-Blokland et al. reported that the prescriber did not support the prescription of an antipsychotic in 8% of cases but was pressured into prescribing the medication by the nursing staff.172 In 2011, Zuidema et al. found that the odds of being prescribed an antipsychotic or anxiolytic were 66% and 62% higher, respectively, if staff
24 | Page Daniel J. Hoyle experienced distress related to resident agitation.143 More recently, the prescribing of other psychotropic medications have also been shown to increase with staff distress.174
2.2.6.3 Specialist availability
Despite the high prevalence of mental health issues in RACFs, GPs generally have limited training in the management of these issues.167, 175 As a result, there is some evidence of difficulties when diagnosing these conditions. For example, a cross-sectional audit of general practice (n=46,515 patients) in Australia found that GPs (n=386) did not recognise the presence of common mental conditions in 56% of patients, particularly in patients with predominantly somatic symptoms.176 As a result of the limited training, many GPs do not feel adequately prepared to manage the complex care issues of RACF residents.165, 177 Referral to a specialist practitioner may be beneficial in more complex cases. For example, a cluster-randomised controlled trial found improvements in depression with specialist mental health consultation compared to usual care for RACF residents with clinical depression.178 However, difficulties facilitating patient access to specialists has been reported. In the 2017 AMA Aged Care Survey, most respondents indicated that providing patient access to specialists was difficult (43.6%) to very difficult (27.6%) and nearly half (49.3%) of respondents called for improved access to specialist care.146 Furthermore, GP participants in qualitative research have indicated extreme difficulty when trying to get specialist interest in their RACF resident.165
Up until May 2018, RACF residents were not able to access subsidised psychological interventions for mental health issues through Medicare.179 After claiming that residents of
RACFs are being denied equal access to psychological services,180 the Australian Government allocated $82.5 million for psychological services for older Australians in RACFs with a diagnosed mental health condition.179 These funds will bring psychologists and other eligible mental health providers into the RACF setting where there has previously been limited funding.179 It is hoped that increased psychological support will help support the implementation of non-
25 | Page Daniel J. Hoyle pharmacological strategies and minimise the use of psychotropic medications.179 However, these psychological services do not include the management of residents with dementia.181
Furthermore, specialist recommendations are often frustrated by systemic factors, such as insufficient staffing and time available to implement recommendations.163
2.2.7 Medication management programs in Australian residential aged care facilities
2.2.7.1 Residential medication management reviews
In Australia, permanent residents of Government-funded RACFs are eligible to receive residential medication management reviews (RMMRs) once every 24 months (unless there is a clinical need for an earlier review in the intervening period).182 RMMRs are defined as “a resident-focussed, collaborative, comprehensive medication review involving the systemic evaluation of the resident’s complete medication regimen and management of that medicine in the context of other relevant clinical information and the resident’s health status”.183
Accredited pharmacists can perform the medication review upon written referral by a resident’s
GP.182 In Australia, accredited pharmacists are registered pharmacists who have undergone training and have demonstrated their competence to perform medication review services.184
Pharmacists can become accredited through two governing bodies; the Australian Association of Consultant Pharmacy or the Society of Hospital Pharmacists of Australia.184
The RMMR is conducted in collaboration with the resident’s GP, as well as the RACF nursing staff, residents and their family members, where appropriate.182 After performing the review, the accredited pharmacist is required to provide a written report to the resident’s GP stating the findings of the review and outlining recommendations (e.g., alterations to medications) to assist in the development of a Medication Management Plan by the GP.182
RMMRs have been remunerated for pharmacists since 1997.183
26 | Page Daniel J. Hoyle A 2019 systematic review investigated the outcomes associated with medication reviews (n=13 studies; eight studies were based on RMMRs).185 This review found that comprehensive medication reviews successfully identified and resolved medication-related issues.185 A small amount of evidence suggested that RMMRs may reduce the use of sedative medications in RACFs.186 However, clinical outcomes associated with RMMRs are lacking.185
2.2.7.2 Quality use of medicines
Quality use of medicine (QUM) services are separate from RMMR services and can be provided by a registered or accredited pharmacist.182 To develop a QUM service, pharmacists work with stakeholders to improve medication management at the facility level.182 Examples of QUM activities include participation in drug use evaluations at the RACF, participation in Medication
Advisory Committees, and provision of in-service education for nursing staff and carers. The impact of these services on medication use and clinical outcomes is unclear.
2.3 Summary
An ageing population with increasingly complex health needs is expected to add enormous pressure on the already overwhelmed aged care industry. Those entering the RACF setting are older and have more complex health needs than ever before.
Many residents have high levels of behavioural symptoms due to increasing rates of dementia, deinstitutionalisation of mental health, reduced mortality of younger people with mental health issues, increased susceptibility of mental health disorders among Baby Boomers, and mental health issues related to physical disabilities. Despite the demanding nature of behavioural symptoms commonly encountered in the RACF setting, most of the direct resident care is provided by PCAs with minimal training. Other systemic factors affecting the management of behavioural symptoms in RACFs includes a high turnover of staff, high resident- to-staff ratios, declining GP visits, and lack of specialist input.
27 | Page Daniel J. Hoyle These shortcomings have created an environment in many RACFs where pharmacological management is often considered the only viable option by RACF staff and GPs.
A major class of medications that have come under scrutiny for their highly prevalent but inappropriate usage in RACFs are psycholeptic medications, including antipsychotics and benzodiazepines. Chapter 3 provides an overview of the use of these medications, their common indications in the RACF setting, potential adverse effects, and barriers to the reduction in their use.
28 | Page Daniel J. Hoyle Chapter 3 Antipsychotic and benzodiazepine use in Australian residential aged care facilities
3.1 Introduction
Chapter 2 introduced several challenges faced by direct-care staff and prescribers when providing care for RACF residents with increasingly prevalent mental health disorders. These challenges are considered to be contributors to the over-reliance on antipsychotics and benzodiazepines in RACFs worldwide.187
Chapter 3 explores the use of antipsychotics and benzodiazepines 1 in RACFs.
Specifically, the reader will be introduced to the mechanism of action for these medications, indications for their use in the RACF setting, efficacy for these indications, adverse effects and their potentially inappropriate use in RACFs. This chapter will conclude with an investigation into the barriers of antipsychotic and benzodiazepine dose reduction in the RACF setting.
The focus on antipsychotic and benzodiazepine medications stems from the overarching aim of this thesis to examine the clinical outcomes of a multifaceted intervention to improve antipsychotic and benzodiazepine review and reduction in RACFs, called RedUSe.50
On the other hand, antidepressant medications were not targeted in RedUSe due to conflicting evidence over the use of this psychotropic class in older people.188, 189 Indeed, some advocates think that depression is undertreated in RACFs.145
1 Note: Z-drugs (e.g., zolpidem, zopiclone) were not included in this study considering their relatively low usage at the time in Australian RACFs.
29 | Page Daniel J. Hoyle 3.1.1 Definitions
Antipsychotics, anxiolytics, and hypnotics and sedatives (‘anxiolytics’ and ‘hypnotics and sedatives’ are predominantly benzodiazepine medication) are classified as ‘psycholeptic’ medications under the World Health Organization’s Anatomical Therapeutic Chemical (ATC) classification code ‘N05’.190 Table 4 lists these drug classes, their ATC classification code, and medications from these groups that are marketed in Australia.
Table 4: Psycholeptic drug groups. Data from the World Health Organisation Collaborating Centre for Drug Statistics Methodology, 2016.190
Psycholeptic drug group ATC classification Medications available in Australia code Antipsychotics N05A Amisulpride, aripiprazole, asenapine, brexpiprazole, chlorpromazine, clozapine, droperidol, flupentixol, haloperidol, lithium, lurasidone, olanzapine, paliperidone, periciazine, quetiapine, risperidone, ziprasidone, zuclopenthixol Anxiolytics N05B Alprazolam, bromazepam, buspirone, clobazam, diazepam, lorazepam, oxazepam Hypnotics and Sedatives N05C Chloral hydrate, dexmedetomidine, flunitrazepam, melatonin, midazolam, nitrazepam, suvorexant, temazepam, thiopental, zopiclone, zolpidem ATC=Anatomical Therapeutic Chemical
For the purpose of this thesis, benzodiazepine-type anxiolytics and hypnotics and sedatives will be called benzodiazepines. Furthermore, antipsychotics and benzodiazepines will collectively be referred to as ‘sedative’ or ‘psycholeptic’ medications.
Clonazepam, a benzodiazepine derivative, although classified under the ATC classification code ‘N03A- antiepileptic’,190 will also be included in this study as a benzodiazepine considering its role in the mitigation of acute agitation, anxiety and psychotic symptoms.27 In
Australia, clonazepam is indicated for status epilepticus (as an intravenous product) and epilepsy refractory to other antiepileptics.191, 192 Whilst reduction of this medication could be considered inappropriate due to an increased risk of seizures, there is evidence clonazepam, particularly oral drops, is used off-label for anxiety and behavioural disturbances in older people.10, 193 Clonazepam has a long half-life (~39 ±8.3 hours),194 which, particularly in older
30 | Page Daniel J. Hoyle people, may lead to increased next-day sedation and adverse effects, such as falls and fractures.195-197 The long half-life also prolongs the time to reach plasma steady-state, which may take weeks to achieve in older people. This limits the utility of clonazepam as a ‘when required’ medication for behavioural symptoms.198
3.2 Antipsychotics
Antipsychotics were discovered in the 1950s when it was noticed the antihistamine, chlorpromazine, improved the clinical symptoms of 38 patients with mania and psychosis.199, 200
Prior to this, the management of psychiatric conditions largely revolved around electroconvulsive therapy, insulin coma, frontal lobotomy, and sedation.200 After further investigation, it became clear that chlorpromazine exhibited its effect by blocking dopamine
199 receptors, particularly postsynaptic dopamine (D2) receptors. This led to the development of what we now know as typical (first-generation) antipsychotics over the next three decades.199
Whilst these medications were effective for aggressive and psychotic symptoms, they were associated with serious neurological adverse effects, such as extrapyramidal adverse effects
(movements disorders, such as dystonia, akathisia, bradykinesia, tremor, tardive dyskinesia).200
Efforts to minimise these adverse effects led to the advent of atypical (second-generation) antipsychotics. The first atypical antipsychotic, clozapine, was gradually introduced across
Europe during the early 1970s.201 However, clozapine was voluntarily withdrawn by the manufacturer in 1975 after cases of agranulocytosis were reported among older and infirm
Finnish patients.201-203 In 1990, clozapine was released on the US market alongside the strict requirement for regular monitoring of white blood cell counts.203, 204
Several atypical antipsychotics have since been developed (e.g., risperidone, olanzapine, quetiapine).199 Whilst the mechanism of atypical antipsychotics is still debated, it has been hypothesised that they, in general, strongly block serotonin (5-HT2A) receptors and weakly block
199 D2 receptors and, thus, are associated with a lower risk of extrapyramidal symptoms.
31 | Page Daniel J. Hoyle However, increased affinity to other receptors can increase the risk of a variety of other adverse effects, such as weight gain and hyperglycaemia.205 Exceptions to this information include amisulpride and risperidone which have significant activity at D2 receptors and may cause movement disorders at higher doses.205, 206 Furthermore, aripiprazole, lurasidone and ziprasidone have little to no effect on weight gain.205
3.2.1 Mechanism of action
Antipsychotics are commonly divided into two groups based on their hypothesised modes of action; Typical (first-generation) and atypical (second-generation) antipsychotics.102, 207 Table 5 describes the antipsychotics available in Australia and the pharmacological group they belong to.
Table 5: Antipsychotic medications used in Australia. Information from the Australian Medicines Handbook, 2018.205
Typical antipsychotics Atypical antipsychotics Chlorpromazine Amisulpride Droperidol Aripiprazole Flupentixol Asenapine Haloperidol Brexpiprazole Periciazine Clozapine* Trifluoperazine Lurasidone Olanzapine Paliperidone Quetiapine Risperidone** Ziprasidone Zuclopenthixol *Listed on the Pharmaceutical Benefits Scheme2 as an authority item for schizophrenia. To receive this item under the Pharmaceutical Benefits Scheme, the patient must be treated by a psychiatrist or by an authorised medical practitioner, with the agreement of the treating psychiatrist.208
**Listed on the Pharmaceutical Benefits Scheme as an authority item for psychotic symptoms and aggression for up to 12 weeks in people with dementia of the Alzheimer type where non-pharmacological methods of treatment have failed.209
2 The Pharmaceutical Benefits Scheme is an Australian Government program that subsidises medicines to make them more affordable (https://www.pbs.gov.au/info/general/faq).
32 | Page Daniel J. Hoyle 210 Whilst the exact mechanism of action is unclear, all antipsychotics antagonise D2 receptors in the CNS to varying degrees.207 In particular, antagonism of dopamine activity in the mesolimbic tract has been shown to reduce psychotic symptoms.211 Blockade of dopamine activity in the nigrostriatal tract and tuberoinfundibular tract is associated with movement
(extrapyramidal) disorders and endocrine effects, respectively.207
Generally, typical antipsychotics have stronger affinities for D2 receptors and atypical
207, 212 antipsychotics are more selective towards D3, D4 and 5-HT2 receptors (Table 6). The variation in receptor affinities may explain differences in adverse effects observed with the use of these medications (Table 7).207 To illustrate, typical antipsychotics (e.g., haloperidol), which have high affinities for D2 receptors, are associated with a higher risk of extrapyramidal side
205 effects. Generally, atypical antipsychotics (e.g., quetiapine), which have lower D2 receptor affinities, are associated with a lower risk of extrapyramidal side effects but may confer a higher risk of dry mouth, postural hypotension and sedation considering higher affinities for other receptors, such as cholinergic, adrenergic and histaminergic receptors.205, 207
Table 6: Example of differences in relative receptor binding affinities for antipsychotics. Adapted from Neil et al., 2003.207
Medication D1 D2 D3 D4 M1 α1 5-HT2 H1 Amisulpride - ++++ ++++ ++ - - - - Chlorpromazine + +++ - - ++ +++ ++ ++ Clozapine ++ +++ ++ ++++ ++++ ++++ ++++ ++++ Haloperidol - ++++ - - - + + - Olanzapine + ++ ++ ++ ++ + ++++ +++ Quetiapine + ++ ++ + + ++ ++ ++ Risperidone - +++ ++ ++ - +++ ++++ +++ ++++=Very high affinity; +++=High affinity; ++=Moderate affinity; +=Weak affinity; -=Absent affinity (D=Dopaminergic; M=Muscarinic; α=Adrenergic; 5-HT=Serotoninergic; H=Histaminergic)
33 | Page Daniel J. Hoyle Table 7: Comparison of antipsychotic adverse effects. Information from referenced sources.205, 206
Medication Movement Sedation Anticholinergic Orthostatic Weight gain Hyperglycaemia Prolactin release Other disturbances effects hypotension Amisulpride + (++ if dose + +/- + + +/- +++ Prolonged QT interval >400mg/day) Chlorpromazine + (++ for +++ +++ +++ ++ ++ ++ Photosensitivity higher doses) Haloperidol +++ + + + + + ++ Prolonged QT interval
Olanzapine + ++ ++ + +++ +++ +
Quetiapine +/- ++ + ++ ++ ++ +/-
Risperidone + (++ if dose + +/- ++ ++ + +++ Sexual/ejaculatory >6 mg/day) problems (due to raised prolactin) +++=common, ++=infrequent, +=rare, -=absent.
NB Prolongation of the QT interval is possible with all antipsychotics. However, amisulpride and haloperidol carry a relatively higher risk.
34 | Page Daniel J. Hoyle 3.2.2 Use in residential aged care facilities
Antipsychotics are primarily indicated for severe psychiatric conditions, such as schizophrenia and bipolar disorder, usually under the supervision of a psychiatrist.213, 214 In patients with schizophrenia or related psychoses, antipsychotics are the cornerstone of pharmacological therapy due to their ability to reduce positive symptoms (e.g., hallucinations, delusions, thought disorder) and prevent relapse.18, 20, 215 Similarly, antipsychotics are effective in the management of acute mania in patients with bipolar disorder and are commonly given alongside mood stabilisers, such as lithium or sodium valproate.19, 21, 216 However, in the RACF setting antipsychotics are more commonly used to manage neuropsychiatric symptoms (Chapter 3.2.3), associated with dementia, usually under the supervision of a GP.25, 213, 214, 217
Antipsychotics, especially risperidone, have short-term evidence of modest efficacy for select neuropsychiatric symptoms in dementia, including agitation, aggression and psychosis
(Chapter 3.2.4).22-24 However, long-term use (>6-12 weeks) or use for other neuropsychiatric symptoms in people with dementia is generally not supported by the literature and, consequently, may be considered potentially inappropriate.28, 31, 106 Considering the modest efficacy of antipsychotics in the management of neuropsychiatric symptoms associated with dementia (Chapter 3.2.4) and risk of severe adverse effects (Chapter 3.2.5), clinical guidelines generally recommend reserving antipsychotics for residents in whom non-pharmacological therapies are ineffective or situations where there is a risk of harm to the resident or others.28,
31, 98, 106, 142, 218-220
There are numerous non-pharmacological therapies for the management of neuropsychiatric symptoms in people with dementia.31 Examples of non-pharmacological therapies include activity and recreation, aromatherapy, family tape-records, music and sound therapy, social interaction, and exercise.31 The choice of non-pharmacological therapy can be guided by careful assessment of the resident to identify possible antecedents (triggers) of the
35 | Page Daniel J. Hoyle neuropsychiatric symptoms and what makes them better or worse.28, 31, 221 There is substantial variability in the reported effectiveness of non-pharmacological therapies for neuropsychiatric symptoms in dementia.222 Recent reviews, however, have suggested that non-pharmacological therapies, particularly music therapy and behavioural management techniques, are generally effective for neuropsychiatric symptoms in people with dementia,222, 223 possibly even more effective than pharmacological agents, like antipsychotics, for reducing aggression and agitation in people with dementia.223
According to the Royal Australian and New Zealand College of Psychiatrists (RANZCP) guidelines,106 antipsychotics should only be used in people “experiencing significant agitation, aggression or psychosis” unresponsive to psychosocial/non-pharmacological interventions or where there is a severe and complex risk of harm. Furthermore, antipsychotic therapy should be avoided for symptoms that are unlikely to respond to them, such as wandering, undressing, inappropriate voiding, verbal aggression or screaming.106
In Australia, risperidone is the only atypical antipsychotic listed on the Pharmaceutical
Benefits Scheme (PBS) for neuropsychiatric symptoms in people with Alzheimer’s dementia.209
It is reserved for patients with Alzheimer’s disease exhibiting psychotic (e.g., hallucinations, delusions) and aggressive symptoms which have not responded to non-pharmacological methods of treatment.28, 209 The treatment duration must be limited to a maximum of 12 weeks.28, 209 In contrast, risperidone is only licensed in the UK for up to six weeks for the treatment of persistent aggression in patients with moderate to severe Alzheimer’s dementia.224 The RANZCP and Psychotropic Expert Group recommend using risperidone at doses of up to 2 mg per day.28, 106 Furthermore, doses should be started low and slowly incremented based on the effect with the aim to use the lowest effective dose for the shortest period of time.28, 106 Other antipsychotics (e.g., quetiapine, olanzapine, aripiprazole) should only be considered if risperidone is not tolerated or is inappropriate.106
36 | Page Daniel J. Hoyle Once the management strategy has been selected, neuropsychiatric symptoms should be monitored to determine the effectiveness of therapy.106 The RANZCP suggest using a validated instrument, such as the Neuropsychiatric Inventory, the Behave-AD or the Cohen-
Mansfield Agitation Inventory, to monitor neuropsychiatric symptoms, although the practicality of using these scales in the RACF environment is questionable.106 If ineffective or causing adverse effects, the antipsychotic should be ceased prior to trialling another medication (i.e., another antipsychotic or other drugs). Since neuropsychiatric symptoms in dementia are generally intermittent, the duration of antipsychotic therapy should be time-limited with review and potential dose-reduction at least every 12 weeks.106 Generally, the antipsychotic medication should be withdrawn if symptoms are controlled.225 Antipsychotic withdrawal is also recommended if there is no response to therapy within four weeks, considering that lack of early improvement in symptoms with an antipsychotic is a marker of subsequent nonresponse.142, 225
When dose reduction is considered, the most recent Canadian antipsychotic deprescribing guidelines recommend tapering the dose by 25 to 50% every one to two weeks.225
During dose tapering, the patient should be monitored for withdrawal events (e.g., psychosis, aggression, agitation, delusions, hallucinations).225 If symptoms relapse, non-pharmacological approaches should be considered (e.g., relaxation therapy, social contact, music therapy, aromatherapy)225 and, if necessary for severe symptoms, the antipsychotic can be restarted at the lowest dose possible or an alternative drug could be considered.225 At least two cessation attempts should be attempted before excluding people from future discontinuation attempts.225
Despite these guidelines and risk of adverse effects, antipsychotics are often used inappropriately on a long-term basis (Chapter 3.4).12, 25 Considering that antipsychotics are often used inappropriately in RACFs for the treatment of neuropsychiatric symptoms, RedUSe aimed
37 | Page Daniel J. Hoyle to promote the appropriate use of antipsychotic medications for neuropsychiatric symptoms associated with dementia.50 Consequently, this thesis will focus on the use of antipsychotics for neuropsychiatric symptoms associated with dementia.
3.2.3 Impact of neuropsychiatric symptoms
Neuropsychiatric symptoms can have a substantial impact on the resident, RACF staff and increase costs to the government.
Neuropsychiatric symptoms have been identified as an independent risk factor for falls among residents of RACFs.171, 172 In 2012, Syllias et al. investigated risk factors for falls in 1,147
RACF residents in Norway.226 Alongside dementia severity, depression severity and use of sedatives, a higher NPI score was identified as the strongest predictor for residents having one or more falls.226 Furthermore, a 2018 Japanese study of 242 residents of RACFs reported an increased risk of falls among residents displaying symptoms of wandering (hazard ratio=2.23,
95%CI=1.35, 3.68) and agitation (hazard ratio=1.94, 95%CI=1.24, 3.04) more than once a week at baseline.227
Neuropsychiatric symptoms are also associated with deterioration in QoL.228-231 In a study of 290 residents with dementia from nine Dutch RACFs, a greater Neuropsychiatric
Inventory-Nursing Home version (NPI-NH) total score consistently predicted lower QoL for the resident.231 The use of antipsychotics did not have a significant effect on QoL.231 Additionally,
Samus et al. found that agitation, depression, apathy and irritability were significant predictors of QoL (explaining 29% of the QoL variance) among 134 US residents residing in 22 RACFs.230
Similarly, Shin et al. reported that QoL was negatively correlated with the NPI total score and domain scores related to agitation/aggression, anxiety, disinhibition, and irritability/lability.229
Lastly, a cross-sectional study of 101 people with dementia found that higher NPI scores suggested that high levels of neuropsychiatric symptoms are associated with decreased QoL.228
A multivariate model which included the results of the NPI, Mini-Mental State Examination,
38 | Page Daniel J. Hoyle General Health Questionnaire and Barthell Index, and age of the patient, was able to account for 35% of the variation in QoL overall.228
Neuropsychiatric symptoms are associated with increased caregiver stress and premature admission to RACFs.232, 233 Once in a RACF, neuropsychiatric symptoms, particularly sleep and night-time behaviour disorders, and delusions and agitation/aggression may distress
RACF care staff and lead to increased sick days, burnout and potentially increased staff turnover.144, 234-236 Qualitative interviews of 25 staff from six UK RACFs reported that staff often felt powerless, frightened and overwhelmed by residents with dementia displaying agitation.237
Generally, staff wanted to deliver person-centred care to manage these symptoms but felt they struggled to do so when feeling overwhelmed and unsafe.
These symptoms may also increase the costs of care due to greater use of health services and direct care costs.238-240 A prospective cohort study of 224 people with Alzheimer’s disease in the UK found that those with agitation had higher health and social care costs.238
After adjusting for demographic variables, coexisting conditions, cognitive impairment, follow- up and individual clustering for repeated measures, the mean costs varied from £29,000 over a
12-month period in residents with no agitation (NPI agitation score=0) to £57,000 in residents with severe agitation (NPI agitation score=12, p=0.01).238 For each unit increase in the NPI agitation domain score, there was a £1,064 increase in costs per patient over a 12-month period
(95%CI=-£34, £2162, p=0.058) when adjusting for covariates.238 An earlier study by Murman et al. reported increased direct health care utilisation costs associated with higher NPI scores among 128 patients with probable Alzheimer’s disease in the US.239 Models estimated that a one-point increment in the NPI score would result in an annual increase in costs between $27 and $409 (depending on the value of the unpaid caregiving).239
39 | Page Daniel J. Hoyle 3.2.4 Effectiveness of antipsychotics for neuropsychiatric symptoms associated with dementia
Despite being commonly prescribed for neuropsychiatric symptoms in people with dementia, the effectiveness of antipsychotics for these symptoms are generally modest.
Most data reporting on the efficacy of typical antipsychotics for neuropsychiatric symptoms in people with dementia relates to haloperidol. A Cochrane systematic review of five randomised controlled trials (RCTs) reported that haloperidol, at doses between 1.2 to
3.5 mg/day, was modestly effective for aggression (effect size=0.31, 95%CI=0.49, -0.13; p=0.0006), as measured with the Multidimensional Observation Scale for Elderly Subjects
(MOSES)-irritability/aggressiveness subscale, the Behavior Pathology in Alzheimer’s Disease
Rating Scale (BEHAVE-AD) and the Behavioural Syndromes Scale for Dementia.241 However, haloperidol was ineffective for the treatment of general behavioural symptoms or agitation when compared to placebo.241 Furthermore, Lonergan et al. found that those taking haloperidol were twice as likely to drop out due to adverse effects compared to placebo.241 Therefore, it is unclear whether these benefits on aggression outweighed the risk of adverse effects associated with haloperidol.
In 2006, a meta-analysis of 15 RCTs compared atypical antipsychotics (aripiprazole, olanzapine, quetiapine and risperidone, n=3,353 patients) to placebo (n=1,757 patients) for the treatment of neuropsychiatric symptoms in people with dementia over eight to 12 weeks.242
Using trials which defined a clinical response on the basis of a 30 to 50% improvement on several rating scales, a number needed to treat between five to 14 was calculated.242 This indicates that between five and 14 patients need to be treated for improvement (a 30 to 50% improvement) in one patient.
In 2006, Ballard et al. performed a Cochrane systematic review (n=16 RCTs) to investigate the effectiveness of atypical antipsychotics for aggression and psychosis in patients
40 | Page Daniel J. Hoyle with Alzheimer’s disease.22 Nine of these studies were included in the meta-analysis. Five trials examined the effectiveness of risperidone over 10-13 weeks at doses between 0.5 to 2.0 mg/day. Risperidone showed statistically significant benefits in overall behaviour (0.5-2.0 mg/day), aggression (1.0 and 2.0 mg/day) and psychosis (1.0 mg/day) compared to placebo.22
Four studies assessed the effectiveness of olanzapine over six to 10 weeks. Olanzapine, at doses of 5-10 mg/day, showed benefit compared to placebo for aggression, anxiety and euphoria.22
Despite modest efficacy of risperidone and olanzapine for aggression and risperidone for psychosis, Ballard et al. discouraged the routine use of antipsychotics in people with dementia unless symptoms are severely distressing or pose a risk of physical harm to those living and/or caring the person.22 These recommendations were made in consideration of the severe adverse effects associated with antipsychotics, such as an increased rate of cerebrovascular events and mortality.22
A 2018 overview of systematic reviews has also reported modest efficacy of atypical antipsychotics on neuropsychiatric symptoms in people with dementia.243 The authors summarised the results of 14 studies (n=3,158 participants). Atypical antipsychotics had a small but statistically significant effect on the overall neuropsychiatric symptoms (effect size=-0.13,
95%CI=-0.21, -0.06).243 However, this systematic review was limited by reviewing atypical antipsychotics as a class and did not review the effects on specific neuropsychiatric symptoms.243
In 2011, Maher et al. pooled the results of 18 placebo-controlled trials reporting on changes in overall neuropsychiatric symptoms, psychosis and agitation between six and 12 weeks in people with dementia.23 Compared to placebo, olanzapine, risperidone and aripiprazole were modestly effective for overall neuropsychiatric symptoms. The pooled effect size was 0.12 (95%CI=0, 0.25) for olanzapine, 0.19 (95%CI=0, 0.38) for risperidone and 0.20
(95%CI=0.04, 0.35) for aripiprazole.23 The effect size for quetiapine (0.11) was not statistically
41 | Page Daniel J. Hoyle significant.23 Risperidone also demonstrated modest efficacy in the management of psychosis with a small pooled effect size of 0.20 (95%CI=0.05, 0.36; five trials). The effect sizes for aripiprazole (0.20), olanzapine (0.05) and quetiapine (-0.03) for psychosis were not significantly more effective than placebo.23 Lastly, risperidone, aripiprazole and olanzapine were all associated with small but statistically significant improvements in agitation with pooled effect sizes between 0.19 and 0.31 compared to placebo. The pooled effect size of quetiapine (0.05,
95%CI=-0.14, 0.25) was not statistically significant.23 In summary, these results suggest that risperidone, olanzapine and aripiprazole are modestly effective for overall neuropsychiatric symptoms and agitation. Additionally, risperidone demonstrated modest efficacy for psychosis.
On the other hand, quetiapine had no evidence to support its use for neuropsychiatric symptoms in people with dementia.
Particular symptoms are more likely to respond to antipsychotic treatment. The Clinical
Antipsychotic Trials of Intervention Effectiveness: Alzheimer’s disease (CATIE-AD) trial is the largest study to date investigating the efficacy and safety of atypical antipsychotics for neuropsychiatric symptoms in people with dementia.244 This multicentred, double-blinded, placebo-controlled trial involved 421 outpatients with Alzheimer’s disease and symptoms of psychosis, aggression or agitation. The patients were randomly assigned to receive olanzapine, quetiapine, risperidone or placebo. The main outcome was time to discontinuation for any reason. Overall, the time to discontinuation was similar between all four groups. However, there were suggestions that olanzapine and risperidone were less likely to be discontinued for lack of efficacy compared to placebo and quetiapine.244
A post-hoc analysis of the CATIE-AD trial compared the change in individual symptoms between baseline and the last observation in Phase 1 of the CATIE-AD trial among participants receiving olanzapine (n=99), quetiapine (=94), or risperidone (n=84), against placebo (n=139).24
Compared to placebo, patients treated with olanzapine (least-squares mean (LSM) difference=-
42 | Page Daniel J. Hoyle 0.4, 95%CI=-0.7, -0.1; p=0.006) or risperidone (LSM difference=-05, 95%CI=-0.8, -0.2; p=0.003) showed greater improvements in the hostile suspiciousness factor (assesses hostility, aggression, mistrust, and uncooperativeness) as rated on the Brief Psychiatric Rating Scale
(BPRS).24 Furthermore, patients treated with risperidone had significant improvements in psychosis, as rated on the BPRS, compared to placebo (LSM difference=-0.5, 95%CI=-0.8, -0.1; p-0.01).24 Conversely, withdrawn depression, as measured with the BPRS, worsened in patients treated with olanzapine compared to placebo (LSM difference=0.3, 95%CI=0.1, 0.5; p=0.003).24
Additionally, quetiapine was no more effective than placebo for any individual symptom.24 This study provides further support that olanzapine and risperidone can be effective for hostile suspiciousness symptoms, like aggression, and risperidone can be used for the management of psychosis.23, 24, 245 Further to this, worsened withdrawn depression with olanzapine use suggests antipsychotics have differential effects in individual symptoms of Alzheimer’s disease.24
Published data on the use of other antipsychotics, including clozapine, paliperidone, asenapine and ziprasidone, for neuropsychiatric symptoms in dementia are lacking considering that these antipsychotics are rarely used for this indication, if at all.246
Overall, evidence suggests that the efficacy of typical and atypical antipsychotics for the treatment of neuropsychiatric symptoms in people with dementia is modest at best and is limited mainly to the treatment of agitation, aggression and psychosis for a duration of up to 12 weeks. In practice, these modest benefits should be weighed up against the risk of severe adverse effects associated with antipsychotics.
3.2.5 Safety concerns
The safety concerns with antipsychotics are wide-ranging and often outweigh the modest benefits derived from their use for neuropsychiatric symptoms in dementia. Antipsychotics are associated with a wide range of possible adverse effects, such as metabolic effects, cardiac conduction abnormalities, falls and movement disorders. Furthermore, there is an increased
43 | Page Daniel J. Hoyle risk of cerebrovascular events and death among older people with dementia. Patient-related factors (e.g., type of dementia, co-prescribed medications) can influence the presentation of these adverse effects. As discussed in Chapter 3.2.1, the adverse effect profile for individual antipsychotics are heavily determined by differences in receptor binding affinity (Table 6 and
Table 7).
A detailed discussion of all possible adverse effects is beyond the scope of this thesis.
Instead, this section is intended to provide the reader with a brief background on the most severe and prevalent issues associated with antipsychotic use.
3.2.5.1 All-cause mortality risk
In 2005, the US Food and Drug Administration (FDA) introduced a ‘black box’ warning for all atypical antipsychotics to advise of an increased risk of mortality in older patients with dementia.247 This warning was made after finding a mortality rate of 1.6 to 1.7-fold during re- analysis of 17 placebo-controlled trials (including results from several unpublished studies).247
Most of the deaths reported in this meta-analysis were cardiac-related (e.g., heart failure) or infections (mostly pneumonia).247 Within the same year, another meta-analysis, including 15 trials (nine unpublished), reported an increased risk of death with use of atypical antipsychotics for neuropsychiatric symptoms in people with dementia, with an odds ratio (OR) of 1.54
(95%CI=1.06, 2.23; p=0.02).248 The causes of death were not reported within this meta- analysis.248
In June 2008, the FDA extended this ‘black box’ warning to include typical antipsychotics based on two observational studies.249 Gill et al. retrospectively reviewed the outcomes of
27,259 people aged over 65 years with a diagnosis of dementia in Ontario, Canada.250 They found that those taking typical antipsychotics had a higher risk of mortality at 30, 60, 120 and
180 days compared to those taking atypical antipsychotics.250 Similarly, a retrospective review of mortality rates for people aged 65 years and over with newly-prescribed typical (n=12,882)
44 | Page Daniel J. Hoyle and atypical (n=24,359) antipsychotics for any indication found typical antipsychotics conferred a comparable and possibly greater risk of death than atypical antipsychotics.251 In particular, compared to risperidone, haloperidol was associated with a greater risk of mortality (mortality ratio=2.14, 95%CI=1.86, 2.45).251 Mortality was greatest among people taking doses higher than the median level and during the first 40 days after starting therapy.251 Neither study could determine the causes of death. However, several plausible mechanisms were suggested including sudden cardiac death and arrhythmias due to a prolonged QT interval, aspiration syndromes (e.g., aspiration pneumonia) and choking due to sedation and extrapyramidal symptoms, cerebrovascular events and mortality associated with fractures from falls.250, 251
Since the release of the ‘black box’ warnings, many studies have reported increased mortality with antipsychotic use.36, 252-255 In 2015, a retrospective case-control study of 90,786 people 65 years or older with dementia investigated the effect of antipsychotics, valproic acid and antidepressants on mortality over 180 days.36 The number needed to harm (NNH) was 26
(95%CI=15, 99) for haloperidol, 27 (95%CI=19, 46) for risperidone, 40 (95%CI=21, 312) for olanzapine, and 50 (95%CI=30, 150) for quetiapine.36 Conversely, a 2019 network meta-analysis of 17 clinical trials found that the risk of death was highest for olanzapine (OR=1.74, 95%CI=0.74,
4.07), followed by aripiprazole (OR=1.66, 95%CI=0.65, 4.25) and quetiapine (OR=1.64,
95%CI=0.74, 3.63). The lowest risk of mortality was with risperidone (OR=1.32, 95%CI=0.77,
2.27).256
In 2018, a systematic review and meta-analysis of 68 observational studies reported that exposure to antipsychotics is associated with ~1.5-fold increased risk of mortality, particularly among residents older than 65 years and those with Parkinson’s disease.252 Another
2018 meta-analysis of 20 studies found that the increased risk of mortality with antipsychotic use extended to older people (generally aged ≥65 years) without a diagnosis of dementia.253
45 | Page Daniel J. Hoyle Lastly, a 2009 report to the Minister of State for Care Services estimated that out of the
180,000 people with dementia treated with antipsychotics per year in the UK, 1,800 additional deaths in this population can be attributed directly to the use of antipsychotic medication.257
3.2.5.2 Cerebrovascular events
Cerebrovascular events, such as stroke, increase the risk of death. However, with improvements in acute stroke care, a greater proportion of patients are surviving the acute cerebrovascular episode and are suffering long-term consequences, such as physical disability, cognitive impairment, and psychological issues.258 In 2004, the Committee on Safety of Medicines advised of a three-fold increased risk of cerebrovascular events with the use of risperidone and olanzapine in older people with dementia compared to placebo.259
A 2019 network meta-analysis of 17 studies found that olanzapine (OR=4.28,
95%CI=1.26, 14.56) and risperidone (OR=3.85, 95%CI=1.55, 9.55) were associated with a greater risk of cerebrovascular events compared to placebo among people with dementia.256 On the other hand, the risk of cerebrovascular events with aripiprazole and quetiapine was not significantly different from placebo among people with dementia.256
A 2008 self-controlled case series (n=6,790 patients) reported an increased risk of stroke with typical (Rate Ratio (RR)=1.69, 95%CI=1.55, 1.84) and atypical antipsychotics
(RR=2.32, 95%CI=1.73, 3.10).260 Furthermore, the use of any antipsychotic drug in people with dementia was associated with a greater risk of stroke (RR=3.50, 95%CI=2.97, 4.12) compared to people without dementia (RR=1.41, 95%CI=1.29, 1.55).260
A 2006 meta-analysis of five placebo-controlled trials found that patients treated with risperidone for neuropsychiatric symptoms associated with dementia were significantly more likely to have a cerebrovascular event (OR=3.64, 95%CI=1.72, 7.69, p=0.0007).22
46 | Page Daniel J. Hoyle The biological mechanisms responsible for an increased risk of ischaemic cerebrovascular events with antipsychotic use are not entirely understood.261, 262 However, it has been proposed that antipsychotic-induced hyperprolactinaemia, which promotes platelet aggregation,263 may explain the increased risk of cerebrovascular events.261 There are also suggestions that antipsychotic-induced sedation may result in venostasis and clot formation.264
Lastly, possible insulin resistance and weight gain associated with antipsychotic use may increase the risk of a cerebrovascular event.265
3.2.5.3 Falls and fractures
A fall is defined by the World Health Organization as “an event which results in a person coming to rest inadvertently on the ground or floor or other lower level”.266 Between 30 and 70% of
RACF residents fall at least once a year.267 Falls may be associated with potentially serious consequences, including fracture, head trauma, soft-tissue injuries and severe lacerations.268 In developed countries, falls account for 85% of injury-related hospital admissions in people over
65 years.269 Hip fractures account for the majority of injuries.270 Over 70% of people report mobility issues after 120 days following a hip fracture.271 Furthermore, falls increase the risk of mortality.272 In a prospective study of 300 Norwegian women, the relative risk of death among people experiencing at least two falls during one year was 1.6 (95%CI=1.1, 2.4) compared to those who did not have a fall.272 Falls may also lead to loss of independence, fear of falling and reduced QoL.559, 560 Whilst risk factors for falls are multifactorial (e.g., gait and balance instability, visual impairment, cognitive impairment),268 use of psychotropic medication is often targeted as a modifiable risk factor.268
Antipsychotics are thought to contribute to an increased risk of falls and fractures due to the increased risk of extrapyramidal side effects (e.g., tremor, rigidity, bradykinesia), reduced bone mineral density related to elevated prolactin levels, sedation, balance issues, vision changes and orthostatic hypotension.273-277
47 | Page Daniel J. Hoyle Several systematic reviews and meta-analyses have demonstrated an increased risk of falls with the use of antipsychotics in older populations.195, 278-281 The latest of these reviews performed a meta-analysis using 248 studies to investigate psychotropics as a risk factor for falls in participants ≥60 years or in studies where participants had a mean age ≥70 years.195
Using adjusted data, antipsychotic use was associated with an increased risk of falls (pooled
OR=1.54, 95%CI=1.28, 1.85).195 Additionally, a retrospective observational study of 2,368 residents from nine Dutch RACFs found that both regular and when required antipsychotic use was associated with an increased incidence of falls (OR=1.97, 95%CI=1.51, 2.59).37
3.2.5.4 Pneumonia
As mentioned above, the US FDA’s investigation found that most deaths associated with antipsychotic use in older people with dementia were related to cardiovascular and infectious causes, including pneumonia.247 Since then, systematic reviews and meta-analyses of observational studies have provided evidence that both typical and atypical antipsychotics are associated with an increased risk of pneumonia.282-284
In 2018, Dzahini et al. performed a meta-analysis on 14 observational studies
(n=206,899 patients), with nine studies restricted to patients over the age of 65 years.282 The risk of pneumonia was increased by the use of atypical antipsychotics (risk ratio=1.93,
95%CI=1.55, 2.41) and typical antipsychotics (risk ratio=1.07, 95%CI=1.34, 2.15). There was no significant difference in the risk of pneumonia between typical and atypical antipsychotics (risk ratio=1.07, 95%CI=0.85, 1.35).282 Three studies reported an increased risk of pneumonia with higher antipsychotic doses,285-287 although in one study this effect was only observed with clozapine.287
Several biological mechanisms have been proposed to explain antipsychotic-associated pneumonia.282, 288 In particular, antipsychotics may increase the risk of aspiration pneumonia.282
Blockade of the dopaminergic pathway by some antipsychotics (e.g., haloperidol) has been
48 | Page Daniel J. Hoyle linked to decreased levels of substance P,289, 290 a key neurotransmitter involved in the initiation of the swallowing and cough reflexes. Similarly, extrapyramidal side effects, such as tardive dyskinesia, resulting from dopamine receptor-2 blockade can affect the respiratory muscles and increase the risk of aspiration.291 Sedation, resulting from histamine-1 receptor blockade by some antipsychotics (e.g., olanzapine) may impair the laryngeal cough reflex.282 Sialorrhoea, a common adverse effect of clozapine that tends to be more prevalent at night,292, 293 can result in inhalation of saliva. Alternatively, antipsychotics, such as olanzapine and clozapine, can impair pharyngeal and oesophageal peristalsis leading to excessive saliva pooling.282, 288
Other explanations for antipsychotic-induced pneumonia include xerostomia leading to an increase in oral bacteria,288 impairment of the immune system (e.g., agranulocytosis with clozapine), and the effects antipsychotics have on the thromboxane A2 receptor and platelet activation factor receptor downstream.288
3.2.6 Summary
Antipsychotics are widely utilised in the RACF setting often to manage neuropsychiatric symptoms in residents with dementia. Clinical guidelines recommend reserving these medications for the symptoms of agitation, aggression and psychosis when non- pharmacological therapies are ineffective and/or if there is a risk of harm to the resident or others. If used for this indication, antipsychotics should be reviewed at least every 12 weeks and the dose reduced or ceased, if possible, considering only limited efficacy for this indication and the risk of severe adverse effects, such as falls and fractures, cerebrovascular events and an increased risk of death.
49 | Page Daniel J. Hoyle 3.3 Benzodiazepines
3.3.1 Mechanism of action
Benzodiazepines act by enhancing the affinity of gamma-aminobutyric acid (GABA) by allosteric modification of the GABAA receptor (Figure 12). When stimulated, the GABAA receptor mediates fast inhibitory synaptic transmission throughout the CNS. Benzodiazepines enhance the response to GABA by facilitating the opening of GABA-activated chloride channels.294 This increases the influx of chloride ions and leads to neuroinhibitory (sedative, anxiolytic and antiepileptic) effects.295
Z-drugs (e.g., zolpidem, zopiclone) have a similar mechanism of action. However, these medications are rarely used in Australian RACFs as they are not covered on the PBS. Only 26
(0.23%) out of 11,368 residents involved in RedUSe were taking a z-drug.10 Consequently, these benzodiazepine-related medications were not included in this study.
Figure 12: Mechanism of benzodiazepines; (A) represents an inactive GABAA receptor with the coupled chloride channel being closed. (B) represents GABA binding to the GABAA receptor which causes the chloride channel to open and leads to hyperpolarisation of the cell. (C) represents the binding of GABA and a benzodiazepine (BZ) to the GABAA receptor. This leads to the greater entry of the chloride ion and increased hyperpolarisation. This makes it more difficult for a cell to depolarise. Thus, there is a subsequent neuroinhibitory response. GABA=γ-aminobutyric acid.
50 | Page Daniel J. Hoyle The benzodiazepines available in Australia are listed in Table 8.
Table 8: Benzodiazepines available in Australia according to the length of action. Adapted from the Australian Medicines Handbook, 2018.296
Length of action Drugs Very short Midazolam Short Alprazolama, oxazepama, temazepamb Medium Bromazepama, lorazepama Long Clobazama, clonazepam, diazepama,c, flunitrazepamb, nitrazepamb aIndicated for anxiety, bIndicated for insomnia, cIndicated for anxiety 3.3.2 Use in residential aged care facilities
Benzodiazepines are commonly prescribed for insomnia and/or anxiety, frequently at doses at the upper limit of the recommended doses in older people.297 They are also used off-label, as an alternative or in addition to antipsychotics, for the management of neuropsychiatric symptoms, such as acute agitation.31, 198, 297-299 Similar to antipsychotics, benzodiazepines have limited evidence of long-term efficacy for these symptoms and may be associated with severe adverse effects (Chapter 3.3.3). Considering that the guidance for their use differs depending on the condition being treated, the place of benzodiazepines in the management of each of these conditions is discussed separately below.
With limited evidence of efficacy in the long-term for insomnia (Chapter 3.3.2.1.1) and risk of severe adverse effects (Chapter 3.3.3), clinical guidelines for the management of insomnia generally recommend reserving benzodiazepines to scenarios where non- pharmacological therapies are ineffective. Furthermore, therapy duration should be limited to two to four weeks.29, 32, 300, 301 Non-pharmacological therapies utilised for insomnia include cognitive therapy, stimulus control, relaxation techniques, sleep hygiene education and sleep restriction.29, 32, 300-302
Similarly, clinical guidelines recommend psychological interventions (e.g., education on relaxation and coping skills, cognitive behavioural therapy) as a first-line treatment for generalised anxiety disorder, panic disorder, and social anxiety disorder, considering their
51 | Page Daniel J. Hoyle ability to reduce symptoms of anxiety in the short- and long-term.30, 32, 303, 304 Whilst benzodiazepines have well established anxiolytic effects (Chapter 3.3.2.2.1), concern regarding their adverse effects and risk of dependence limit their use (Chapter 3.3.3). Generally, benzodiazepines should be restricted to those with severe or treatment-resistant cases of generalised anxiety disorder, social anxiety disorder and panic disorder or as short-term adjunctive therapy to minimise an exacerbation in symptoms in the first few days to weeks after initiating antidepressant therapy.32, 303, 304 Antidepressants, particularly selective serotonin reuptake inhibitors, venlafaxine, tricyclic antidepressants and selective monoamine oxidase inhibitors, are considered appropriate for long-term pharmacological management of anxiety.30,
303, 304 When benzodiazepines are used for anxiety they should be used for up to two weeks and the dose gradually reduced over six weeks.30 If ongoing use is required, it should be provided on an ‘as required’ basis.30 Long-term use of benzodiazepines for anxiety should only be pursued, under specialist advice, if both psychological therapies and other pharmacological options have not provided significant improvement.30, 303
Alongside antipsychotics, benzodiazepines are commonly used in the management of behavioural symptoms associated with dementia.28, 31, 198, 297, 299 Although not licensed for behavioural symptoms in people with dementia, clinical guidelines suggest benzodiazepine therapy can be used intermittently when required for severe agitation or anxiety that cannot be managed effectively with non-pharmacological therapies.28, 3132, 305 Recent evidence of a reduction in risperidone use for neuropsychiatric symptoms in people with dementia in
Australia306 has been marred by concern that psychotropics, such as benzodiazepines, are being used as a substitute.307 There is, however, only limited evidence of efficacy for this indication,198,
298 particularly when used long-term. Furthermore, benzodiazepine use in older people is associated with significant adverse effects, such as falls and fractures, cognitive impairment and paradoxical worsening in agitation.142, 195, 308, 309 Regular use (e.g., daily) of benzodiazepines for the treatment of neuropsychiatric symptoms is not endorsed by guidelines, especially in people
52 | Page Daniel J. Hoyle with dementia considering their cognitive impairing effects. However, one systematic review has suggested that benzodiazepines may be used in circumstances where non-pharmacological therapies are ineffective and other psychotropic medications are considered inappropriate due to medication allergies or tolerability issues.198
3.3.2.1 Insomnia
Insomnia is a complaint of dissatisfaction with sleep quantity or quality, associated with difficulty initiating sleep (sleep-onset insomnia), difficulty maintaining sleep (characterised by frequent awakenings or problems returning to sleep after awakenings; sleep maintenance insomnia), and/or early-morning awakening with an inability to return to sleep (late insomnia).310 Difficulties maintaining sleep is the most common insomnia symptom, followed by difficulty falling asleep. A combination of these symptoms is the most common presentation overall.311-313
In primary care, insomnia is prevalent in 10-20% of individuals.311, 312 However, within the RACF setting, the prevalence of insomnia symptoms is reported to be between 17-65% depending on the definition of insomnia, the study design, and study population.314-317 The high prevalence of insomnia in RACFs is likely attributable to environmental factors (e.g., noise, lighting, temperature, nursing times) and resident characteristics (discussed below).26 A cross- sectional study in 76 Belgian RACFs found that the most common indication for chronic benzodiazepine use was insomnia (59% of benzodiazepine users).297
Insomnia can be situational, persistent, or recurrent.310 Situational insomnia is usually transient, lasting a few days or weeks and is associated with events or rapid changes in sleep schedules or environment (e.g., transferring to a RACF).310 Situational insomnia usually resolves once the precipitating event subsides. However, for individuals more vulnerable to sleep disturbances, insomnia may persist long after the triggering event, possibly due to heightened arousal.310 For example, a person who is bedridden due to injury and has difficulty sleeping may
53 | Page Daniel J. Hoyle develop negative associations for sleep leading to conditioned arousal and persistent insomnia.
Studies have found chronic insomnia to develop in 40 to 75% patients that report symptoms at baseline.318, 319
Insomnia symptoms are more prevalent among older adults, particularly older females.320 Further, the types of insomnia symptoms change with age. Whilst older people are more likely to report difficulties maintaining sleep, younger patients are more likely to report problems initiating sleep.311-313 The increased incidence of insomnia in older adults can be partly explained by the higher incidence of physical health problems with ageing (e.g., pain). Increasing age is also associated with decreased rapid eye movement sleep time, increased amount of time spent in light sleep phase, and shorter, more fragmented, more easily disturbed sleep, and an altered circadian rhythm.321-323
Cognitive impairment and dementia have also been linked to sleep disorders. A review of 18 studies reported that 14 to 59% of people with mild cognitive impairment exhibited sleep disturbances.324 Furthermore, disrupted circadian rhythm (e.g., fragmented nocturnal sleep, excessive daytime sleepiness, sundowning) are common symptoms among people with
Alzheimer’s disease. People with Lewy body dementia also commonly present with rapid eye movement sleep behaviour disorder, disrupted night sleep, and hypersomnia.325, 326
Sleep disturbances in RACF residents are associated with a range of consequences. For example, a study of 49 residents from four Australian RACFs reported an increase in verbal and physically nonaggressive agitation with reduced sleep.327 Furthermore, night-time behaviours have been linked to high amounts of distress for care staff, particularly among residents with young-onset dementia.144, 234 Lastly, insomnia has been linked to increased health service costs,328 daytime fatigue329 and mood disturbances,330 and reduced cognitive function,331 functional status330 and QoL.330, 332
54 | Page Daniel J. Hoyle 3.3.2.1.1 Effectiveness of benzodiazepines for insomnia
Evidence from systematic reviews and meta-analyses have suggested that benzodiazepine use may result in small but statistically significant improvements in sleep.35, 333, 334 However, these benefits are often outweighed by the risk of adverse effects (discussed further in Chapter
3.3.3).35, 333, 334
In 2005, Glass et al. performed a systematic review and meta-analysis of 24 RCTs of hypnotic medications for insomnia in older people (mean age >60 years).35 Overall, benzodiazepine use was associated with a significant improvement in sleep quality (mean effect size=0.37, 95%CI=0.01, 0.73), amount of sleep (mean=34.2 minutes, 95%CI=16.2, 52.8 minutes) and a decrease in the mean number of awakenings by 0.60 (95%CI=-0.41, -0.78) compared to placebo.35 The researchers calculated that sleep is improved in one person for every 13
(95%CI=6.7, 62.9) treated with a hypnotic.35 However, adverse effects were twice as common, with one in every six people treated with a hypnotic exhibiting adverse effects, such as drowsiness or fatigue, headache, nightmares, and nausea or gastrointestinal disturbances.35
In 2007, Buscemi et al. performed a systematic review and meta-analysis of 105 randomised double-blind, placebo-controlled trials on the pharmacological treatments for chronic insomnia in adults.333 The researchers reported that benzodiazepine use resulted in a reduction in sleep latency by an average of 10 minutes (95%CI=-16.6, -3.4 minutes) when measured with polysomnography. The improvements were more prominent when measured with a sleep diary. Sleep quality and sleep efficiency were greater with benzodiazepine use compared to placebo. However, the differences in wakefulness after sleep and total sleep time were not significantly different. The researchers suggested that the results are more applicable to short-term treatment of chronic insomnia because few studies have evaluated long-term use.333
55 | Page Daniel J. Hoyle The translation of results from short-term studies to long-term use is fraught due to the rapid development of tolerance to benzodiazepines. Evidence suggests that the efficacy for insomnia can diminish in less than four weeks, particularly among shorter-acting benzodiazepines.335 However, adverse effects tend to persist.335 Supporting this, a study within
10 Belgian RACFs found no difference in self-reported time to fall asleep or hours asleep in long- term users (>3 months) of benzodiazepines (n=178) compared to nonusers (n=122).336 In this study, long-term benzodiazepine users reported more next-day sedation, difficulties falling asleep, and midnight awakenings, and lower sleep quality than non-users.336 Whilst there is evidence to suggest changes in sleep architecture with chronic benzodiazepine use leads to a reduction of time in deep restorative stages of sleep,337 the perception of difficulties falling asleep among residents using long-term benzodiazepines may be explained by pseudoinsomnia; a misperception where sleep latency is over-estimated and duration of sleep is under- estimated.338
3.3.2.2 Anxiety
“Anxiety disorders include disorders that share features of excessive fear and anxiety and related behavioural disturbances. Fear is the emotional response to a real or perceived imminent threat, whereas anxiety is the anticipation of future threat”.339 Whilst fear and anxiety overlap, fear is more often associated with the fight or flight response and anxiety is more often associated with muscle tension and vigilance in preparation for future danger and cautious or avoidant behaviours.339
Anxiety disorders, especially generalised anxiety disorder and specific phobias, are common psychological issues among older people, particularly within the RACF setting.340-342
The overall prevalence of anxiety disorders among community-dwelling older people is estimated to be between 1.4 and 17.1%.341, 343 In comparison, a 2016 systematic review of 18 studies (n=5,927 participants) found that the overall prevalence of anxiety in RACFs ranged from
56 | Page Daniel J. Hoyle 3.2 to 20%.340 In 2016, a systematic review by Creighton et al. found that the most common anxiety disorder amongst residents of RACFs was generalised anxiety disorder.340 The prevalence of generalised anxiety disorder across four studies was 0.9 to 11.2%.340
More recently, Creighton et al. investigated the prevalence and treatment of anxiety in residents (n=180) from 12 Australian RACFs using validated measures of anxiety.342 They found that 19.4% (n=35) of their sample met the threshold for at least one anxiety disorder. However, only 40% of these residents (n=14) had a documented diagnosis of anxiety on their medical file.
Sixty per cent of residents with at least one anxiety disorder were receiving a benzodiazepine yet only 8.6% were receiving psychological treatment.342
As discussed earlier in this chapter, benzodiazepines are commonly used to manage anxiety. A 2012 cross-sectional study in 76 Belgian RACFs found that anxiety was the second most common indication for chronic benzodiazepine use (17% of users), after insomnia (59% of users).297
3.3.2.2.1 Effectiveness of benzodiazepines for anxiety
Benzodiazepines exert an anxiolytic effect by decreasing vigilance and somatic symptoms, such as muscle tension.344 Studies have reported prolonged efficacy when used for generalised anxiety disorder and panic disorder345, 346; unlike the sedative, cognitive impairing and amnesic effects of benzodiazepines, development of tolerance to the anxiolytic effects of benzodiazepines appear to develop much slower if at all.335 Alternative treatments, such as antidepressants, are generally preferred for long-term treatment of anxiety disorders.30 Whilst benzodiazepines provide a more rapid anxiolytic effect, they are also associated with a greater risk of harm (e.g., falls, cognitive impairment, dependence) in older people.347, 348 Additionally, antidepressants can be used to address comorbid depression which occurs in up to 90% of patients with anxiety.349, 350
57 | Page Daniel J. Hoyle A 2019 systematic review and network meta-analysis, published in the Lancet, pooled the results of participants involved in RCTs of pharmacological treatments for generalised anxiety disorder (n=25,441 participants across 89 studies).351 The network meta-analysis found benzodiazepines and other treatments (including buspirone, citalopram, duloxetine, sertraline, venlafaxine) had similar efficacy in the treatment of generalised anxiety disorder. However, unlike most other drugs, participants were more likely to drop out of the study compared to participants given placebo.351
Another 2019 network meta-analysis of 91 RCTs (n=14,812 participants) investigated the effectiveness of pharmacological and psychological interventions for generalised anxiety disorder.352 Whilst benzodiazepines were found to be more effective than placebo (effect size=-
0.40, 95%CI=-0.65, -0.15), other pharmacological therapies (e.g., bupropion, mirtazapine, selective serotonin reuptake inhibitors, serotonin and noradrenaline reuptake inhibitors) were relatively more efficacious.352 However, compared to selective serotonin reuptake inhibitors and serotonin and noradrenaline reuptake inhibitors, benzodiazepines had a faster onset of action.352
Supporting the long-term use of antidepressants in preference to benzodiazepines for anxiety, a randomised, placebo-controlled trial assessed the effectiveness of imipramine, trazodone and diazepam over an eight-week treatment (n=230 patients).353 Whilst patients treated with diazepam showed the most improvement in anxiety during the first two weeks, particularly in somatic symptoms, imipramine was more efficacious and trazodone had comparable effects on anxiety from week three onwards.353
Considering their fast onset of action, benzodiazepines may be beneficial as an adjunctive therapy during the first four weeks of antidepressant therapy given the slow onset of effect and increased risk of initially worsened symptoms with antidepressant medication.354-
356 However, Bower et al. argue that the augmentation of antidepressant therapy using
58 | Page Daniel J. Hoyle benzodiazepines should be avoided considering the poor safety profile and insufficient evidence to guide augmentation pharmacotherapy in older patients with anxiety.357
3.3.2.3 Behavioural symptoms
Benzodiazepines are commonly prescribed acutely to manage agitated and aggressive behaviour.31, 198, 297, 299 Whilst many guidelines for the management of neuropsychiatric symptoms in people with dementia do not mention benzodiazepine use, there are suggestions by the International Psychogeriatric Association, the Australian Psychotropic Expert Group, and the Royal Australian College of General Practitioners that short-acting benzodiazepines, such as oxazepam or lorazepam, can be used for short durations (up to two weeks) at low doses in an
‘as needed’ capacity for severe anxiety and agitation when behaviours escalate.28, 31, 221
3.3.2.3.1 Effectiveness of benzodiazepines for behavioural symptoms
Evidence for the efficacy of benzodiazepines in the management of neuropsychiatric symptoms in dementia is limited. In 2017, a systematic review by Tampi et al. found no meta-analyses reporting on the effect that benzodiazepines have in the management of this indication.298
However, two systematic reviews have reported a handful of studies related to this query.27, 198
In 2015, Defrancesco et al. performed a systematic review of prospective and retrospective studies to investigate the frequency of benzodiazepine use, and their effects on cognitive function and neuropsychiatric symptoms associated with dementia.27 Overall, 18 studies met the inclusion criteria with five studies reporting on the effect that benzodiazepines have on neuropsychiatric symptoms.27 The results of these five studies were inconsistent. One study found increased night-time wandering and daytime restlessness with brotizolam use.358
Another found no significant effects of triazolam on total sleep time, latency to sleep onset, number of arousals or time asleep during the day compared to placebo.359 Two studies reported small but non-significant improvements in agitation with oxazepam and lorazepam use.360, 361
The only randomised, placebo-controlled trial found significant improvements in agitation
59 | Page Daniel J. Hoyle following intramuscular administration of lorazepam.362 Overall, Defrancesco et al. concluded that it is not possible to provide clear recommendations for the evidence-based use of benzodiazepines in people with dementia.27
In 2014, Tampi et al. reviewed the evidence of efficacy and tolerability of benzodiazepines for the management of neuropsychiatric symptoms in dementia within
RCTs.198 Five studies were included in this review; one trial compared diazepam to thioridazine,363 one trial compared oxazepam to haloperidol and diphenhydramine,361 one trial compared alprazolam to lorazepam,360 one trial compared alprazolam to haloperidol,364 and one trial compared intramuscular lorazepam to olanzapine and placebo.362 Four of these studies had small sample sizes (40 to 68 participants).360, 361, 363, 364 However, the study by Meehan et al. included 272 participants.362 Overall, four of these studies reported no difference in efficacy between the active drug groups to treat the neuropsychiatric symptoms in people with dementia.360-362, 364 Though the study by Covington et al. suggested that thioridazine was superior to diazepam for affect and insomnia.363 Two studies reported improvements in agitated symptoms with the use of pharmacological agents compared to placebo.361, 362 The researchers of this review highlighted that the trials were short, lasting between 24 hours and eight weeks.198 Overall, the researchers highlight that whilst there is only limited data to support the use of benzodiazepines in the treatment of neuropsychiatric symptoms in people with dementia, these medications may be appropriate for severe agitation or anxiety in situations where other psychotropic medications are unsafe (e.g., in patients severe Parkinson’s disease,
Dementia with Lewy bodies, severe cardiac arrhythmias) or where there are allergies or tolerability issues to other medications.198
Considering there is little to no evidence that benzodiazepines are effective for the treatment of neuropsychiatric symptoms and there is a high risk of severe adverse effects associated with their use (see Chapter 3.3.3), benzodiazepines should generally be avoided for
60 | Page Daniel J. Hoyle the treatment of neuropsychiatric symptoms except for cases of severe agitation or anxiety when other medications cannot be used.
3.3.3 Safety concerns
The Beer’s Criteria, a list of potentially inappropriate medication use in older adults updated by the American Geriatrics Society every three years, strongly recommends the avoidance of all benzodiazepines in older adults based on moderate-quality evidence.348 This recommendation is made on the basis that older adults have increased sensitivity and decreased ability to clear benzodiazepines, which may predispose the patient to an increased risk of adverse effects.218,
279297 In older people, hepatic clearance of some benzodiazepines (e.g., diazepam) may be reduced due to impaired hepatic blood flow and atrophy of the liver.365, 366 Additionally, a relative increase in total body fat and subsequently increased volume of distribution of benzodiazepines partially accounts for the prolonged effects of benzodiazepines in older people.295
In addition to the general warning against the use of benzodiazepines in older people, the Beer’s Criteria strongly recommends the avoidance of benzodiazepines in older people with: delirium or at high risk of delirium because of potential induction or worsening of delirium; dementia or cognitive impairment due to adverse CNS effects; and a history of falls or fractures due to adverse CNS effects.348
In 2005, a meta-analysis of 24 RCTs (n=2,417 participants) investigated the risks and benefits of pharmacological therapies for the treatment of insomnia in people aged 60 years and over.35 On the basis of adverse events reported in 16 studies (n=2,220 participants), the
NNH for sedative-hypnotics compared to placebo was 6 (4.7 to 7.1) suggesting that one person in every six treated with a hypnotic sedative will suffer an adverse effect.35 Considering the number needed to treat for improvement in sleep with hypnotic sedatives was 13,35 it is evident that an older person who is treated with a hypnotic sedative are more likely to suffer an adverse
61 | Page Daniel J. Hoyle effect than have their sleep improved.35 The meta-analysis found that the most common adverse events were drowsiness or fatigue, headache, nightmares and gastrointestinal disturbances.35 There were also notable cognitive adverse effects and psychomotor side effects
(e.g., dizziness or loss of balance), particularly with higher doses.35
Whilst it is beyond the scope of this thesis to discuss all possible adverse effects associated with benzodiazepine use, information on some of the more severe adverse effects
(frailty, pneumonia, cognitive impairment, dementia, falls and fractures, and mortality) is provided below.
3.3.3.1 Frailty
Use of sedative agents, including benzodiazepines, have been linked to higher levels of frailty.367-369 Frailty is a “state of increased vulnerability to poor resolution of homeostasis following stress, which increases the risk of adverse outcomes”.370 In 2001, Friedman et al. defined frailty as a clinical syndrome when three or more of the following criteria were present: unintentional weight loss; self-reported exhaustion; weakness (grip strength); slow walking speed; and, low physical activity.371 People with frailty are at an increased risk of falls, hospitalisations, difficulties with mobility, disability, cognitive decline, and death.370, 372
In 2015, a cross-sectional study investigated the association between sedative use and frailty among 3,446 people aged ≥65 years in Ireland.367 Overall, 19.4% (n=668) of the sample took a sedative medication. Specifically, 6.5% (n=223) and 2.6% (n=88) took hypnotic and anxiolytic medications, respectively. Among 1,718 participants who had their frailty status assessed, frail (adjusted OR=1.30, 95%CI=1.02, 1.64; p=0.023; n=72) and pre-frail participants
(i.e., subclinical frailty; adjusted OR=1.27, 95%CI=1.11, 1.46; p<0.001; n=672) were more likely to use sedatives than nonfrail participants (n=974) after adjusting for relevant confounders.367
Furthermore, multiple linear regression models to control for covariates, confirmed that higher sedative loads were associated with higher frailty index scores (adjusted β=1.77, 95%CI=1.13,
62 | Page Daniel J. Hoyle 2.42; p≤0.001).367 Considering the cross-sectional study design, it was not possible to assess the causality of this relationship. Nonetheless, the high prevalence of sedative use among frail older people is concerning considering that both sedative use and frailty pre-dispose older people to adverse health outcomes, such as falls and fractures.367
In 2019, a prospective, population-based, cohort study investigated the relationship between potentially inappropriate medications and frailty among 2,865 community-dwelling people aged ≥60 years in Germany between October 2008 and September 2016.368 At baseline,
261 participants were identified as frail (“prevalent frailty”) and a further 423 became frail
(“incident frailty”) during follow-ups (average follow-up= 4.8 years). The identification and grouping of potentially inappropriate medications were based on several sources.373-375 For example, the BEERS criteria “to avoid in patients with dementia or cognitive impairment” sub- list (“BEERS dementia sub-list”), which includes anticholinergics, benzodiazepines, H2-receptor antagonists, z-drugs and antipsychotics, was included as one group of potentially inappropriate medications. Only the BEERS dementia sub-list, had a statistically significant association with prevalent frailty (adjusted OR=1.51, 95%CI=1.04, 2.17).368 However, given the small sample size, re-analysis on the benzodiazepine class alone (n=91 users) found no significant relationship with prevalent frailty after adjustment for confounders. Additionally, none of the potentially inappropriate medication groups were significantly associated with incident frailty, indicating the effect of these medications may be acute.368 Larger studies involving new users of potentially inappropriate medications are required to confirm this effect.
Overall, it remains unclear whether sedatives, such as benzodiazepines, influence frailty in older people or if frail older people have underlying conditions that increase the risk of sedative prescription. Whilst it is biologically plausible that benzodiazepines may contribute to frailty through mechanisms such as muscle relaxation, which can lead to decreased muscle strength, gait speed, physical activity and increase the feeling of exhaustion, evidence of a
63 | Page Daniel J. Hoyle causal relationship between benzodiazepine use and frailty has not yet been elucidated.
Nonetheless, it is advisable that prescribers perform frailty assessments before and after prescribing sedatives, and consider ceasing this medication if there are signs of deterioration.368
3.3.3.2 Pneumonia
There have been concerns that benzodiazepine use is associated with an increased risk of pneumonia.376 It has been hypothesised that benzodiazepine use may suppress the immune system or relax the oesophageal sphincter leading pulmonary aspiration, thus contributing to pneumonia.377, 378 However, the exact mechanism is unknown. Whilst several studies have suggested a possible association between benzodiazepine use and pneumonia risk,379-383 others have not.384, 385
A 2019 systematic review and meta-analysis of six case-control studies and four-cohort studies (n>120,000 pneumonia cases) found benzodiazepine use to be significantly associated with an increased risk of pneumonia (pooled OR=1.25, 95%CI=1.09, 1.44; p<0.001).376 Further analysis found that the risk for pneumonia was highest in people who had received a prescription for a benzodiazepine in the previous 30 days (pooled OR=1.4, 95%CI=1.22, 1.6) or received their prescription in previous the 31 to 90 days (pooled OR=1.38, 95%CI=1.06, 1.8) compared to those who received their prescription more than 90 days ago (pooled OR=1.11,
95%CI=0.96, 1.27).376 However, it should be noted that this sub-analysis was based on very few studies and may not be reliable. Another sub-analysis found that people aged ≥65 years old had a lower risk of pneumonia (pooled OR=1.29, 95%CI=1.15, 1.45) compared to those less than 65 years old (pooled OR=1.8, 95%CI=1.39, 2.34).376 Again, the researchers questioned the reliability of this result, suggesting that the sample size is insufficient to provide a reliable interpretation of the relationship between age and pneumonia risk with benzodiazepine use.376
There is concern that the risk of pneumonia may be even higher than reported in this review.376 Studies included in this review did not specifically report on the associations between
64 | Page Daniel J. Hoyle pneumonia and benzodiazepine use in RACF residents despite evidence that RACF residents, particularly those with advanced dementia, have a higher risk of pneumonia.386-388 Considering that a substantial proportion of RACF residents have dementia7 and utilise benzodiazepines at a greater rate than their community-dwelling counterparts,26, 297, 389-393 further studies are needed to investigate the risk of pneumonia in this vulnerable group.
Overall, there is some evidence that benzodiazepine use may increase the risk of pneumonia. However, more research is required to confirm this relationship.
3.3.3.3 Cognitive decline and dementia
Benzodiazepine use has been linked to cognitive deficits and, arguably, the development of dementia. It has been well documented that benzodiazepines contribute to acute cognitive and memory deficits.308, 394-397 However, the long-term effects remain largely unknown. Whilst there is literature suggesting that there are no long-term effects on cognitive decline,396, 398, 399 recent literature proposes an association between benzodiazepine use and the development of
Alzheimer’s disease and other dementias.400-404
A case-control study investigated the association between benzodiazepine use and cognitive decline in 3,777 community-dwelling people 65 years and older over an eight-year period.405 Overall, 150 people developed dementia and were compared with matched controls.
Use of benzodiazepines at any time was associated with a significantly increased risk of dementia (adjusted OR=1.7, 95%CI=1.2, 2.4).405 Furthermore, participants who had their benzodiazepine ceased in the previous two or three years prior to a diagnosis of dementia were also at higher risk of dementia (adjusted OR=2.3, 95%CI=1.2, 4.5).405 However, there was no association between the development of dementia and the use of benzodiazepines at the time of diagnosis.405 Whilst there are suggestions that benzodiazepine use may have been started to treat prodromal symptoms of dementia, the researchers indicated that if that were the case it would be expected that benzodiazepine use would be greater at the diagnosis of dementia.405
65 | Page Daniel J. Hoyle However, there are arguments that the duration of follow-up was insufficient in this study to rule out that benzodiazepines were being used to treat prodromal symptoms of dementia.402
More recently, Billioti de Gage et al. performed a large case-control study in Quebec,
Canada, to investigate the association between Alzheimer’s disease and benzodiazepine use initiated at least five years before diagnosis (n=1,796 cases and 7,184 controls matched on sex, age group and duration of follow-up).402 The use of benzodiazepines at any time point was associated with an increased risk of Alzheimer’s disease (adjusted OR=1.51, 95%CI=1.36, 1.69).
However, no association with Alzheimer’s disease was found in scenarios where the cumulative exposure to benzodiazepines was less than three months.402 Instead, the risk of Alzheimer’s disease was increased with greater exposure to benzodiazepines and with benzodiazepines with longer half-lives (≥20 hours).402 Diagnoses of anxiety, depression and insomnia were further controlled for in multivariable regression models but were not found to substantially affect the results, suggesting that it is unlikely benzodiazepines were being utilised for prodromal symptoms of Alzheimer’s disease.402 Furthermore, by only including people who had their benzodiazepine started at least five years prior to a diagnosis of Alzheimer’s disease, it is less likely that the benzodiazepine was started to treat prodromal symptoms of dementia.
Reviews of the literature308, 401 and meta-analyses406, 407 have suggested that whilst some smaller, short-term studies have not found evidence of increased dementia risk with long- term benzodiazepine use,308 the majority of studies suggest that benzodiazepine use is associated with an increased risk of dementia. In particular, a meta-analysis of six studies found that any use of benzodiazepines was associated with a 49% increased risk of dementia compared to non-users.406 Furthermore, the risk of dementia increased by 22% for every 20 defined daily doses per year.406
In contrast, three recent studies have found no evidence of increased dementia risk with benzodiazepine use.408-410 In 2019, a UK study found no relationship between the incidence
66 | Page Daniel J. Hoyle of dementia and use of benzodiazepines or anticholinergic agents among 3,045 people aged
≥65 years.410 Over 10 years, 220 of the 3,045 participants (7%) developed dementia. After adjusting for potential confounders, the incidence rate ratio for any benzodiazepine use was
1.06 (95%CI= 0.72, 1.06), the incidence rate ratio for new benzodiazepine use (started after baseline) was 0.65 (95%CI= 0.27, 1.60), and the incidence rate ratio for recurrent benzodiazepine use (taking at baseline and after two years) was 1.30 (95%CI 0.79, 2.14).410
Whilst there was no evidence that benzodiazepines were associated with the incidence of dementia, the number of benzodiazepine users was relatively small (n=227 participants used benzodiazepines at least once during the study), limiting statistical power.410 Furthermore, the type of benzodiazepine used (short- versus long-acting), the dose used, or the exact duration of use could not be ascertained from this study, despite evidence of their influence in the relationship between benzodiazepine use and dementia.402, 406
In 2019, a case-control study explored the relationship between benzodiazepine use and dementia.409 A total of 40,770 people with dementia diagnosed between April 2006 and
July 2015 were matched to 283,933 control participants on the basis of age, sex, available data history and deprivation, using the UK’s Clinical Practice Research Datalink. Benzodiazepine use was ascertained four to 20 years prior to the diagnosis of dementia. Whilst models that were unadjusted for potential confounders found a positive relationship between benzodiazepine use and dementia (OR=1.09, 95%CI=1.06, 1.12), models that were adjusted for potential confounders (e.g., number of recent physician consultations, anxiety, insomnia, depression, antidepressant prescriptions) at the end of the drug-exposure period indicated a negative relationship between benzodiazepine use and dementia (adjusted OR=0.89, 95%CI=0.86, 0.92).
The researchers suggested that the variation in results possibly indicates under- and over- estimation of the relationship between benzodiazepine use and dementia. The adjusted models likely underestimated the relationship through the unintended adjustment of colliders
(common consequences of the exposure and the outcome). The unadjusted models may have
67 | Page Daniel J. Hoyle overestimated the relationship via protopathic bias (initiation of a drug to treat the symptoms of the (currently undiagnosed) disease of interest). Taken together, the researchers suggested no causal link between benzodiazepine use and dementia.409
Lastly, post-mortem examinations of 298 brains did not reveal any neuropathological changes related to dementia based on exposure to benzodiazepine (n=49 patients) and/or anticholinergic medications (n=51 patients) at least one year prior to dementia diagnosis.408
Similar levels of Alzheimer’s disease neuropathologic change was seen between benzodiazepine users (64% with intermediate/high levels) and non-users (53% with intermediate/high levels of change). After adjusting for potential confounders and accounting for multiple testing, no significant relationships were found between benzodiazepine use and neuronal loss in the entorhinal cortex (adjusted OR=2.47, 95%CI=0.90, 6.76), nucleus basalis (adjusted OR=4.63,
95%CI=1.11) or substantial nigra (adjusted OR=3.30, 95%CI=1.02, 10.68). Similarly, a reduction in cortical microinfarcts with recurrent benzodiazepine use (adjusted OR=0.11, 95%CI=0.02,
0.63) was not statistically significant after accounting for multiple testing.408 Unfortunately, the small sample size limited statistical power. Furthermore, detailed information (e.g., duration of treatment, adherence to the drug) was unavailable.408
Biological mechanisms have been proposed to explain the potential relationship between benzodiazepine use and dementia. First, benzodiazepines are associated with amnesic and non-amnestic cognitive impairment.397 Amnestic cognitive impairment has been associated with a faster progression of Alzheimer’s disease.411, 412 Second, there is evidence that chronic benzodiazepine treatment is associated with uncoupling between the GABAA and benzodiazepine binding sites.413 A small study investigating the presence of flunitrazepam binding sites in the brains of seven people with Alzheimer-type dementia and seven people without dementia found that benzodiazepine binding sites were significantly lower (i.e., uncoupling) in the frontal and temporal cortex and hippocampus of dementia affected brains.414
68 | Page Daniel J. Hoyle Lastly, Billioti de Gage et al. suggested that chronic benzodiazepine use may limit the capacity of the cognitive reserve by reducing the utilisation of accessory neuronal networks.402 These accessory networks are important in the preservation of cognitive function in brains affected by early-phase brain lesions.415
Overall, there is some research to suggest an increased risk of dementia with long-term benzodiazepine use, particularly those with a long half-life. However, further research is required to determine the exact biological mechanism and to rule-out potential protopathic bias within earlier studies.
3.3.3.4 Falls and fractures
Similar to antipsychotics, benzodiazepines are also associated with an increased risk of falls in older people.195, 279, 281, 416-423 Benzodiazepines are thought to increase the risk of falls through several mechanisms, including psychomotor impairment (e.g., increased reaction time, impaired balance, gait), sedation, negative effects on cognition, and impaired vision.197, 295, 416
Several systematic reviews and meta-analyses have investigated the relationship between benzodiazepine use and falls and fractures.195, 279, 281, 417, 419 The most recent meta- analysis reported an increased risk of falls with both adjusted (n=14 studies) (OR=1.42,
95%CI=1.22, 1.65) and unadjusted data (n=26 studies) (OR=1.73, 95%CI=1.49, 2.01).195 These findings were similar to an earlier meta-analysis of adjusted data by Woolcott et al. in 2009.279
Seppala et al. also reported an increased risk of falls with long-acting benzodiazepines compared to short-acting benzodiazepines.195 This aligns with concern that longer-acting benzodiazepines may increase the risk of ‘hangover effects’ to a greater extent than shorter- acting benzodiazepines.196, 197, 424 ‘Hangover effects’ are adverse psychomotor and cognitive effects that are detected the day after being treated with a benzodiazepine.196, 197, 424
Nonetheless, results from other reviews have found no differences in the risk of falls based on
69 | Page Daniel J. Hoyle the half-life of the benzodiazepine used.278, 425 Seppala et al. attributed the differences in the results to differences in the selection of papers and the use of adjusted data.195
Unlike antipsychotics, benzodiazepine initiation and dose increase have been associated with an immediate risk of falls.416, 423 A 33-month longitudinal, observational study reviewing the risk of falls with benzodiazepine and antipsychotic use in RACF residents (n=594) found that the odds of a fall within 24 hours following benzodiazepine initiation was 3.79-fold greater than other times.416 Furthermore, stopping a benzodiazepine was associated with a significant reduction in fall risk (OR=0.26, 95%CI=0.08, 0.91).416 In contrast, the odds of having a fall within 24 hours following the initiation of an antipsychotic was not elevated compared to other times.416 Differences in the mechanisms whereby benzodiazepines and antipsychotics increase the risk of falls were suggested as a possible reason for the differences in results.416 In conflict with these results, another observational, longitudinal study of RACF residents (n=851) found that the risk of falls was elevated by a similar amount in the seven days following change
(initiation or increased dose) in any psychotropic drug (antidepressants, antipsychotics and benzodiazepines).423 Furthermore, the risk of falls was elevated four days before the drug change, possibly due to changes in the underlying psychiatric condition.423
A common complication of falls is the risk of fracture. A meta-analysis of 25 studies found that benzodiazepine use increases the risk of fracture by 1.26-fold in people aged ≥65 years.417 When only including hip fractures, the relative risk increased to 1.35.417 As discussed earlier (Chapter 3.2.5), hip fractures are associated with an increased risk of death272 and restriction of mobility,271 which may have consequences on independence and QoL.559, 560
3.3.3.5 Mortality
Several studies have identified an increased risk of all-cause mortality associated with chronic benzodiazepine use.426-429 A 2016 systematic review and meta-analysis of 25 studies (24 cohort and one case-control studies) reported a 43% increased risk of death among hypnotic and
70 | Page Daniel J. Hoyle anxiolytic users compared to non-users.427 Two of the studies included in the meta-analysis suggested that the risk of death may be more than three-fold greater compared to no treatment.426, 430 In fact, Kripke et al. reported an increased risk of death by more than five-fold in people taking >132 doses per year.430 After excluding studies that investigated the risk of mortality with z-drugs, the risk of mortality was 60% higher for benzodiazepine users compared to non-users.427 The small amount of evidence available suggested that the increased risk of death was associated with suicides, cancers and cardiovascular-related causes.427 However, causality could not be determined. It is possible that the results of these studies may be subject to protopathic bias where the underlying indication for the benzodiazepine treatment may be contributing to the increased mortality rather than the benzodiazepine itself. For example, the increased risk of suicides may be related to underlying depression rather than benzodiazepine utilisation; benzodiazepines are often used in people with depression to treat sleep disturbances. Depression is also associated with an increased risk of cardiovascular-related mortality.431
More recently, a large matched cohort study of community-dwelling people with
Alzheimer’s disease (n=10,380 benzodiazepine users, n=20,760 controls) investigated the impact benzodiazepines and related drugs had on mortality rates.429 Overall, the risk of death in users of benzodiazepines and related drugs was 41% greater compared to non-users. This relationship was significant for the first 120 days of benzodiazepine use but not use beyond this time.429 It should be noted that while benzodiazepine use was associated with an increased mortality rate, use of z-drugs were not.
Conversely, a large retrospective cohort study found no relationship between benzodiazepine initiation and all-cause mortality.432 Using a de-identified US commercial health care database (n=2,500,000), people starting benzodiazepines were matched with benzodiazepine non-initiators based on propensity score matching to control for potential
71 | Page Daniel J. Hoyle confounding. Prior to propensity score matching, analyses estimated a 79% increased risk of mortality in people initiating benzodiazepines compared to non-initiators. However, after propensity score matching, there was no increased risk of death with benzodiazepine initiation.432 Additionally, sensitivity analyses found no increased risk of all-cause mortality with age or the use of benzodiazepines with a longer half-life. Instead, there was a small but statistically significant increased risk of all-cause mortality in younger patients (hazard ratio=1.09, 95%CI=1.02, 1.15), those using shorter-acting benzodiazepines (hazard ratio=1.06,
95%CI=1.02, 1.10) and with a longer follow-up period (hazard ratio (12 month follow-up)=1.04,
95%CI=1.01, 1.08; hazard ratio (18 month follow-up)=1.05, 95%CI=1.02, 1.07).432 Patorno et al. suggested that residual confounding rather than the true effect of benzodiazepines could be driving the small increase in mortality risk.432 Other studies in older people have found no increased risk of death in older people taking benzodiazepines beyond the mortality attributed to falls and fractures.433, 434
3.3.4 Dependence and withdrawal syndrome
One of the major issues with prolonged benzodiazepine use is the rapid development of dependence. At the neural level, dependence to benzodiazepines is thought to develop through activation of dopaminergic neurons in the ventral tegmental area by modulation of GABAA receptors in nearby interneurons.435 Generally, other drugs (e.g., opioids) that increase dopamine release in the mesolimbic region of the brain are potentially addictive.436, 437
Benzodiazepine dependence in older people after long-term use has been characterised by the following set of symptoms:
• rebound anxiety and insomnia on withdrawal of the drug;
• a strong desire to use benzodiazepines;
• use of benzodiazepines despite falls;
• driving while under the influence of benzodiazepines;
72 | Page Daniel J. Hoyle • use of benzodiazepines in addition to other hypnotics; and,
• continuing to use benzodiazepines despite a prescriber’s recommendations.438
Although dependence is often seen in practice, the 12-month prevalence of benzodiazepine use disorder is estimated to be 0.04% in people 65 years and older.439 This is possibly explained by preference to diagnose this disorder in younger people who are polysubstance users rather than older users of benzodiazepine medications.439 In 2019, a prospective observational study of hospitalised older people in Norway (n=246) found that increased pain and use of multiple CNS drugs was related to an increased risk of dependence.440
Physical dependence to benzodiazepines can make it difficult to withdraw a benzodiazepine due to the risk of adverse drug withdrawal events. Adverse drug withdrawal events are defined as a ‘clinically significant set of symptoms or signs caused by the removal of a drug’.441 They may present as a return of the condition that the drug was treating (e.g., insomnia upon benzodiazepine withdrawal) or symptoms associated with physiological alterations caused by the medications (e.g., seizure after benzodiazepine withdrawal).442
Furthermore, these symptoms can vary depending on whether they present in the acute or chronic stage (Table 9). Withdrawal symptoms are likely to occur in patients who have been taking a benzodiazepine continuously for three to four weeks.443
73 | Page Daniel J. Hoyle Table 9: Symptoms of acute and chronic benzodiazepine withdrawal syndrome. Information from several sources.196, 443, 444
System Signs and symptoms Acute withdrawal Cognition Impaired concentration Motor Motor restlessness, tremor, muscle pain (limbs, back, neck, jaw), muscle twitching, gait impairment Autonomic Tachycardia, hypertension, sweating Life-threatening Generalised tonic-clonic seizures, hyperpyrexia, psychosis Chronic withdrawal Cognition Impaired concentration, reduced attention, amnesia Motor Muscle spasms Neurological Tinnitus, dizziness, headache, paraesthesia and other perceptual disturbance, visual disturbance Gastrointestinal Nausea, bloating, irritable bowel syndrome, anorexia Psychological Insomnia, anxiety, delirium
Considering withdrawal symptoms may present as worsening in the original indication
(e.g., worsened insomnia, anxiety), the patient or prescriber may think that the benzodiazepine was having a beneficial effect on the symptoms.445 Withdrawal symptoms develop faster with short-acting agents (within two to three days) compared to long-acting agents (within five to 10 days).446 Rebound of the original symptoms is the mildest form of withdrawal symptoms and tends to last a few days to weeks after discontinuation.446, 447
Rebound symptoms are more likely to occur with abrupt cessation of benzodiazepine use.447 Therefore, deprescribing guidelines generally recommend tapering the dose of the benzodiazepine by 20-25% every two weeks followed by a slower taper of 12.5% every two weeks near the end of stopping.448, 449 However, there is evidence that rebound and withdrawal symptoms may be less severe in older people compared to their younger counterparts.450
Schweizer et al. postulated that slower benzodiazepine clearance attenuates withdrawal symptoms and reduced neuronal capacity in older people results in less rebound hyperactivity with benzodiazepine withdrawal.450
74 | Page Daniel J. Hoyle 3.3.5 Summary
Benzodiazepines are widely utilised in the RACF setting for insomnia, anxiety and acute agitation.
Clinical guidelines generally do not endorse the use of benzodiazepines as a first-line option for these symptoms. However, benzodiazepines may be used if non-pharmacological strategies are ineffective. If used, the duration of use should be limited considering the risk of severe adverse effects, such as pneumonia, cognitive decline and dementia, falls and fractures, an increased risk of all-cause mortality, the development of tolerance to sedative and amnesic effects and dependence and withdrawal syndromes. If used for insomnia, guidelines generally recommend limiting the duration of continued use to less than two weeks. If used for anxiety, benzodiazepines can be used with a tapered dose over a six-week period as an adjunct alongside initial antidepressant therapy. Benzodiazepines should be limited to intermittent use if required for severe anxiety or agitation in people with dementia.
3.4 Inappropriate use of antipsychotic and benzodiazepine in residential aged care facilities
Clinical guidelines, discussed earlier in Chapter 3, encourage judicious use and ongoing review of antipsychotics, when used for neuropsychiatric symptoms associated with dementia, and benzodiazepines. Broadly, clinical guidelines recommend reserving antipsychotics to cases where the resident may harm themselves or others, or where non-pharmacological measures are ineffective.28, 31, 98, 106, 142, 218-220 These guidelines recommend that the antipsychotic should be reviewed, and the dose reduced if possible, every 12 weeks.31, 106 Furthermore, benzodiazepine use should be limited to two to four weeks when used for insomnia32, 448 and six weeks (with withdrawal after two weeks) when used for anxiety.30, 303 These guidelines reflect the modest efficacy that these medications have for these conditions and the risk of severe adverse effects. Despite these guidelines, there continue to be high prevalence rates and
75 | Page Daniel J. Hoyle long-term use of antipsychotics and benzodiazepines in RACFs internationally suggesting that residents are being inappropriately sedated.10, 16, 25, 41, 138, 451-470
3.4.1 High prevalence of antipsychotic and benzodiazepine prescribing
Internationally, a high prevalence of psychotropic prescribing in RACFs has been reported for more than three decades.10, 13, 15, 16, 39, 453, 458, 460, 464, 470-477 In 2019, Hasan et al. performed a systematic review to determine the international prevalence of CNS medications in RACFs within papers published between 2000 and 2018.471 A total of 89 articles were identified. The studies were conducted in Europe (n=32), Scandinavia (n=23), North America (n=17), Australia and New Zealand (n=15), and South East Asia (n=2). Fifty-seven studies reported on antipsychotic prevalence and 70 studies described benzodiazepine prevalence in RACFs. On average, a quarter of residents were prescribed antipsychotics (26.1%, 95%CI=25.1, 27.2%) and
36.2% (95%CI=32.2, 40.4%) were prescribed benzodiazepines (Table 10).471 However, it was unclear whether this systematic review solely included regularly charted (e.g., daily) CNS medications or if ‘when required’ (or pro re nata) CNS medications were also accounted for.471
Furthermore, different countries have different aged care systems (e.g., the presence or absence of a RACF medical director varies internationally).478 Differences in these systems may partly account for the prevalence rates of CNS medications.
76 | Page Daniel J. Hoyle Table 10: Use of antipsychotics and benzodiazepines by residents of residential aged care facilities according to country. Adapted from Hasan et al., 2019.471
Country Number Drug class Pooled proportion (%) 95% confidence of interval (%) studies Australia 7 Antipsychotic 24.8 21.3, 28.5 6 Benzodiazepine 27.4 17.4, 38.7 Austria 2 Antipsychotic 45.9 44.3, 47.5 2 Benzodiazepine 35.0 33.5, 36.5 Belgium 2 Antipsychotic 24.2 10.6, 41.0 3 Benzodiazepine 54.1 51.8, 56.4 Canada 5 Antipsychotic 28.6 24.5, 32.9 4 Benzodiazepine 23.9 12.3, 37.7 Finland 6 Antipsychotic 37.0 31.9, 42.2 4 Benzodiazepine 40.8 26.6, 55.8 France 3 Antipsychotic 26.0 23.8, 28.3 2 Benzodiazepine 48.4 38.4, 58.4 Germany 2 Antipsychotic 31.6 25.0, 38.7 New Zealand 3 Antipsychotic 20.3 16.6, 24.4 2 Benzodiazepine 35.1 32.6, 37.7 Norway 10 Antipsychotic 23.5 22.5, 24.5 10 Benzodiazepine 44.8 39.2, 50.4 Spain 2 Benzodiazepine 41.4 32.2, 50.9 Sweden 5 Antipsychotic 20.3 16.4, 24.7 4 Benzodiazepine 37.1 34.0, 40.3 United Kingdom 5 Antipsychotic 18.9 17.5, 20.3 United States 10 Antipsychotic 22.8 20.7, 24.9 5 Benzodiazepine 14.7 8.8, 21.8
In Australian RACFs, specifically, high rates of psychotropic prescribing have been reported for over 20 years.13 In 1995, a landmark paper by Snowdon et al. found that 58.9% of residents from 46 Sydney RACFs (n=2,414) were taking at least one psychotropic regularly; 27.4% were taking antipsychotics and 32.3% were taking benzodiazepines.15 This was the highest reported use of psychotropic medications from geriatric institutions internationally at that point in time and led to the formation of several Government committees, including the NSW
Ministerial Taskforce who released a report entitled, ‘Psychotropic Medication Use in Nursing
Homes’.479 Since then, follow-up studies of RACFs from central Sydney have reported an increase in antipsychotic prevalence following a period of reduction between 1993 and 2003, and a decrease in hypnotic and anxiolytic prescribing (Table 11).13, 39, 472
77 | Page Daniel J. Hoyle Table 11: Use of psychotropic medication in Sydney residential aged care facilities in 1993, 1998, 2003 and 2009. Adapted from Snowdon et al., 2011.13
Year (number of Antipsychotic Hypnotic prevalence Anxiolytic prevalence residents) prevalence 1993 (n=2414) 662 (27.4%) 641 (26.6%) 207 (8.6%) 1998 (n=1975) 447 (22.6%) 335 (17.0%) 123 (6.2%) 2003 (n=3093) 730 (23.6%) 350 (11.3%) 127 (4.1%) 2009 (n=2465) 690 (28.0%) 272 (11.0%) 117 (4.7%)
A 2018 literature review by Westaway et al. investigated the prescribing of antipsychotics in Australian RACFs between 2000 and 2017.464 Antipsychotic prescribing rates varied substantially between studies, ranging between 13 and 42%, with an average of 26% of residents prescribed at least one antipsychotic (Figure 13).464 Westaway et al. highlighted that the two studies with the highest prevalence of antipsychotic use were conducted in RACFs where the residents had limited proficiency in English and where residents were of greater age.464 The authors concluded that antipsychotic use in Australian RACFs has not decreased over the last two decades464 despite evidence of reduced risperidone use for dementia in Australia since the Therapeutic Goods Administration amended the indications for risperidone in June
2015.306 These amendments restricted risperidone prescribing to patients with moderate to severe Alzheimer’s disease only, for up to 12 weeks of treatment.480
45 42 40 38 33 33 35 30 30 28 27 24 24 23 23 24 25 21 22 20 14 15 13 10 5 Proportion(%) of residents 0
Figure 13: Prevalence of antipsychotic use in Australian residential aged care facilities by year of publication. Adapted from Westaway et al., 2018.464
78 | Page Daniel J. Hoyle More recently, the national expansion of RedUSe involving 11,368 residents from 139
Australian RACFs found that 21.8% of residents were charted antipsychotics and 21.6% of residents were charted benzodiazepines on a regular basis (e.g., daily).10 Additionally, 11.1% of
(n=1,261) and 30.5% of residents (n=3,467) were charted antipsychotics and benzodiazepines, respectively, on a ‘when required’ basis.10 It was suggested that the relatively lower rate of regularly charted antipsychotics is a result of the release of guidelines, safety warnings and prescribing restrictions.10 However, the authors warn of a compensatory increase in the use of other medications for neuropsychiatric symptoms, including benzodiazepines (22%), sedating antidepressants (17%) and anticonvulsants (5%).10, 189, 198, 481
3.4.2 Inappropriate duration of antipsychotic and benzodiazepine therapy
In their study of 559 residents with dementia residing in Dutch dementia special care units, Van der Spek et al. reported that only 9.6% of psychotropics were prescribed appropriately,138 according to a validated and reliable medication appropriateness index designed specifically for psychotropic use in dementia.482 In particular, only 4.7%, 2.6% and 5.4% of residents were prescribed anxiolytics, hypnotics and antipsychotics appropriately, respectively.138 An inappropriate therapy duration was a common reason for psychotropic prescribing to be considered inappropriate (Figure 14).138
100% 90.6% 90% 76.4% 80% 68.4% 70% 60% 50%
duration 40% 30% 20% 10% Proportion of residents prescribed psychotropicsan forinappropriate 0% Antipsychotics Anxiolytics Hypnotics
Figure 14: Proportion of residents prescribed antipsychotics, anxiolytics and hypnotics for an inappropriate length of time. Data from van der Spek et al., 2016.138
79 | Page Daniel J. Hoyle Prolonged antipsychotic and benzodiazepine use has also been reported in Australian studies.12, 469 In 2010, a longitudinal study of 2,389 residents from 40 Tasmanian RACFs found that 58.6% and 62.4% of residents were taking the same dose and agent of antipsychotics and benzodiazepines, respectively, between 2006 and 2007.12 This greatly exceeds the current 12- week review period for antipsychotics when used for neuropsychiatric symptoms and the recommended two to six-week period when benzodiazepines are used.28-32, 106, 142 Furthermore, approximately only a quarter of residents had their antipsychotic or benzodiazepine dose reduced or ceased between 2006 and 2007.12
More recently, Kalisch Ellett et al. estimated the duration of antipsychotic and benzodiazepine treatment based on Australian Government Department of Veterans’ Affairs administrative claims for residents (n=14,237 residents) of Australian RACFs between 2015 and
2016.469 Overall, the median duration of antipsychotic treatment was 180 days (~26 weeks) and the median duration of benzodiazepine treatment was 240 days (~34 weeks).469 However, the translation of this data to the general RACF population is questionable. First, only 14% of
Australia’s RACF population are veterans.483 Furthermore, the Department of Veterans’ Affairs funds interventions to improve the health of the veterans (e.g., the Veterans’ MATES program which provides written advice to GPs and pharmacists about potential medication problems with their veteran patients).484, 485
The high prevalence and long-term use of antipsychotics and benzodiazepines in RACFs despite the risk of adverse effects and modest efficacy for indications often presenting in RACFs suggests that there are significant barriers to antipsychotic and benzodiazepine dose reduction.
3.5 Barriers to antipsychotic and benzodiazepine reduction
Barriers to reducing antipsychotic and benzodiazepine use in RACFs are complex and wide- ranging.43-45, 143, 162, 187, 486-490 Whilst it was outside the scope of this thesis to perform a detailed exploration of these barriers, the major barriers have been briefly summarised to contextualise
80 | Page Daniel J. Hoyle why clinical outcome monitoring within antipsychotic and/or benzodiazepine dose reduction programs are important in improving the uptake of such programs.
In 2018, Sawan et al. performed a review of the literature to investigate the role of organisational culture on the use of psychotropic medications in RACFs.187 The review suggested that the perceived risk of an increase in behavioural symptoms and sleep disturbances, which could increase the workload of on-site staff, is a major barrier to the reduction and/or cessation of psychotropic medications.187 This barrier was identified within four studies including a qualitative study of 40 on-site and visiting staff from eight Australian RACFs,170 a cross-sectional survey of nurses and GPs responsible for 51 residents prescribed an antipsychotic for more than one month across four Belgian RACFs,44 a cross-sectional survey of nurses and GPs responsible for 109 residents prescribed a benzodiazepine for more than three months across five Belgian
RACFs,45 and a qualitative study of 29 on-site and visiting staff from 12 Dutch RACFs.491
The concern of a recurrence in neuropsychiatric symptoms with antipsychotic discontinuation is interesting considering evidence, from a Cochrane systematic review of 10
RCTs (n=632 participants), that antipsychotics can be safely discontinued in older people with dementia without an increase in symptom frequency or severity.104 In fact, there was evidence that discontinuation in people with mild neuropsychiatric symptoms reduced agitation.492 With that said, the included studies were considered to be of low quality with small sample sizes limiting the possibility of detecting statistically significant differences in symptoms.104
Furthermore, two of the larger studies included in the review reported an increase in neuropsychiatric symptoms upon antipsychotic discontinuation in people with more severe symptoms at baseline and people who initially responded well to antipsychotic treatment.492,
493 Lastly, it was unclear whether the residents were representative of the general RACF population. For example, this review specifically investigated the impact of antipsychotic discontinuation when used in people with dementia.104 However, antipsychotics are commonly
81 | Page Daniel J. Hoyle utilised inappropriately in residents without a documented diagnosis of dementia.138 Larger studies to clarify the impact that antipsychotic reduction has on the neuropsychiatric symptoms of residents with and without a documented diagnosis of dementia may help to minimise this barrier to antipsychotic reduction.
There are no systematic reviews of RCTs to reduce benzodiazepine use. However, a small study (n=38 residents) of benzodiazepine discontinuation within five Belgian RACFs has reported improvements in self-perceived sleep quality (p=0.013) and midnight awakenings
(p=0.041) among residents who had successfully discontinued their benzodiazepine at eight months (n=24).494 Furthermore, withdrawal symptoms, QoL, functionality and medication use was unchanged among successful discontinuers.494 Larger prospective studies of benzodiazepine reduction examining the effects on sleep disturbances are necessary to minimise the aforementioned barrier to benzodiazepine dose reduction.187
Neither the above systematic review of antipsychotic discontinuation104 nor a pilot study of benzodiazepine discontinuation494 investigated the effects on staff workload despite being identified as a major barrier to psychotropic reduction.187 Generally, staff outcomes associated with antipsychotic and/or benzodiazepine dose reduction (e.g., staff distress, job satisfaction) are not routinely investigated within interventions to reduce antipsychotic and/or benzodiazepine use. Instead, staff outcomes tend to be qualitative evaluations of staff experiences with aspects of the intervention rather than the impact of psychotropic reduction.495-497 Whilst there is some evidence that antipsychotic dose reduction within two educational interventions did not result in an increase in staff distress (studies are discussed further in Chapter 4),498, 499 there appears to be a lack of investigation into the effects that benzodiazepine dose reduction has on staff outcomes.
Another barrier to psychotropic discontinuation involves possible effects on QoL. In the
2014 study by Azermai et al., the perceived risk of deterioration in QoL was considered to be
82 | Page Daniel J. Hoyle the biggest barrier to antipsychotic discontinuation among nurses and GPs caring for 51 residents.44 This is somewhat confusing considering that psychotropic use generally reduces
QoL.500 However, the concern regarding QoL may reflect the fact that undertreated neuropsychiatric symptoms may have deleterious effects on QoL.491 Therefore, if the antipsychotic is effective in managing the neuropsychiatric symptoms, it may also be having a positive effect on QoL. Two studies in the Cochrane systematic review by Van Leeuwen et al. reported no difference in QoL between people who had their antipsychotic discontinued and those who continued their antipsychotic.104 However, the quality of the evidence for this outcome was low. Future studies are needed to investigate the possible effects that antipsychotic and/or benzodiazepine reduction has on QoL in order to minimise this barrier to discontinuation.
In summary, major barriers to antipsychotic and/or benzodiazepine reduction, including concern of escalation in neuropsychiatric symptoms and sleep disturbances, increased workload, and deterioration in QoL, have not been fully addressed in discontinuation trials to date. Whilst there is some evidence to suggest that reduction in these medications does not result in worsened neuropsychiatric symptoms, increased workload or deterioration in QoL, the studies have often utilised small sample sizes and restrictive selection criteria preventing generalisation to the general RACF population. Our systematic review, presented in Chapter 4, investigated the reporting of these clinical outcomes, among others, within real-world interventions to improve antipsychotic and/or benzodiazepine use in RACFs.
83 | Page Daniel J. Hoyle Chapter 4 Clinical and economic outcomes of interventions to reduce antipsychotic and benzodiazepine use within residential aged care facilities
Part of the research contained within this chapter has been published as Hoyle DJ, Bindoff IK,
Clinnick LM, Peterson GM, Westbury JL. Clinical and economic outcomes of interventions to reduce antipsychotic and benzodiazepine use within nursing homes: a systematic review. Drugs and Aging. 2018;35(2):123-34.
Reprinted with permission from: Springer Nature, Drugs & Aging, Clinical and Economic
Outcomes of Interventions to Reduce Antipsychotic and Benzodiazepine Use Within Nursing
Homes: A Systematic Review. Hoyle DJ, Bindoff IK, Clinnick LM, Peterson GM, Westbury JL. 2018
4.1 Introduction
As discussed in Chapter 3, antipsychotic and benzodiazepine medications are widely prescribed in RACFs.10, 16 Antipsychotics are commonly used to manage neuropsychiatric symptoms, such as aggression, agitation or psychosis,25 and benzodiazepines are widely used to manage acute agitation, anxiety and sleep disturbances.27 Antipsychotics and benzodiazepines are only modestly effective for these indications22, 35, 243 and pose a risk of severe adverse effects, particularly when used in older people.36, 37
Generally, clinical guidelines recommend reserving antipsychotics to cases where the resident may harm themselves or others, or where non-pharmacological measures are ineffective.31, 106 Furthermore, guidelines recommend that they should be reviewed, and reduced if possible, every 12 weeks31, 106 and benzodiazepine use should be limited to two to six weeks when used for these indications.32
84 | Page Daniel J. Hoyle Withdrawal or dose reduction of these medications may minimise the risk of adverse effects. However, these medications are often used long-term.12, 25 The risks associated with neuropsychiatric symptoms are substantial, including lower QoL and increased risk of physical injuries to the residents,226, 227, 231 increased distress and turnover of RACF staff,144, 234 and increased costs of care due to greater use of health services.238, 239 It is not surprising then that a significant barrier to the reduction of antipsychotic and/or benzodiazepine use is the belief among nursing staff and prescribers that the initial symptoms, such as neuropsychiatric symptoms, and QoL may worsen44, 45, 187 despite evidence suggesting otherwise.104, 494
Numerous interventions have been designed to encourage the reduction of antipsychotic and/or benzodiazepine medications in RACFs. However, reviews of these interventions have been focussed primarily on medication changes as a measure of their effectiveness, with little examination of the resultant outcomes for the resident, RACF staff, or health care system.46, 48, 501-504 Others have highlighted the need for more reporting of clinical and economic endpoints within these interventions.48, 503, 505-507 Also, reviews that have reported on the clinical outcomes of interventions to reduce antipsychotic or benzodiazepine use in older people have only included RCTs, despite many interventions utilising non-randomised study designs,508 or have focussed on interventions outside of the RACF setting.509
An investigation into the clinical and economic impact of these interventions may help to minimise barriers to reducing antipsychotic and/or benzodiazepine use. Therefore, this systematic review aimed to review interventions to reduce antipsychotic and/or benzodiazepine use within RACFs to examine the outcomes reported and the impact on residents, staff, and the health care system. This review was published in Drugs & Aging in
2018.47
85 | Page Daniel J. Hoyle 4.2 Methods
4.2.1 Definitions
In this review, the term ‘intervention’ refers to a strategy implemented in a RACF designed to reduce antipsychotic and/or benzodiazepine use. In particular, this review focussed on professional and organisational interventions, as defined by the Cochrane Effective Practice and
Organisation of Care (EPOC) Review Group.510 Examples include distribution of educational materials, educational meetings, auditing and feedback, and the formation of clinical multidisciplinary teams. Indirect interventions (such as changes to regulations or access to medication) were not included. Clinical outcomes were defined as indicators of health status
(e.g., neuropsychiatric symptoms, QoL, falls) for the residents.
4.2.2 Search strategy and selection criteria
The search strategy aimed to identify studies on interventions to reduce antipsychotic and/or benzodiazepine use in RACFs. A comprehensive search strategy was developed using MeSH and free-text terms. The strategy was developed for PubMed and adapted for other databases
(Scopus, ProQuest, CENTRAL, EMBASE, and CINAHL). The PubMed search utilised the following keywords: (‘antipsychotic*’ OR ‘benzodiazepine*’ OR ‘psychotropic*’ OR ‘sedative*’ OR
‘tranquili*’ OR ‘psychoactive*’ OR ‘anxiolytic*’ OR ‘anti anxiety’ OR ‘neuroleptic*’ OR ‘hypnotic*’
OR ‘anti-anxiet*’ OR ‘antipsychotic agents [pharmacological action]’ OR ‘benzodiazepines [mh]’
OR ‘psychotropic drugs [pharmacological action]’ OR ‘hypnotics and sedatives [pharmacological action]’ OR ‘tranquillizing agents [pharmacological action]’ OR ‘anti-anxiety agents
[pharmacological action]’) AND (‘nursing homes’ OR ‘nursing home’ OR ‘long term care’ OR
‘aged care facilities’ OR ‘housing for the elderly’ OR ‘residential care’ OR ‘residential treatment’
OR ‘assisted living facilities’ OR ‘care home’ OR ‘geriatric care’ OR ‘institutional*’ OR ‘dementia home*’ OR ‘dementia care’ OR ‘nursing homes [mh]’ OR ‘long-term care [mh]’ OR ‘homes for the aged [mh]’ OR ‘institutionalization [mh]’) AND (‘intervention*’ OR ‘project*’ OR
86 | Page Daniel J. Hoyle ‘implementation*’ OR ‘program*’ OR ‘program evaluation’ OR ‘health care’ OR ‘healthcare’ OR
‘program evaluation [mh]’) AND (‘reduc*’ OR ‘inappropriate prescribing’ OR ‘withdraw*’ OR
‘decreas*’ OR ‘inappropriate medication’ OR ‘inappropriate drug’ OR ‘inappropriate prescribing’) NOT (‘animals [mh]’ NOT ‘humans [mh]’) NOT (‘child [mh]’ OR ‘infant [mh]’ OR
‘adolescent [mh]’ NOT ‘aged [mh]’). Searches were conducted from the inception of the database to 25 July 2017. The reference lists of all included articles were checked for additional relevant studies.
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed.511 Two reviewers (DH and LC) independently screened the titles, abstracts, and full texts. Discrepancies were resolved by discussion. Selected papers were assessed against the following inclusion criteria: (1) interventions aimed at reducing prescribing of regularly administered antipsychotic and/or benzodiazepine medication; (2) RACF setting; and (3) written in English. Papers were not excluded based on study design. Articles were excluded under the exclusion criteria ‘wrong study design’ if the intervention did not match our definition of an intervention (e.g., placebo-controlled withdrawal trials, regulatory interventions were excluded), did not specifically target antipsychotic and/or benzodiazepine medications, or did not differentiate the effect that the intervention had on specific medication classes (e.g., results indicating a reduction in psychotropic medication without clarifying the specific effects on antipsychotics and/or benzodiazepines). Articles published as theses, abstracts, reviews, opinion pieces, and protocols were excluded. Interventions performed in the outpatient, hospital, or dementia-specific unit setting, or in residents with severe psychiatric conditions (e.g., bipolar disorder, schizophrenia) were excluded.
Following the initial analysis of the included studies, interventions that reported clinical and/or economic outcomes were examined further.
87 | Page Daniel J. Hoyle 4.2.3 Data extraction and quality assessment
Information was extracted from included articles that reported on clinical and/or economic outcomes. The following information was extracted by DH and checked by IB for accuracy and completeness: (1) author, country, year, study design, and time to follow-up; (2) number of participants (residents and RACFs) in intervention and control groups; (3) change in antipsychotic and/or benzodiazepine use; and (4) impact of the intervention on clinical and/or economic outcomes.
A quality assessment was conducted for each study using the Effective Public Health
Practice Project Quality Assessment Tool for Quantitative Studies,512 which evaluates studies based on six criteria: (1) selection bias, (2) study design, (3) confounders, (4) blinding, (5) data collection methods, and (6) withdrawals and drop-outs. Scoring was completed independently by two reviewers (DH and JB) and discrepancies were resolved by discussion or, in the case of one study,513 by a third reviewer (GP).
4.2.4 Data synthesis
Changes in antipsychotic and benzodiazepine use and clinical outcomes were reported in a variety of formats across the studies. Furthermore, the tools used to assess the clinical outcomes and the length of follow-up periods varied substantially between studies.
Consequently, pooling of the data for meta-analysis was not considered possible. Therefore, the data were tabulated, grouped per the intervention design, and discussed narratively.
4.2.5 Protocol
A predefined protocol was developed and registered with PROSPERO (PROSPERO 2016:
CRD42016038601).
88 | Page Daniel J. Hoyle 4.3 Results
A total of 1,028 unique articles were retrieved from the electronic searches. Thirty-five articles describing interventions to reduce antipsychotic or benzodiazepine use remained after excluding articles that did not meet the inclusion criteria (Figure 15). Only 14 of these articles
(describing 12 interventions) reported on clinical outcomes related to the residents, RACF staff, and/or the health care system; all further analyses were limited to these articles (Table 12). One intervention was represented across three articles, which described changes in neuropsychiatric symptoms,514 QoL,515 and apathy.516 No studies reported on the economic outcomes.
Figure 15: Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flowchart of the study selection process.
89 | Page Daniel J. Hoyle 4.3.1 Study characteristics
Four interventions (reported across six articles) utilised randomised, controlled designs.514-519
The remaining studies were either non-randomised, controlled trials,520, 521 interrupted time series,513 or before-and-after studies.498, 499, 522-524 The studies were published between 1992 and
2017, and were conducted in the United Kingdom (n=5), the United States (n=4), Canada (n=4), and Australia (n=1). The studies involved between 46 and 678 residents; in total, 2,852 residents were included. This does not include the study by Tjia et al.,513 which did not report on the number of residents included. None of the studies reported on the duration of medication use prior to withdrawal, although four of the interventions were targeted at residents who had taken antipsychotics for at least three months.498, 499, 514-517
4.3.2 Intervention characteristics
Articles were grouped into three categories according to their intervention: educational interventions (n=8), psychiatric support (n=1), and multicomponent interventions (n=5). Using the Cochrane EPOC checklist (Table 13), educational interventions498, 499, 513, 519, 521-524 mainly entailed educational meetings, distribution of educational materials, and educational outreach.
Within the psychiatric support intervention,520 staff referred residents directly to a psychiatric support service. Additionally, specific care plans were developed to assist antipsychotic withdrawal.520 All multicomponent interventions incorporated educational outreach for care staff.514-518 In addition, in 2016 Ballard et al.514-516 randomly assigned RACFs to receive an antipsychotic review and/or social interaction and/or exercise interventions. Given the social interaction and exercise interventions did not aim to reduce antipsychotic use, exploration of the results was limited to the antipsychotic review intervention. Gilbert et al.518 provided relaxation courses for the residents. Fossey et al.517 reported on a psychiatrist review intervention which recommended cessation of psychotropic drugs that had been prescribed for
90 | Page Daniel J. Hoyle more than three months or when behavioural problems had resolved. There was no apparent relationship between intervention type and changes in medication use or clinical outcomes.
4.3.3 Study quality
Table 14 describes the quality of the included studies. The global assessment for most studies was weak (n=7). The remaining studies were of moderate (n=6) and strong quality (n=1). The main area of weakness was the lack of blinding of participants and/or researchers to the intervention allocation. However, it needs to be acknowledged that blinding of participants would not be possible given the nature of the interventions. Furthermore, most studies did not account for differences between treatment groups, such as baseline neuropsychiatric symptoms, initial antipsychotic dose, and RACF size. There were no apparent differences in medication use and clinical outcomes based on study quality.
4.3.4 Outcomes
There were no clear differences in clinical outcome measures between interventions reporting significant reductions in antipsychotic or benzodiazepine use and those that did not.
4.3.4.1 Medication use
The aim of 11 of the interventions was to minimise the use of antipsychotics in RACFs.498, 499, 513-
517, 519-524 One study targeted benzodiazepine use exclusively.518 Overall, six interventions demonstrated statistically significant reductions in antipsychotic use498, 513-517, 519, 521 and one study significantly reduced the use of benzodiazepines.518
4.3.4.2 Neuropsychiatric symptoms
All interventions aimed at antipsychotic use reported on changes in neuropsychiatric symptoms.498, 499, 513, 514, 516, 517, 519-524 Outcomes that were allocated under the heading
‘neuropsychiatric symptoms’ included anxiety, depression, psychiatric symptoms, agitation, physically aggressive behaviour, verbally abusive behaviour, prevalence of rejecting care, and
91 | Page Daniel J. Hoyle apathy. Additionally, two studies514, 520 assessed changes in overall neuropsychiatric symptoms using the Neuropsychiatric Inventory psychometric test.525
Neuropsychiatric symptoms generally remained stable or improved marginally when compared to pre-intervention symptoms or a control group in most studies.498, 499, 513, 517, 520-524
However, three studies reported a significant deterioration in symptoms with reduced antipsychotic usage.514, 516, 519 Two of these studies assessed changes in neuropsychiatric symptoms within a multicomponent antipsychotic review intervention.514, 516 The antipsychotic review intervention resulted in worsened neuropsychiatric symptoms and apathy.514, 516
Furthermore, residents who discontinued their antipsychotic and had a Neuropsychiatric
Inventory score of greater than 14 at baseline experienced a mean deterioration of greater than
20 points on the Neuropsychiatric Inventory.514, 516 However, agitation and depression levels were unaffected. The effects of the antipsychotic review intervention were mitigated by the addition of a social interaction intervention.514, 516 In the third study, Avorn et al.519 reported a significant increase in depressive symptoms with trends towards increased behavioural issues and anxiety with antipsychotic reduction.
4.3.4.3 Additional health care utilisation
Two studies519, 520 reported on the utilisation of health care services. Avorn et al. reported no changes to the rates of hospitalisation or level of resident care required.519 Ballard et al. found that GP visits decreased and days that residents were admitted as a psychiatric inpatient remained unchanged as a result of their intervention.520 No studies reported on potential health care savings or costs.
4.3.4.4 Other
Eight studies498, 499, 514, 515, 517, 519, 520, 524 reported on the impact that antipsychotic reduction had on other outcomes, including memory, cognitive function (measured with the Mini-Mental
State Examination), sleep disturbances, mortality, ADLs, abnormal involuntary movement, QoL,
92 | Page Daniel J. Hoyle falls, care staff stress, and expressive language skills. Outcomes remained unchanged in all but three studies.515, 519, 520 Apparent improvements in cognitive function, memory, expressive language skills, the number of GP visits, deteriorations in sleep disturbances, and QoL were associated with antipsychotic reduction.515, 519, 520 The effect of antipsychotic reduction on QoL was inconsistent between studies. One antipsychotic review intervention reported a non- significant decline in QoL.515 However, the addition of a social interaction intervention mitigated the deterioration in this outcome measure.515 Other studies found no change in the QoL.517, 520
One study investigated the effect that benzodiazepine reduction had on sleep, cognitive function, subjective health score and emotional responsiveness.518 All outcome measures, besides an improvement in emotional responsiveness, were unchanged upon benzodiazepine reduction.518
93 | Page Daniel J. Hoyle Table 12: Summary of clinical outcomes reported in included publications.
Source, Country; Number of residents Change in Change in clinical outcomes Study Design; Follow- and/or residential aged AP and BZ up period care facilities included in use the analysis AP BZ Neuropsychiatric symptoms Additional health care Other Overall Other utilisation Educational interventions 519 Avorn, 1992, USA; All residents ↓ Anxiety: *worsened Hospitalisation rates: = Memory: *improved RCT; 5 months n=349 [I], 6 RACFs Depression: worsened Change in level of care: = Cognitive function: *improved n=329 [c], 6 RACFs * Behaviour: worsened Sleep disturbances: *worsened Mortality: = 521 Ray, 1993, USA; All residents ↓ = Behaviour: = non-RCT; 4 months n=194 [I], 2 RACFs n=184 [c], 2 RACFs 524 Thapa, 1994, USA; All residents receiving ↓* Behaviour: = Activities of daily living: = non-RCT (analysis antipsychotics Psychiatric symptoms: improved Abnormal involuntary movement: BA); 6 months n=271, 6[I] and 6[c] = RACFs Depression: = Cognitive function: = 523 Earthy, 2000, All residents = = Behaviour: *improved Canada; BA; 6 months n=198, 1 NH 499 Monette, 2008, All residents with ↓* = Behaviour: improved Care staff stress: = Canada; BA; 7 months dementia receiving antipsychotics regularly n=81, 1 RACFs 522 Vida, 2012, Residents with dementia ↓* Behaviour: = Canada; BA; 5 months receiving antipsychotics regularly n=46, 1 RACF 498 Monette, 2013, Residents with dementia ↓ = Behaviour: = Care staff stress: = Canada; BA; 6 months n=293, 2 RACFs 513 Tjia, 2017, USA; ITS NHs ↓ = Physically aggressive behaviour: = with control group; 6 n=93 RACFs [I] months n=831 RACFs [c] Verbally abusive behaviour: = Prevalence of rejecting care: =
94 | Page Daniel J. Hoyle Source, Country; Number of residents Change in Change in clinical outcomes Study Design; Follow- and/or residential aged AP and BZ up period care facilities included in use the analysis AP BZ Neuropsychiatric symptoms Additional health care Other Overall Other utilisation Psychiatric support 520 Ballard, 2002, UK; All residents with ↓ = GP visits: fewer Quality of life: = non-RCT; 9 monthsa dementia Days admitted as a Activities of daily living: = n=136 [I], 6 RACFs psychiatric inpatient: = n=88 [c], 3 RACFs Expressive language skills: improved Multicomponent interventions 518 Gilbert, 1993, Aus; All residents without ↓ Sleep: = RCT; 3 month dementia, acute illness Cognitive function: and who were not bed- = bound Subjective health score: = n=27 [I], 1 RACF Emotional responsiveness: improved n=33 [c], 1 RACF 517 Fossey, 2006, UK; All residents ↓ Agitation: = Quality of life: = RCT; 12 months n=176 [I], 6 RACFs n=170 [c], 6 RACFs Falls: = 514 Ballard, 2016a, All residents with ↓ worsened Depression: = Mortality: = Ballard, 2016b,515 and dementia Apathy: worsened Overall quality of life: c worsened Rajkumar, 2016,516 n=146 [antipsychotic Agitation: UK; RCT; 9 monthsb review], 8 RACFs = n=131 [no antipsychotic review], 8 RACFs AP antipsychotic, BZ benzodiazepine, RCT randomised controlled trial, non-RCT non-randomised controlled trial, BA before-and-after, ITS interrupted time series, RACFs residential aged care facilities, [I] intervention, [c] control, ↓ Reduced medication use, Improved outcome, Worsened outcome, * Indicates statistical significance was not reported, = No meaningful change in outcome aAntipsychotic prevalence was reduced in intervention and control homes; Neuropsychiatric symptoms are reported for residents who had their antipsychotic stopped. All other outcomes are reported as a comparison between the intervention and control homes. bResults displayed only for antipsychotic review versus no antipsychotic review as the other arms of the intervention were directed at the management of neuropsychiatric symptoms. cIndicates p-value is between 0.05 and 0.1
95 | Page Daniel J. Hoyle Table 13: Intervention descriptions (summary using Cochrane EPOC Data Collection Checklist).510
Study Professional Organisational Provider orientated interventions Structural educational material of Distribution meetings Educational Educational leaders opinion Local Audit and feedback Reminders teams multidisciplinary Clinical care of Continuity services of integration Formal site of service delivery Changes to the setting services and benefits of nature in and scope Changes monitoring mechanisms quality of organisation Presence and
outreach
Educational interventions Avorn 1992519 Ray, 1993521 Thapa, 1994524 Earthy, 2000523 Monette 2008499 Vida, 2012522 Monette, 2013498 Tjia, 2017513 Psychiatric Support Ballard, 2002520 Multicomponent interventions Gilbert, 1993518 Fossey, 2006517 Ballard, 2016a,514 Ballard, 2016b,515 and Rajkumar, 2016516
96 | Page Daniel J. Hoyle Table 14: Indicators of study quality.
Study Selection Study design Confounders Blinding Data collection Withdrawals Global bias method and drop- assessment outs Educational interventions Avorn, 1992519 Moderate Strong Weak Moderate Strong Moderate Moderate
Ray, 1993521 Moderate Strong Weak Weak Strong Moderate Weak Thapa, 1994524 Moderate Moderate Weak Weak Strong Strong Weak Earthy, 2000523 Weak Moderate NA Weak Weak Weak Weak
Monette, 2008499 Moderate Moderate NA Weak Strong Strong Weak Vida, 2012522 Moderate Moderate NA Weak Strong Strong Weak
Monette, 2013498 Moderate Moderate NA Weak Strong Weak Weak Tjia, 2017513 Moderate Moderate Moderate Weak Weak Moderate Weak
Psychiatric support Ballard, 2002520 Moderate Strong Strong Moderate Strong Moderate Strong
Multicomponent interventions Gilbert, 1993518 Moderate Strong Strong Weak Strong Moderate Moderate Fossey, 2006517 Moderate Strong Weak Moderate Strong Moderate Moderate
Ballard, 2016a514 Weak Strong Strong Moderate Strong Moderate Moderate Ballard, 2016b515 Weak Strong Strong Moderate Strong Moderate Moderate Rajkumar, 2016516 Moderate Strong Weak Moderate Strong Moderate Moderate
NA Not assessable
97 | Page Daniel J. Hoyle 4.4 Discussion
Thirty-five articles describing interventions to reduce antipsychotic and/or benzodiazepine use in RACFs were identified. However, only 14 of these articles reported on outcomes for the resident and/or staff, representing 12 intervention studies in total. Antipsychotic reduction was the target for 13 of these articles and invariably reported on changes in residents’ neuropsychiatric symptoms. Only one small multicomponent intervention exclusively aimed to reduce benzodiazepine use.518 None of the studies reported on the potential economic impact on the health care system.
This review demonstrates several key issues that have been mentioned in previous systematic reviews from related areas.48, 501, 503, 504, 509 First, there is a paucity of reporting of clinical and economic outcomes in studies aimed at reducing psychotropic use in RACFs. Second, when these outcomes are documented, there is substantial heterogeneity in their measurement and reporting. The heterogeneity in the reporting of antipsychotic and benzodiazepine use made it difficult to directly compare interventions. For example, changes in medication use were expressed using point-prevalence, percentage reduction, and weekly proportion of users. Third, there is a lack of high-quality research in this setting, mainly because of unclear methodologies, inappropriate statistical analyses, or omissions when reporting statistical significance.
Residents’ neuropsychiatric symptoms remained stable or improved in nine of 11 antipsychotic reduction interventions.498, 499, 513, 517, 520-524 The lack of deterioration in neuropsychiatric symptoms is similar to what has been observed in placebo-controlled trials of antipsychotic withdrawal.101, 104 Ballard et al.’s antipsychotic review intervention demonstrated an increase in neuropsychiatric symptoms upon antipsychotic withdrawal.514, 516 The authors suggested that the increase in symptoms was due to changes in prescribing behaviour over the last five years where prescribers are more likely to reserve antipsychotic treatment for those
98 | Page Daniel J. Hoyle residents with severe neuropsychiatric symptoms,514, 516 who may be at increased risk of deterioration upon antipsychotic withdrawal.492 Interestingly, the increase in symptom severity was mitigated by co-administering a social interaction intervention, emphasising the benefits of non-pharmacological interventions when performing antipsychotic reviews.514, 516
Although a wide variety of other clinical outcomes were documented, papers rarely reported on the same outcome. This affected our ability to compare and contrast studies.
Examples of these outcomes include QoL, which was only assessed in three studies,515, 517, 520 and falls, which were reported in just one.517 The lack of interventions reporting on these outcomes was surprising given that improved QoL and reduced falls are considered potential benefits of antipsychotic and/or benzodiazepine reduction, and evidence to show this may promote uptake.422, 526 Additionally, many of these outcomes could have been utilised in health economic evaluations. Unfortunately, cost-effectiveness analyses were not reported in any studies. Zwijsen et al.527 highlighted challenges associated with economic analyses in this setting, including difficulties utilising quality-adjusted life years for residents with dementia, given that the tools are relatively insensitive in these residents and improvements in life expectancy are often minimal.
This review highlights a lack of clinical and economic reporting from interventions to reduce benzodiazepine use in RACFs. Despite the high prevalence of use,11, 297 severe adverse effects,35, 402 and concern regarding the substitution of antipsychotics with benzodiazepines,528 only one intervention specifically aiming to reduce benzodiazepine use was included in our analysis.518 It is likely that most interventions have targeted antipsychotic use because of significant risks associated with use in people with dementia,248 safety warnings,249 updated professional guidance,22, 31, 142 and consequent worldwide political and media attention surrounding these agents.
99 | Page Daniel J. Hoyle This systematic review followed best practice guidelines for systematic reviews529 and is reported according to the PRISMA statement.511 Electronic searches were not limited by date or study design and, given that only two articles were excluded on the basis of being written in a language other than English, we are confident that we have captured most, if not all, the international data on this topic.
4.5 Conclusion
There is a need for interventions to report on clinical outcomes and, in particular, economic outcomes to justify the expenditure of scarce health and research resources. It is difficult to make definitive clinical recommendations given the scarcity and heterogeneity in outcome reporting. However, we found little evidence to suggest that deleterious effects will result from interventions to reduce antipsychotic or benzodiazepine medications. It seems reasonable, however, to suggest that non-pharmacological strategies, such as social interaction interventions, should be co-administered with antipsychotic review interventions to mitigate any potential negative effects of the latter on neuropsychiatric symptoms.514-516
In 2017, Millar et al.506 developed a set of core outcomes to assess trials aimed at optimising prescribing in older residents in RACFs to improve comparisons between interventions. This core outcome set includes items such as falls, QoL, and admissions to hospital (and associated costs). Adoption of these core outcomes in future studies will allow the comparison between interventions on a variety of clinical and economic outcomes.
4.6 Recent interventions
Since the above systematic review has been published, several interventions have reported on the clinical and/or economic outcomes of antipsychotic and/or benzodiazepine dose reduction.
These studies were identified through a non-systematic search of the literature immediately prior to the finalisation of this thesis in February 2020.
100 | Page Daniel J. Hoyle In 2018, Brodaty et al. implemented the Halting Antipsychotic use in Long Term care
(HALT) study across 23 RACFs in Sydney, Australia.530 Residents (n=139) were recruited if they were taking an antipsychotic regularly for three or more months without a primary psychotic illness (e.g., schizophrenia, bipolar disorder) or severe neuropsychiatric symptoms at baseline
(defined as an NPI-NH total score ≥50; domain scores of 12 for at least two of delusions, hallucinations, agitation/aggression, anxiety or disinhibition, and occupational disruptiveness score ≥4 for at least two of these domains).530 Overall, data were available for 133 residents following the intervention. Antipsychotic use was ceased in 126 residents (94.7%) at any time point; 86.2% were not taking a regular antipsychotic at three months, 79.1% at six months and
81.7% at 12 months post-intervention. Furthermore, 93.5%, 87.3% and 90.3% of residents had a dose reduction of ≥50% of the original dose at three, six and 12 months, respectively.530
Neuropsychiatric symptoms decreased slightly, albeit non-significantly, with antipsychotic deprescription when measured with the NPI-NH (total NPI-NH score was one point lower upon antipsychotic deprescription) and Cohen-Mansfield Agitation Inventory (CMAI)
(total CMAI score was 1.7 points lower upon antipsychotic deprescription) (Table 15). However, social withdrawal was slightly, although not significantly, increased with antipsychotic deprescription (Multidimensional Observation Scale for Elderly Subjects- Withdrawal subscale score was 0.27 points higher upon antipsychotic deprescription) (Table 15).530 The proportion of mobile residents who fell at least once decreased, non-significantly, in the six-months post- intervention compared to the six-months pre-intervention (44.7% versus 54.2%, p=0.32). Lastly, there was a non-significant reduction in the proportion of residents who were hospitalised at least once in the six-months post-intervention compared to the six-months pre-intervention among those who had their antipsychotic deprescribed (8.8% versus 18.8%, p=0.14).530
101 | Page Daniel J. Hoyle Table 15: Estimated effect of antipsychotic deprescription on clinical outcomes within the Halting Antipsychotic use in Long Term care (HALT) study. Data from Brodaty et al., 2018.530
Outcome Estimate (β) 97.5% confidence interval p-value NPI-NH total score -1.0 point -5.03, 3.02 0.58 CMAI total score -1.7 points -5.77, 2.47 0.37 MOSES-withdrawal subscale 0.27 points -0.67, 1.22 0.52 NPI-NH=Neuropsychiatric Inventory-Nursing Home version, CMAI=Cohen-Mansfield Agitation Inventory, MOSES=Multidimensional Observation Scale for Elderly Subjects.
In our systematic review,47 we reported on the clinical outcomes of residents involved in the proof-of-concept studies for the Well-Being and Health for People with Dementia (WHELD) program.514-516 Since then, the optimised WHELD intervention, including training in person- centred care for care staff, promotion of tailored person-centred activities and social interactions, and a system for triggering a review of antipsychotic medications by the prescribing physician at each RACF, has been implemented in 69 UK RACFs (n=847 residents with dementia).531 Residents were randomly allocated to the WHELD intervention (n=404 residents) or treatment-as-usual (n=443 residents) and assessed at baseline and nine-months.
Compared to residents receiving treatment-as-usual, residents receiving the WHELD intervention had a statistically significant 2.54-point benefit in QoL (95%CI=0.81, 4.28, p=0.0042), measured with the Dementia Quality of Life (DEMQOL)-proxy tool, a statistically significant 4.27-point benefit on agitation (95%CI=-7.39, -1.15, p=0.0076), measured with the
CMAI, and 4.55-point benefit on overall neuropsychiatric symptoms (95%CI=-7.07, -2.02, p<0.001), measured with the NPI-NH.531 However, antipsychotic use remained stable across the
WHELD and treatment-as-usual groups, although there was a low baseline use of antipsychotic medications (<10%).531 When the data were restricted to residents with clinically significant levels of agitation (CMAI>40), the WHELD intervention conferred a significant improvement in agitation (5.16, 95%CI=1.11, 9.20; p=0.01) and QoL (3.23, 95%CI=-5.54, -0.93; p=0.01) compared to treatment-as-usual.532
Cost analyses of the optimised WHELD intervention demonstrated an overall advantage for WHELD over treatment-as-usual due to savings in total health and social care costs.531 Total
102 | Page Daniel J. Hoyle health and social care costs saved with WHELD were approximately £4,740
(95%CI=-6,129, -3,156) compared to treatment as usual.531 In residents with clinically significant agitation (CMAI total score >40), health and social care costs tended to be lower in WHELD compared to treatment-as-usual, although statistical significance was not reached (£2,103,
95%CI=-13, 4,219).532
In 2019, Richter et al. reported on the adaptation of a UK intervention by Fossey et al.517 to a German setting within a cluster-randomised controlled study over 12 months.533 The intervention involved antipsychotic review by experienced physicians specialised in psychotropic drug treatment for older people at three, six and nine months for all permanent residents without schizophrenia or bipolar disorder. In addition, all physicians involved with the prescribing of antipsychotics were provided education (workshops or brochures). Both the control and intervention groups received the antipsychotic review intervention to minimise contamination due to physicians being involved in the care of residents in multiple sites.
However, only the intervention RACFs received training for selected staff on person-centred care.533 Overall, 18 RACFs (n=439 residents) were allocated to the intervention group and 19
RACFs (n=603 residents) were in the control group.533
Antipsychotic prescribing reduced in the control group (39.8% at baseline and 33.3% at
12 months) but was stable in the intervention group (44.6% at baseline and 44.8% at 12 months).533 After excluding residents who were recruited post-randomisation, a reduction in antipsychotic prescription rates were similar between the intervention and control groups at 12 months (prevalence difference 7.4%, 95%CI=-2.9 to 17.7; p=0.156).533 QoL (both self-assessed and proxy-assessed), agitation, and the proportion of residents having at least one fall were similar between the intervention and control groups when comparing outcomes at 12- months.533
103 | Page Daniel J. Hoyle A 2019 Spanish study retrospectively investigated the impact of three interventions to reduce psychotropic use and the effect this has on falls and mortality rates in residents with dementia.534 The interventions included multidisciplinary medication reviews, application of
“STOPP/START” criteria, and use of a decision aid to guide the use of psychotropics.534 Overall,
606 residents with dementia taking a psychotropic received one of the aforementioned interventions. The outcomes for these residents were compared to 1,047 residents with dementia taking a psychotropic who did not receive an intervention. All three interventions were associated with statistically significant lower antipsychotic, anxiolytic (predominantly benzodiazepines) and antidepressant dosages after four weeks (Table 16).534 Weeks et al. found that the number of documented falls in the intervention groups was not greater than in the control group.534 However, supporting data were not provided. Furthermore, none of the residents who died during the intervention period were receiving an intervention.534
Table 16: Percentage reduction in antipsychotic and anxiolytic doses based on the intervention used in the study by Weeks et al., 2019.534
Intervention Percentage reduction in Percentage reduction in antipsychotic dose (based on anxiolytic dose (based on milligram-equivalent daily doses of milligram-equivalent daily chlorpromazine) (95%CI) doses of diazepam) (95%CI) Multidisciplinary 23.3% (13.9, 32.8) 23.1% (18.3, 28.0) medication reviews Application of 27.7% (22.4, 33.0) 39.5% (35.5, 43.5) “STOPP/START” criteria Decision aids 24.8% (15.6, 33.9) 31.8% (25.5, 38.2) 95%CI=95% confidence interval
Lastly, a study called “Optimizing Practices, Use, Care and Services- Antipsychotics”
(OPUS-AP) was implemented across 24 Canadian RACFs in March 2017.535 The intervention involved guideline dissemination followed by training, coaching, clinical tools, evaluation of clinical practices, and a change management strategy to improve dementia-related care, including antipsychotic deprescribing. Antipsychotic dose reduction and/or cessation was attempted in 220 of 344 residents over nine months. Overall, 52.7% of residents had their antipsychotic ceased and 32.7% had their dose reduced.535 Residents who had their antipsychotic dose reduced or ceased were more likely to have their benzodiazepine ceased as
104 | Page Daniel J. Hoyle well. In terms of clinical outcomes, residents who had their antipsychotic ceased were more likely to have reductions in agitation/aggression (measured with the CMAI). Hallucinations and delusional symptoms were no more likely to present in residents who had their antipsychotic dose reduced or ceased. Lastly, the change in the number of falls residents had in the 30 days prior to the baseline and nine-month data collection periods did not differ according to changes in antipsychotic use.535 Currently, phase two of this intervention is being rolled out across 129
RACFs in Quebec, Canada.
105 | Page Daniel J. Hoyle Chapter 5 General methods
5.1 Aim of the thesis
The overall aim of this thesis was to explore associations between changes in the antipsychotic and/or benzodiazepine dose and clinical outcomes for the residents, disruptiveness to RACF staff, and health care-related costs, within the national expansion of RedUSe.
To achieve this aim, several objectives were developed. The primary objectives were to investigate associations between changes in antipsychotic and/or benzodiazepine doses and changes in:
• neuropsychiatric symptoms; and
• QoL.
The secondary objectives were to investigate associations between changes in antipsychotic and/or benzodiazepine doses and:
• changes in social engagement;
• changes in ADLs;
• the rate of falls;
• the rate of behavioural events;
• the rate and length of hospitalisations; and,
• the rate of GP consultations;
Additionally, RACF staff-related outcomes included investigation into:
• the relationship between changes in the occupational disruptiveness of
neuropsychiatric symptoms on RACF staff and the magnitude of change in
antipsychotic and/or benzodiazepine doses; and
• differences in job satisfaction among RACF staff at baseline and four months.
106 | Page Daniel J. Hoyle Lastly, medication-, hospitalisation- and GP-related costs associated with antipsychotic and/or benzodiazepine dose reduction were investigated.
5.2 Introduction
The purpose of this chapter is to describe the general methods used in this study. This chapter begins with a summary of the Reducing Use of Sedatives project before discussing the research approach taken to assess the impact of antipsychotic and benzodiazepine reduction on residents and RACF staff (i.e., RNs, ENs and PCAs). Following this, the process of recruitment, choice of measures, and methods of analysis are discussed, alongside challenges encountered.
5.3 Intervention
RedUSe originated in 2008 as the central component of Associate Professor Juanita Breen’s
(previously Westbury) doctoral thesis entitled, ‘Roles for pharmacists in improving the quality use of psychotropic medicines in Residential Aged Care Facilities’.497 The main aim of RedUSe was to reduce the inappropriate use of antipsychotics and benzodiazepines in RACFs.
In June 2013, Associate Professor Breen and her team at the University of Tasmania received a grant from the Government Department of Health and Ageing to expand RedUSe nationally (as of November 2015, the grant was administered by the National Aged Care Grants
Hub at the Department of Health as part of the ‘Dementia and Aged Care Service Fund’).217
Whilst the intervention is briefly summarised below, detailed information on the intervention is reported in Appendix A and earlier publications.50, 217
5.3.1 Components
The national expansion was implemented in 150 Australian RACFs between March 2014 and
April 2016 across six states and one territory over four implementation waves. Initially, key stakeholder groups (i.e., BUPA care, Southern Cross Care) were approached to commit their facilities. Additionally, expressions of interest were sought from RACFs by advertising the
107 | Page Daniel J. Hoyle project through publications put out by RACF advocacy groups, aged care conferences and ‘cold calling’ (this method was used to recruit facilities in rural locations). Over 300 expressions of interest were received and the 150 RACFs involved in RedUSe were purposely selected to ensure a representative sample based on the setting (e.g., urban versus rural), ownership (e.g., for-profit versus not-for-profit), and size. For logistical reasons, a minimum RACF size of 35 beds was set.10, 217 Considering expressions of interest were sought through multiple avenues of advertisement (e.g., aged care conferences, online publications put out by RACF advocacy groups), it is unclear how many RACFs were approached. Consequently, it was not possible to calculate a conversion rate.
The intervention comprised three main components: auditing, benchmarking and feedback of antipsychotic and benzodiazepine prescribing (baseline, three months and six months); the provision of interactive education to nursing and care staff (baseline and three months), pharmacists (baseline) and prescribers (baseline) on the evidence-based use of antipsychotics and benzodiazepines; and a multidisciplinary review of antipsychotic and benzodiazepine prescribing (sedative review; baseline and three months) for individual residents taking these agents. Figure 16 illustrates the implementation of the intervention alongside the time points for collecting clinical data for the evaluation described in Chapter 5.4.
108 | Page Daniel J. Hoyle
Figure 16: Implementation and clinical data collection timeline within residential aged care facilities (RACFs). Adapted from Westbury et al., 2016.217 QUM=Quality Use of Medicine, GP=General Practitioner, RedUSe=Reducing Use of Sedatives.
109 | Page Daniel J. Hoyle 5.3.1.1 Auditing and benchmarking of antipsychotic and benzodiazepine prescribing
A dedicated medication audit data-mining program was installed at each pharmacy supplying medications to the RACFs (‘supply’ pharmacies). The program extracted prescribing information from the pharmacies’ medication package software and uploaded the information to a secure website.217
The e-Health data-mining program was used to collect information about medication prescribing in each participating RACF at three time points (baseline, three months, and six months). The champion nurses (an RN or EN nominated from each RACF to deliver and promote
RedUSe in their RACF) subsequently verified dosing details for prescribed antipsychotics and benzodiazepines. In addition, the champion nurse removed details of residents who had died or left the RACF, indicated whether residents were receiving palliative care or respite care, and added residents not captured by the e-Health data-mining program (i.e., new residents and those receiving medications not supplied by the main ‘supply’ pharmacy for that RACF).
Psycholeptic medications that were not packed or identified by the packing program (e.g., clonazepam drops, risperidone wafers) were entered into the database separately.217 Post- intervention evaluations later revealed that the e-Health data-mining program identified over
95% of psycholeptic medications being taken.536
A major feature of the RedUSe data-mining program was the ability to produce an individualised report for each RACF. The report benchmarked the rates of antipsychotic and benzodiazepine prescribing, graphically, against the average antipsychotic and benzodiazepine prevalence rates reported at baseline for that wave of delivery. This report was presented to participating RACFs at baseline, three months and six months as a five-page document and embedded into an educational PowerPoint presentation for delivery during the staff education.217
110 | Page Daniel J. Hoyle 5.3.1.2 Education
5.3.1.2.1 Project pharmacist training
A project pharmacist was appointed in each of the six states/territories involved in RedUSe and one pharmacist was given the role of rural project pharmacist. The project pharmacists provided training to the champion nurses and QUM pharmacists (pharmacists who provided the training and performed the sedative reviews within each RACF) in their assigned states/territories. They were also required to liaise between the research team, RACFs, champion nurses, QUM pharmacists and supply pharmacies.217
Project pharmacists received two to three days dedicated training in Hobart on the project, change management theory, teaching strategy and team leadership.
5.3.1.2.2 Quality Use of Medicine pharmacist training
QUM pharmacists were required to complete a dedicated continuing education booklet on
RACF psychotropic use, followed by a full day’s dedicated training delivered by their project pharmacist. Components of the nursing staff training sessions were outlined during the training and QUM pharmacists were taught how to deliver the training and identify and manage barriers to change.217
5.3.1.2.3 Champion nurse training
The champion nurse was provided training on the nursing staff educational session and the project. After agreeing to participate in RedUSe, each champion nurse was required to read a four-page amended article called ‘Psychotropic medication use amongst older adults: what all nurses need to know?’537 and attend a four-hour training session delivered by their project pharmacist. During this training session, champion nurses were taught about the project, theory about psychotropic medication and strategies to identify and manage barriers to change.217
111 | Page Daniel J. Hoyle 5.3.1.2.4 Nursing staff training
Education was provided to the nursing staff over two one-hour workshops by the QUM pharmacists; the first session was at baseline and the second at three months. Whilst PCAs were permitted to attend the training sessions, only RNs and ENs were offered gift cards to support their attendance.217
The first nursing staff session was designed to generate enthusiasm and willingness toward psycholeptic reduction. The main objective was to challenge the belief that psycholeptic medications improve resident QoL. The workshop started with didactic education by the QUM pharmacist in the form of a seven-minute video and a short PowerPoint presentation. Nursing staff were also shown graphs that benchmarked antipsychotic and benzodiazepine prescribing in their own facility against the average use in other RACFs involved in the project. The workshop involved small-group work and used a case study to promote conversation over whether psycholeptic medications improve the QoL of residents. Additionally, each participant was given a supporting workbook, printed guidelines, and a booklet of the slides.217
The second nursing staff workshop (also delivered by the QUM pharmacist) included further information on non-pharmacological strategies, case studies, and provided repeated information on the benefits and risks of antipsychotic and benzodiazepine medication for those who attended the first session as well as vital information for new attendees. Additionally, a dedicated handout on common psycholeptic agents and side effects was provided to the attendees, along with a workbook.217
A 10-item Older Age Psychotropic (OAP) quiz538 was used to assess knowledge about psycholeptic medication before and after training sessions, and training evaluation forms were provided at the end of all sessions.217
112 | Page Daniel J. Hoyle 5.3.1.2.5 Prescriber academic detailing
Face-to-face visits to prescribers, commonly referred to as ‘academic detailing’, involved trained facilitators, from the National Prescribing Service (NPS) MedicineWise (for Victoria,
Tasmania, Queensland and New South Wales) and Drug And Therapeutic Information Service
(for South Australia), visiting prescribers at their place of practice to provide information on antipsychotic and benzodiazepine use, and discuss psychotropic prescribing in RACFs and the
RedUSe project.217
Prescribers were identified for academic detailing by their RACF. Each RACF were asked to provide a list of five to 10 prescribers who serviced their RACF most frequently. Each prescriber was sent an information letter about RedUSe, the RedUSe guidelines, a promotional flier, and a sample sedative review plan. The prescriber’s contact details were sent to NPS
MedicineWise, who subsequently invited each prescriber via fax and phone to an academic detailing session. The NPS also developed their own invitation leaflet for RACF staff to provide to attending prescribers who expressed interest in the RedUSe project.217
5.3.1.3 Sedative review plan
The first data-mining audit produced sedative review plans for residents taking regular doses of antipsychotics and/or benzodiazepines. The plan described the resident’s details, psycholeptic agent(s) taken and doses, and three sections for sequential recommendations by the QUM pharmacist, champion nurse and the resident’s regular prescriber. The pharmacist was required to provide the first recommendation, followed by the nurse, then the prescriber.
The sedative review plan generated by the RedUSe website included the option to use pre-programmed comments in the pharmacist and champion nurse sections of the plan to facilitate and accelerate the review process. The pre-programmed comments (detailed in
Appendix A.3.5) were available to select from a drop-down list on the RedUSe website for residents taking regular doses of medications with the ATC classification codes N05A, N05B
113 | Page Daniel J. Hoyle (except N05AB04 and N05AN01), N05C, and N03AE01. The pre-programmed comments were informed through guideline-based use of antipsychotics and benzodiazepines (e.g., if a resident was taking a long-term benzodiazepine, the pharmacist was offered the pre-programmed comment, “Tolerance develops to the hypnotic effects of benzodiazepines within two weeks.
May I suggest a gradual reduction in dose”). Whilst the pre-programmed comments were available, the reviewing pharmacist had the opportunity to use free text to create individualised recommendations.217
5.3.1.4 Other components of the RedUSe project expansion
5.3.1.4.1 Guidelines
Guidelines used in the national expansion of RedUSe were produced for the original RedUSe trial. The guidelines were based on recommended best-practice for antipsychotic and benzodiazepine use, from the International Psychogeriatric Association, the RANZCP, and the
Royal Australian College of General Practitioners.31-33, 77, 106 The guidelines included information on the risk/benefit profile for antipsychotics and benzodiazepines and guidance on recommended doses and duration of use. To assist with reducing these medications, a sample dosage reduction schedule was included.217 The guidelines were printed professionally on thick card and provided to more than 1,400 nursing staff and 80 pharmacists at the training sessions.
The guidelines were also distributed to approximately 600 prescribers.217
5.3.1.4.2 Video
An eight-minute video was made to introduce RedUSe, explain key strategies, and outline the benefits of the project for residents, doctors, nurse prescribers and nursing staff. The video was played at each of the training events for the QUM pharmacists, champion nurses, and in the first staff education. Additionally, the video was uploaded to YouTube and embedded onto the
RedUSe website.217 The video can be viewed at this link: https://www.youtube.com/watch?v=yIxC3IKu5PU.
114 | Page Daniel J. Hoyle 5.3.1.4.3 Resident and relative involvement
Resident and relative pamphlets promoting the appropriate use of benzodiazepines and a relative pamphlet promoting the appropriate use of antipsychotics for neuropsychiatric symptoms of dementia were developed.217 Additionally, small reply-to-sender postcards were produced for distribution to relatives and residents by the RACFs. The postcard contained website details and could be sent back to the research team as a request for further information.217
5.4 Study design
We used an observational, prospective study design to investigate the impact that antipsychotic and/or benzodiazepine reduction had on residents and nursing staff of RACFs involved in waves two, three and four of the national expansion of RedUSe.
Assessments of neuropsychiatric symptoms, QoL, social engagement, and ADLs were conducted as close as possible to the RedUSe baseline nursing staff educational program. On average, the initial assessments took place 26.9 days (standard deviation= 19.9 days) after the
RedUSe baseline nursing staff educational session. Follow-up took place approximately four months after baseline data collection (Figure 17). Resident outcomes were assessed through face-to-face (and, in two cases, phone) interviews with nursing staff using validated psychometric measures described below. Additionally, residents were personally interviewed with a QoL measure if they could provide informed consent. The QoL questionnaire was also sent to the ‘person responsible’ for a resident (e.g., family member, legal guardian) to complete about the resident (Figure 17). For the two cases where a phone interview was used, analyses were run with and without these residents, but no substantial differences were detected.
Therefore, these residents were included in the analyses to increase the power of our study.
Data on the number and severity of falls and behavioural events, number and reason for GP consultations, and reason for and length of time of hospitalisations for participating
115 | Page Daniel J. Hoyle residents were collected by the champion nurse between baseline and four months and the information was validated by the thesis author at four months (Figure 17).
Lastly, occupational disruptiveness related to neuropsychiatric symptoms was measured during psychometric testing with the NPI-NH and job satisfaction was measured anonymously by PCAs, ENs and RNs at baseline and four months through self-completion of the adjusted-Measure of Job Satisfaction (Figure 17).
Figure 18 illustrates which tools/elements were used for the above clinical evaluations and the results sub-chapter where the evaluation is reported.
A follow-up period of four months was chosen for several reasons. First, a shorter follow-up period may have reduced the number of antipsychotic and/or benzodiazepine reductions considering that there would be a time delay between the implementation of
RedUSe and alterations in prescribing. Second, a shorter follow-up period would have increased the risk that the collected data could reflect the transient withdrawal adverse effects that are sometimes seen with de-prescribing rather than a stabilised clinical outcome. A longer follow- up was deemed inappropriate given that the high rates of attrition in the RACF environment would mean a larger sample size would be required in an environment where it is difficult to recruit participants.539
It should be noted that there was some variability in follow-up time. Generally, residents were followed up between three and four months (mean follow-up time=118.2 days,
95%CI=116.7, 119.8 days). However, two residents were recruited late, well after the first training, and were followed up after 78 days. Analyses were run with and without these residents, but no substantial differences were detected. Therefore, these residents were included in the analyses.
116 | Page Daniel J. Hoyle
Figure 17: Data collection flowchart. *Resident consent sought if able to give informed consent otherwise consent sought from the ‘person responsible’. **Resident only interviewed if able to provide informed consent. PCA=Personal Care Assistant, EN=Enrolled Nurse, RN=Registered Nurse, MJS=Measure of Job Satisfaction, NPI-NH=Neuropsychiatric Inventory-Nursing Home version, CMAI=Cohen-Mansfield Agitation Inventory, MOSES=Multidimensional Observation Scale for Elderly Subjects-withdrawal subscale, PSMS=Physical Self-Maintenance Scale, AQoL- 4D=Assessment Quality of Life-4D, RACF=Residential aged care facility.
117 | Page Daniel J. Hoyle
Figure 18: Relationship between the tool or element used for each evaluation and the results sub-chapter they are reported. GP=General Practitioner.
5.5 Setting
5.5.1 Residential aged care facilities
The study was performed in the RACF setting. The Directors of Nursing from a convenience sample of 47 of the 150 RACFs (31%) involved in the national expansion of RedUSe were approached to participate in the clinical outcomes study (wave two=5 RACFs, wave three=21
118 | Page Daniel J. Hoyle RACFs, wave four=21 RACFs). The selection of RACFs was based on logistical and financial considerations. Most data collection was performed by a single investigator (thesis author) with occasional support from a research assistant. The data collection was performed across several states and air travel, car hire and accommodation were necessary in the majority of cases. To minimise travel time and, consequently, the cost of data collection, the decision was made to select RACFs close to major centres, where possible, and limit the involvement of smaller RACFs in isolated areas. Whilst convenience sampling techniques are widely used in research on older populations,540 this technique may introduce sampling bias where the study sample is not representative of the target population. This is further discussed in Chapter 6.1.
Thirty-one out of 47 (66%) RACFs consented to participate in this study (wave two=2
RACFs, wave three=16 RACFs, wave four=13 RACFs). Table 17 describes the location and number of RACFs in each wave. Three RACFs (one in wave three and two in wave four) failed to recruit any residents. Given the lack of engagement, these RACFs were considered to be withdrawn by non-participation. This left a total sample of 28 RACFs.
Table 17: Location and number of residential aged care facilities consenting to participation.
Wave State Number of residential aged care facilities Two Tasmania 2 Three New South Wales 3 South Australia 8 Queensland 5 Four South Australia 8 Victoria 5
5.6 Participants
5.6.1 Residents
Permanent residents were eligible to participate in the study if they were prescribed an antipsychotic and/or benzodiazepine for regular use (e.g., daily).
119 | Page Daniel J. Hoyle Residents were excluded from the study by the champion nurse if they had a documented diagnosis of a severe psychiatric condition, specifically schizophrenia or bipolar disorder, where antipsychotic therapy should generally be maintained long-term under the supervision of a psychiatrist. Reduction of antipsychotics and benzodiazepines for residents with these conditions was not a target of RedUSe.217 Residents were also excluded if they were receiving respite care given that they were unlikely to be present at the conclusion of the study.
Lastly, residents were excluded if they were deemed to be in the final stages of a palliative illness (as determined by the champion nurse), given that regularly charted antipsychotics and/or benzodiazepines are often clinically required in these scenarios.
5.6.2 Nursing staff
All RACF nursing and care staff (RNs, ENs and PCAs) from consenting RACFs (as described above) were eligible to complete an anonymous job satisfaction survey. The nursing and care staff were invited to complete the survey by the champion nurses of their respective RACFs.
5.6.3 Sample size justification
Estimation of the required sample size was calculated based on comparing changes in the primary outcomes (overall change in neuropsychiatric symptoms, measured with the NPI-NH total score, and QoL utility score, measured with the Assessment of Quality of Life-4D (AQoL-
4D) tool)525, 541 between residents who had their antipsychotic and/or benzodiazepine dose reduced (‘reducers’) versus those that did not (‘non-reducers’). Changes in the antipsychotic dose were calculated by converting the total daily antipsychotic dose to a total daily chlorpromazine dose equivalent. The total daily benzodiazepine dose was converted to a total daily diazepam dose equivalent. Several key references were used to calculate these conversions.542-544
Overall, it was calculated that 28 residents per group (antipsychotic/benzodiazepine reducers and non-reducers) were required to detect a clinically important change of 11 in the
120 | Page Daniel J. Hoyle NPI-NH total score,545 given a mean of 17.4 (standard deviation=14.6),546 with a significance level (alpha) of 0.05 and a power (beta) of 0.80.
Additionally, 44 residents per group (antipsychotic/benzodiazepine reducers and non- reducers) were required to detect a clinically meaningful change of 0.06 in the AQoL-4D utility score,547 given a mean of 0.06 (standard deviation=0.10),548 with a significance level (alpha) of
0.05 and a power (beta) of 0.80.
Based on the initial RedUSe trial,49 it was estimated that 30% of residents would have their psycholeptic medication(s) reduced or withdrawn during the course of the project.49
Therefore, 147 residents were needed to have at least 44 reducers. Additionally, an attrition rate of approximately 20% was expected. To take attrition into account, a total number of 184 residents per group taking antipsychotics and benzodiazepines were considered sufficient.
On average, each RACF in Australia accommodates approximately 75 residents.61 Earlier work by Associate Professor Breen found that 21% of residents took antipsychotics and 43% took benzodiazepines in a sample of Tasmanian RACFs.11 Under the assumption that RACFs recruited into this study would accommodate 75 residents with approximately 15 taking an antipsychotic and 32 taking a benzodiazepine, we calculated that at least 13 RACFs would be needed for the required sample size, assuming a 100% recruitment rate.
Considering that a 100% recruitment rate from each RACF is unlikely, we worked on the assumption that 40% of eligible residents would consent to the research. Using this recruitment rate, 31 RACFs would be required to achieve a sufficient sample size.
5.6.4 Recruitment
Directors of Nursing from a convenience sample of 47 RACFs involved in waves two, three and four of the national expansion of RedUSe were contacted, provided with a letter of invitation, information sheet and consent form for the study and asked to participate by a project
121 | Page Daniel J. Hoyle pharmacist or the thesis author (Appendix C). Follow-up phone calls to the RACF were made by the thesis author to answer any queries. Upon the provision of consent, the champion nurse was contacted so that the clinical outcomes research/recruitment methodology could be explained. The champion nurses were required to screen all residents and distribute letters of invitation, information sheets and consent forms to residents who were eligible to participate in the study (Appendix C). The champion nurse was required to perform this role as per ethical obligations to maintain the privacy of resident medical records. If the champion nurse believed that the resident had insufficient cognitive function to make an informed decision, the letter of invitation, information sheet, consent form, and a reply-paid envelope was sent to the ‘person responsible’ (usually a family member or enduring guardian) for the resident (Appendix C).
Additionally, the champion nurses were encouraged to call the ‘person responsible’ for the resident to ensure that the letters were received and to answer queries they may have about the research.
All nursing and care staff were eligible to participate in the job satisfaction survey. This survey was made available to the nursing and care staff by the champion nurse at each RACF.
The survey was anonymous, thus written consent was not required as stated in section 2.2.5
General Requirements for Consent in the National Statement on Ethical Conduct in Human
Research 2007 (“Consent may be expressed… by some other means (for example, return of survey, or conduct implying consent)”).549 In this case, completing the paper-based survey implied consent.
5.6.4.1.1 Recruitment rate
To accurately assess the resident recruitment rate, the champion nurses were asked to record the number of residents who were taking an antipsychotic and/or benzodiazepine regularly.
From these residents, the champion nurse was asked to record the number of residents that did not meet the inclusion criteria due to being classified as ‘palliative’ or ‘respite’ residents, or
122 | Page Daniel J. Hoyle residents who had a severe psychiatric diagnosis, such as schizophrenia or bipolar disorder. The champion nurse was also asked to record the number of residents approached for informed consent and the number of letters sent to the ‘person responsible’ for eligible residents.
Overall, record-keeping by the champion nurse was often incomplete due to high workload and competing priorities. Therefore, it was not possible to assess recruitment rate using this method. Instead, recruitment rate was estimated by comparing the number of antipsychotic and/or benzodiazepine users participating in the study to the total number of antipsychotic and/or benzodiazepine users at baseline in participating RACFs using the overarching RedUSe data. The results of this analysis are presented in Chapter 6.1.
5.7 Data collection methods
5.7.1 Outcome measures
A question booklet (Appendix D) and data collection workbook (Appendix E) were created to facilitate data collection at baseline and four months. Additionally, a register (Appendix F) was created for the champion nurse to record information on falls, behavioural episodes, GP consultations, and hospitalisations between baseline and four months. Data recorded in the register by the champion nurse was validated by the thesis author through an audit of the progress notes and GP consultation records at the RACF.
Collected data were entered into a custom-built Microsoft Access® database (Microsoft
Corporation, Redmond, WA, USA). Validation rules were added to numerical fields to minimise transcription errors when transcribing responses from the paper-based data collection tools to the Microsoft Access® database. Data entry was also double-checked for accuracy.
The thesis author was responsible for data collection, with additional assistance provided by two researchers (Donnamay Brown and Sue Edwards).
123 | Page Daniel J. Hoyle 5.7.1.1 Medications
Medication information was collected from the medication charts at baseline and four months and classified according to the ATC classification system.190 The ATC classification system, developed by the World Health Organization, groups “active medical substances according to the organ or system on which they act and according to their therapeutic, pharmacologic and chemical properties.”550 This allowed for the retrieval of medication-related data according to the individual drug or medication class (e.g., information on both risperidone or antipsychotics as a class could be extracted during analysis).
Information collected included the generic name of the drug, the prescribed strength, instructions, date last reviewed by the prescriber, and if the drug was prescribed for use ‘when required’ (also known as ‘PRN’ use). If the drug was intended for when required use, the number of times it was administered in the previous month was recorded.
Daily doses (in milligrams) of antipsychotics and benzodiazepines prescribed for regular use were converted to chlorpromazine and diazepam daily dose equivalents, respectively, using key references.542-544 It was noted that ‘drops’ were the unit of administration for clonazepam
2.5 mg/mL oral liquid. According to the Australian Medicines Handbook, the dose was calculated as 0.1 mg/drop.192
At four months, current and previous medication charts were audited for changes in prescribing of antipsychotic and benzodiazepine medications. The date and prescribed dose were recorded, where possible, for all changes in antipsychotic and benzodiazepine prescribing.
5.7.1.2 Medical history
The resident’s medical history (diagnoses) was obtained from the resident’s medical progress notes at baseline and four months and classified according to their International Classification of Diseases (ICD)-10 classification.76 The ICD-10 classification system, developed by the World
Health Organization, defines diseases, disorders, injuries and other health-related issues.551
124 | Page Daniel J. Hoyle Similar to the ATC system, the ICD-10 classification system is organised into standard groupings of diseases.551 This allowed for data retrieval based on specific aetiologies (e.g., residents with
Alzheimer’s dementia) or broader disease classifications (e.g., residents with dementia).
The resident’s medical history was used to calculate the Charlson Comorbidity Index552 at baseline and four months. Conditions within the Charlson Comorbidity Index were isolated by using ICD-10 codes as suggested by Quan et al.553 The index was then scored using an updated weighting system suggested by Quan et al. in 2011.554 The Charlson Comorbidity Index, developed by Charlson et al.,552 is a tool used for calculating one-year mortality rates based on comorbid conditions. For the purposes of the thesis, the Charlson Comorbidity Index was utilised as a measure of ‘illness’ of the residents.
5.7.1.3 Demographic details
Demographic details were collected at baseline from information available in the RACF.
Information collected included each resident’s:
• birth date;
• gender;
• country of birth;
• use of mobility aids;
• use of hearing aids;
• use of sight aids; and,
• date of admission to the RACF
5.7.1.4 Neuropsychiatric symptoms
5.7.1.4.1 Neuropsychiatric Inventory-Nursing Home version
Neuropsychiatric symptoms were assessed at baseline and four months through interviews with nursing staff familiar to the resident using the NPI-NH.555 The NPI-NH is widely used in both
125 | Page Daniel J. Hoyle research and clinical care settings and has been assessed for validity and reliability (Appendix
B.1.1).490, 498, 499, 501, 503 It has been specifically developed for professional caregivers in RACFs to evaluate overall neuropsychiatric symptoms throughout the previous two weeks in residents with and without dementia.525, 556 The NPI-NH assesses 12 domains of neuropsychiatric symptoms, including delusions, hallucinations, aggression/agitation, depression, anxiety, euphoria/elation, apathy, disinhibition, irritability, aberrant motor behaviour, sleep, and eating disturbances. The frequency and severity of the NPI-NH domains, over the previous two weeks, were rated on a four-point (1-4) and three-point scale (1-3), respectively.
The overall scores for each domain were calculated by multiplying the frequency and severity scores. The score for each domain ranged from 0 to 12, where higher scores indicated more severe and/or frequent symptoms. The total NPI-NH score was calculated by summing the 12 domain scores (overall score 0 to 144), with higher scores indicating the presence of more severe and/or frequent symptoms.525
Features of the NPI-NH tool include that it is:
• caregiver-based, therefore, does not require patient cooperation, and can be used in
very disturbed or advanced-disease patients;
• a screening-question strategy that minimises administration time;
• a tool that assesses both frequency and severity of neuropsychiatric disorders;
• a tool that assesses caregiver distress associated with individual neuropsychiatric
abnormalities;
• a tool that assesses conventional types of psychopathology that are readily recognised
by clinicians and commonly require treatment;
• a tool with well-established psychometric properties (ability to measure mental traits,
abilities and processes);
• sensitive to drug-induced behavioural changes;
126 | Page Daniel J. Hoyle • comprehensive; including an instructional module describing administration and
scoring techniques.557
5.7.1.4.1.1 Factor structure
In 2004, Lange et al. investigated the psychometric properties of the NPI-NH among 204 older people admitted to a neuropsychiatric hospital in British Columbia.558 Using exploratory principal axis analysis, the researchers reported a five-factor structure composed of agitation, mood, psychosis, sleep/motor activity, and elevated behaviour (Table 18).558
Table 18: Factors of the Neuropsychiatric Inventory-Nursing Home version.558
Factors Domains Factor 1- Agitation Agitation Irritability/lability Factor 2- Mood Depression/dysphoria Apathy/indifference Anxiety Factor 3- Psychosis Delusions Hallucinations Factor 4- Sleep/motor activity Aberrant motor behaviour Night-time behaviour Factor 5- Elevated behaviour Euphoria/elation Disinhibition
5.7.1.4.1.2 Clinically meaningful change in score
Several scores have been suggested as the minimally clinically important change in score.
Initially, Kaufer et al. proposed a clinically meaningful change score on the NPI of ±9 based on the estimate of change which represented a reduction or improvement of more than
50% in the average baseline NPI scores of their study participants.559 However, Iverson et al. identified that this method did not consider the reliability of the test instrument.560 Therefore,
Iverson et al. applied a method of estimating the change in the NPI-NH score by calculating the degree of measurement error on each subscale on two occasions of testing. This method aimed to establish a reliable change in score which is not attributable to measurement error. Within a sample of 52 geriatric inpatients in a Canadian psychiatric hospital, Iverson et al. established
127 | Page Daniel J. Hoyle that a change in the NPI-NH total score by ±22 points is required to reliably detect a clinically meaningful change in score.560 Zuidema et al. repeated the methods used by Iverson et al. in a cohort of 105 RACF residents with dementia from five Dutch RACFs. They concluded that a change of ±11 points is sufficient to detect a reliable (unlikely to be due to measurement error) and clinically meaningful change in the NPI-NH total score in this setting.545
For the purposes of this thesis, a change of ±11 points in the NPI-NH total score was considered clinically meaningful.
5.7.1.4.2 Cohen-Mansfield Agitation Inventory
Resident agitation was specifically assessed at baseline and four months through interviews with nursing staff familiar to the resident using the Cohen-Mansfield Agitation Inventory (CMAI).
The CMAI has been widely used in geriatric research, and the reliability and validity for this measure are well established in RACFs (Appendix B.1.2).545, 561, 562 The CMAI assesses the frequency (ranging from never, 1, to several times an hour, 7) of 29 agitated behaviours during the previous two weeks.563 The total CMAI score was calculated by summing all 29 behaviours
(overall score 29 to 203), where higher scores indicated more frequent behaviours.
Despite some overlap in symptoms rated by the NPI-NH and CMAI, both were necessary to comprehensively assess the neuropsychiatric symptoms. The CMAI can sensitively detect small changes in agitation and aggression (symptoms mainly targeted for with antipsychotic therapy) but cannot be used to assess other neuropsychiatric symptoms.545 On the other hand, the NPI-NH can be used to monitor changes in a wide range of neuropsychiatric symptoms but it is less sensitive to change.545 Utilisation of both scales also allowed comparison with a similar antipsychotic deprescription project (HALT) performed over the same time period.530 To minimise the impact of scale fatigue, the staff were regularly asked if they needed to take a break from psychometric testing and the CMAI was consistently rated before the NPI-NH.
128 | Page Daniel J. Hoyle 5.7.1.4.2.1 Factor structure
A three-factor structure has been proposed for the CMAI.561, 564 The CMAI factors include aggressive behaviour (hitting, kicking, pushing, scratching, tearing things, cursing or verbal aggression, and grabbing; total score ranges from 7 to 49), physically nonaggressive behaviour
(pacing, inappropriate robing or disrobing, trying to get to a different place, handling things inappropriately, general restlessness and repetitious mannerism; total score ranges from 6 to
42) and verbally agitated behaviour (complaining, constant requests for attention, negativism, repetitious sentences or questions, and screaming; total score ranges from 5 to 35).561, 564 The questions related to these factors are described in Table 19.561
Table 19: Factor structure of the Cohen-Mansfield Agitation Inventory.561
Factor Question Question number Factor 1- Hitting Q7 Aggressive behaviour Kicking Q8 Pushing Q10 Scratching Q15 Tearing things Q25 Cursing or verbal aggression Q4 Grabbing Q9 Biting Q14 Spitting Q3 Factor 2- Pacing Q1 Physically nonaggressive behaviour Inappropriate robing or disrobing Q2 Trying to get to a different place Q16 Handling things inappropriately Q22 General restlessness Q29 Repetitious mannerisms Q26 Factor 3- Complaining Q18 Verbally agitated behaviour Constant requests for attention Q5 Negativism Q19 Repetitious sentences or questions Q26 Screaming Q13
5.7.1.4.2.2 Clinically meaningful change in score
Zuidema et al. have proposed clinically meaningful changes in the CMAI total and factor scores by calculating the reliable change index (a true change in score that is not attributable to measurement error between observers or between two observation moments).545 They proposed that a change in total score by 8 and change in the factor-based scores of physically
129 | Page Daniel J. Hoyle aggressive behaviour (CMAI factor 1), physically nonaggressive behaviour (CMAI factor 2) and verbally agitated behaviour (CMAI factor 3) by 3, 6 and 4, respectively was sufficient to detect a statistically and clinically meaningful change in score.545 For the purposes of this thesis, these aforementioned changes in the CMAI total and factor scores were considered clinically meaningful.
5.7.1.5 Quality of life
Resident health-related QoL was assessed at baseline and four months through interviews with nursing staff familiar to the resident using the AQoL-4D psychometric tool.541 AQoL-4D is a validated and reliable (Appendix B.1.3) 12-item, multi-attribute utility tool which provides an overall utility and dimension scores.541, 565-569 There are four dimensions that can be derived from the AQoL-4D including independent living (self-care, household tasks, mobility), social relationships (friendships, isolation, family role), physical senses (sleeping, worrying, pain) and psychological wellbeing (seeing, hearing, communication). The AQoL-4D was chosen to measure QoL because it is capable of providing utility scores, which were necessary for the calculation of quality-adjusted life years in a basic cost-utility analysis included in the final report for RedUSe.217, 570
In addition to assessment by the nursing staff, the residents with capacity were interviewed with the AQoL-4D tool by researchers at baseline and four months if they had provided informed consent to participate in the study. The ‘person responsible’ for the residents were also mailed a copy of the AQoL-4D to complete about the resident at baseline and four months. The ‘person responsible’ for the resident was given six weeks to complete the questionnaire. A reminder letter was sent to non-responders after three weeks.
The AQoL tool was originally developed for self-completion;571 however, the developers of the AQoL-4D understand that questions may have to be answered by a proxy and suggest
130 | Page Daniel J. Hoyle that this is preferable to excluding cases from a study.571 Hawthorne et al. also recommend a stratified analysis be undertaken to investigate the impact of proxy scores on the findings.
5.7.1.5.1 AQoL-4D scoring
There are two ways to score the AQoL-4D utility and dimension values. The method of scoring depends on whether the AQoL-4D is being used as a multi-attribute ‘psychometric’ instrument, used to provide profile scores,571 or as a multi-attribute ‘utility’ instrument, used for economic evaluations.572 In this study, QoL was scored as a utility instrument to facilitate use in cost-utility analyses (within the overarching RedUSe project) and comparisons with other studies who had used utility scores.547, 548, 573-578 The lack of other studies reporting scores using the unweighted
(psychometric) approach to scoring the AQoL-4D also prevented comparison and limited discussion of the results.
Scoring of QoL as a utility instrument requires different weights to be applied to the dimensions of the AQoL-4D. In this study, the weighted AQoL-4D utility and dimension scores were calculated in the Statistical Package for Social Sciences (SPSS) version 25 (IBM Corporation,
Armonk, NY, USA)579 using a syntax created by the developers of the AQoL-4D.572 The AQoL-4D utility and dimension scores range from -0.04 (QoL worse than death) to 0 (QoL equivalent to death) to 1 (best QoL).541
5.7.1.5.2 Clinically meaningful change in score
Hawthorne and Osborne suggested a minimally important difference in the AQoL utility score of 0.06547 based on the mean change in utility score from the re-analysis of four longitudinal studies.575, 580-582 Given this, a mean change in the utility score of ≥0.06 was considered clinically significant within this thesis.
5.7.1.6 Social engagement
Social engagement was measured at baseline and four months through interviews with nursing staff familiar to the resident using the Multidimensional Observation Scale for Elderly Subjects
131 | Page Daniel J. Hoyle (MOSES)-withdrawal subscale.583 The MOSES-withdrawal subscale is an eight-item questionnaire that has been developed for the proxy assessment of the resident’s engagement in the previous week. The MOSES-withdrawal subscale has been shown to be valid and reliable in the RACF setting (Appendix B.1.4).523, 525 Each item is rated from 1 (lowest level of social withdrawal) to 4 (highest level of social withdrawal). The total score was calculated by summing the scores from each item. Scores vary from 8 to 32, where lower scores indicate higher levels of social engagement.584
Two of the questions (questions 4, ‘Friendships with other residents’, and 8, ‘Helping other residents’) on this scale contain five options where the fifth option was ‘Does Not Apply’.
The literature advises that residents with the “Does Not Apply” rating on these items are, in reality, generally less socially engaged than those receiving the classification for the most impairment.585 Previous research had either excluded residents with the “Does Not Apply rating”583 or had changed the rating to that representing the lowest level of engagement (i.e., a score of 4).585 It was not considered feasible to exclude residents from our relatively small sample; hence, the decision was made to change the rating for people who scored “Does not apply” to the rating representing the lowest level of social engagement (i.e., a score of 4).
5.7.1.6.1 Clinically meaningful change in score
A clinically meaningful change in score in the MOSES withdrawal subscale has not been reported in the literature. Therefore, a clinically meaningful change in the MOSES-withdrawal subscale was estimated based on the 0.5 of a standard deviation rule, using baseline scores. Using this method, a change in the MOSES-withdrawal subscale of ≥3.1 points was considered clinically meaningful. This aligns with the findings from a systematic review on reminiscence therapy for dementia, which estimated, using the 0.5 of a standard deviation rule, that a change of ≥3.1 points was needed for a clinically meaningful change in score.586
132 | Page Daniel J. Hoyle 5.7.1.7 Activities of daily living
The ability of a resident to perform ADLs was measured through psychometric testing of nursing staff at baseline and four months using a validated tool called the Physical Self-Maintenance
Scale (PSMS).587 The PSMS is a valid and reliable tool to assess the function of older adults
(Appendix B.1.5).587, 588 The PSMS contains six-items with five options from which to choose.
However, for this piece of research, we utilised an adjusted-PSMS with seven-items to facilitate comparison with another antipsychotic deprescription project (the HALT study),530 which was funded under the same grant opportunity. The question regarding toileting was duplicated for bowels and bladder to assess the residents’ toileting in more detail (in the original PSMS, a single question was used to assess bowels and bladder continence simultaneously). To allow comparison with the broader research literature, the toileting questions were collapsed during analysis and the score related to the lowest level of function was retained as the overall score on the toileting item.
There are two ways to calculate the PSMS score. The first involves assigning a score of
1 to responses corresponding to the highest level of functioning with all other responses receiving a score of 0. Once all items are summed, a total score ranging from 0 to 6 remains, where higher scores represent a greater ability to complete the ADL.587 Alternatively, the responses to each item can be assigned a value from 1 to 5, where higher scores represent greater dependence.589 All items may then be summed to provide a total score ranging from 6 to 30, where higher scores represent greater dependency levels. For this thesis, the latter scoring method was utilised considering that residents involved in this study are unlikely to be able to perform most ADLs independently, thus using the latter scoring method would increase the sensitivity to detect minor improvements in ADLs.
During data collection, the five options for each item were arbitrarily assigned a score from 0 to 4, where higher scores represented higher levels of functioning. For analysis, scores
133 | Page Daniel J. Hoyle were recoded in SPSS version 25 (IBM Corporation, Armonk, NY, USA)579 so that the highest level of function was assigned a value of 1 and the lowest level of function was assigned a value of 5, as mentioned in the previous paragraph.
5.7.1.7.1 Clinically meaningful change in score
A clinically meaningful change in the PSMS scale has not been evaluated in detail.590 However,
Lawton and Brody suggest that changes in the overall score by one or two points are clinically visible to nursing staff.587 Thus, we considered a change of one point in the PSMS scale to be clinically meaningful.
5.7.1.8 Falls
The champion nurse was requested to indicate the date, name of the resident, and severity of falls incurred by residents participating in this research project. To indicate the severity of the fall, the champion nurse was required to tick the box that most closely related to the severity of the fall (Table 20).
Table 20: Fall severity assessment.
Severity Example Minor Minor injury that does not require significant medical intervention (e.g., small skin tear, minor bruise) Moderate Injury requiring GP or outpatient management and not resulting in permanent disability Severe Severe injury resulting in hospitalisation and requiring substantial rehabilitation GP=General Practitioner
The definitions for the different levels of severity were adapted from the definitions for
‘accident and injury’ from the consequences table in ‘The Economic Value of Home Medicine
Reviews’ (VALMER) study.591
5.7.1.9 Behavioural episodes
The champion nurse was requested to indicate the date, name of the resident, and severity of behavioural episodes incurred by residents participating in this research project. To indicate the
134 | Page Daniel J. Hoyle severity of the behavioural episode, the nurse was required to tick the box that most closely related to the severity of the episode (Table 21).
Table 21: Behavioural episode severity assessment.
Severity Example Minor Managed in-house – No addition treatment needed (e.g., excessive wandering, calling out) Moderate GP needed, or treatment started (e.g., self-harm or struck staff/other residents) Severe Specialist visit Hospitalisation required GP=General Practitioner
The definitions for the different levels of severity were adapted from the definitions for
‘psychosis’ from the consequences table in the VALMER study.591
5.7.1.10 General practitioner consultations
The champion nurse was requested to indicate the date, name of resident, and reason for a GP consultation for residents participating in the clinical outcomes research project. GP characteristics were not collected as they were not readily available within the RACF records.
5.7.1.11 Hospitalisations
The champion nurse was requested to indicate the date, name of the resident, the reason for a hospitalisation admission and length of hospitalisation for residents participating in this research project.
At the four-month data collection time point, progress notes and GP consultation records for all participating residents were comprehensively audited by the thesis author to ensure the accuracy and completeness of data recorded by the champion nurse (including falls, behavioural events, GP consultations, and hospitalisations). Discrepancies were discussed with the nurse and records amended where necessary.
It needs to be acknowledged that whilst the original intention was for the champion nurse to collect the required information, issues with record keeping resulted in most data
135 | Page Daniel J. Hoyle being collected by the author through comprehensive auditing of the resident’s progress notes and GP consultation records. In many cases, the champion nurse simply did not have time to collect the required information. In other cases, the champion nurse utilised filters within their electronic records to identify the number of events. However, utilising filters in this manner missed events that were not recorded under the correct code leading to the need for a comprehensive review of all progress notes.
5.7.1.12 Job satisfaction
The adjusted-Measure of Job Satisfaction (MJS) is a 22-item tool used to measure job satisfaction of PCAs, ENs and RNs of RACFs by self-completion of the tool.592 Chou 2002 has shown the MJS to be a reliable and valid measure of staff satisfaction within the Australian RACF environment.592
Copies of the MJS survey were mailed to the champion nurse for distribution to nursing and care staff (RNs, ENs and PCAs) approximately two weeks prior to the researcher visiting the
RACF.
5.7.1.12.1 Factor structure
A five-factor structure has been proposed for the MJS.592 The factors include personal satisfaction, workload, team spirit, training, and professional support (Table 22).
136 | Page Daniel J. Hoyle Table 22: Factor structure of the adjusted-Measure of Job Satisfaction.592
Factor Item Personal satisfaction Q1. The feeling of worthwhile accomplishment I get from my work Q2. The extent to which I can use my skills Q3. The extent to which my job is varied and interesting Q4. The amount of personal growth and development I get from work Q5. The amount of independent thought and action I can exercise in my work. Workload Q6. The time available for resident care Q7. My workload Q8. Overall staffing levels Q9. The amount of time spent on administration Q10. The degree to which I am fairly paid for what I contribute to this organisation Team spirit Q11. The people I talk to and work with Q12. The contact I have with colleagues Q13. The value placed on my work by colleagues Training Q14. The opportunity to attend courses Q15. Time off to attend courses Q16. Being funded for courses Q17. The extent to which I have adequate training for what I do Professional support Q18. The amount of support and guidance I receive Q19. The opportunities I have to discuss my concerns Q20. The support available to me in my job Q21. The overall quality of the supervision I receive in my work Q22. The degree of respect and fair treatment I receive from my boss
5.7.1.12.2 Scoring
Each item was rated between 1 and 5, where higher scores indicated greater job satisfaction.592
To calculate factor scores, item scores within each factor were averaged.592 This provided a factor score between 1 and 5, where higher scores indicated greater job satisfaction.
Additionally, a total adjusted-MJS score was calculated by summing the scores for all
22-items. This provided a score between 22 and 110, where higher scores indicated greater job satisfaction.
5.7.1.12.3 Meaningful change in score
A meaningful change in score in the adjusted-Measure of Job Satisfaction has not been reported in the literature. Therefore, a meaningful change in the adjusted-Measure of Job Satisfaction was estimated based on the 0.5 of a standard deviation rule, using baseline scores. A meaningful minimum change in factor and total scores are reported in Table 23.
137 | Page Daniel J. Hoyle Table 23: Meaningful minimum change in factor and total scores for the adjusted-Measure of Job Satisfaction, utilising the 0.5 of a standard deviation rule.
Factor/total score Meaningful minimum change in score Personal satisfaction 0.3 Workload 0.4 Team spirit 0.3 Training 0.4 Professional support 0.4 Total 6.1
5.7.1.13 Occupational disruptiveness of neuropsychiatric symptoms
Occupational disruptiveness was assessed as a component of the NPI-NH.525 For each of the 12
NPI-NH domains, the proxy-assessor (nurse or PCA) reported the amount of disruption each symptom had on the nurse/PCA using a Likert scale ranging from 0 (not at all) to 5 (very severely or extremely).525
5.7.1.13.1 Scoring
A total score was calculated by summing the NPI-NH occupational disruptiveness score for each of the 12 NPI-NH domains. This gave a total score between 0 and 60 where higher scores indicated more severe occupational disruptiveness.
Considering the evidence that different neuropsychiatric symptoms have variable effects on the distress of staff,593 the NPI-NH occupational disruptiveness score for individual domains were also included in the analyses.
5.8 Data validation
Each record was checked for changes in psycholeptic medications including changes in the dose and the addition and cessation of regular (e.g., daily) and ‘when required’ psycholeptic medication between baseline and four months. The original data collection forms were rechecked, and data entry errors amended after clarification with the champion nurse when changes in antipsychotic and benzodiazepine use between the baseline and four-month
138 | Page Daniel J. Hoyle collection points were not consistent with the record of dose/medication changes between these time points.
The record of medical diagnoses was also checked, confirmed and retrospectively amended to improve consistency in the coding of diagnoses between baseline and four months.
5.9 Data storage
The data were stored in a password-protected, custom-built database in Microsoft Access 2016
(Microsoft Corporation, Redmond, WA, USA).
5.10 Statistical methodology
5.10.1 General analyses
The statistical methods used in this thesis were developed with the assistance of two experienced statisticians. A significant proportion of time was invested in developing models of analysis to minimise potential confounding inherent in our observational research.
Statistical analyses were performed using two statistical programs; SPSS version 25
(IBM Corporation, Armonk, NY, USA)579 for Windows and R statistical package version 3.4.3 (R
Foundation for Statistical Computing, Vienna).594
Analyses can be broken down into three main sections: representativeness of the convenience sample; identification of potentially confounding factors; and, clinical outcomes.
These are discussed in more detail below.
5.10.1.1 Representativeness of the convenience sample
Residents involved in the clinical outcomes study described in this thesis were recruited from a convenience sample of RACFs taking part in RedUSe. To explore the representativeness of the convenience sample of RACFs, we investigated similarities and differences between RACFs recruited into the clinical outcomes study and the RACFs involved in the overall RedUSe project
139 | Page Daniel J. Hoyle (see Chapter 6.1). Basic chi-squared (χ2) tests were used to compare categorical information and independent t-tests were used to compare numerical data between RACFs participating in the clinical outcomes study and the remaining RACFs involved in RedUSe. Similarly, χ2 tests were used to compare categorical information between RACFs participating in the clinical outcomes study and RACF-level data from the Australian Institute of Health and Welfare (AIHW).
Considering RedUSe did not collect demographic details from all residents, we also compared the residents recruited into the clinical outcomes study to resident-level data from the AIHW.595 Chi-squared tests were used to compare resident-level data from the clinical outcomes study to 2015 aged care data from the AIHW.595
5.10.1.2 Identification of confounding variables
Considering that this was an observational study, resident characteristics that may have influenced prescribing could also be related to the clinical outcomes seen with antipsychotic and/or benzodiazepine dose reduction. Therefore, it was important to identify these characteristics and control for their effect statistically, where appropriate, within analyses investigating the clinical outcomes associated with dose changes. The methods for the identification of potentially confounding variables are detailed in-depth in Chapter 6.2.
Linear mixed effects models were used to identify the potentially confounding variables.
Initially, a list of baseline resident characteristics that may contribute to antipsychotic and/or benzodiazepine prescribing were selected from the literature as fixed effects.156, 393, 400, 488, 489
Pair plots and domain-specific knowledge was used to assess multicollinearity (when two or more independent variables within a regression model are correlated) between variables.
Variables violating this assumption were removed from the analysis. To account for clustering, the resident and RACF identification number were included in the model as random effects.
Backwards stepwise selection with α=0.10 provided evidence for retaining variables or not.
140 | Page Daniel J. Hoyle Potentially confounding variables identified in the linear mixed effects models were included in multiple linear regression models, where appropriate, when investigating associations between the percentage change in antipsychotic and benzodiazepine doses and change in psychometric outcomes. The potentially confounding variables utilised in the antipsychotic models, where appropriate, included the CMAI factor 2 (physically nonaggressive behaviour) and CMAI factor 3 (verbally agitated behaviour) baseline scores. The potentially confounding variables utilised in the benzodiazepine models, where appropriate, included the
CMAI factor 2 (physically nonaggressive behaviour) baseline score and whether or not the resident had a documented diagnosis of dementia.
5.10.1.3 Psychometric outcomes
Associations between antipsychotic and/or benzodiazepine dose change and change in neuropsychiatric symptoms (measured with the NPI-NH and CMAI), QoL (measured with the
AQoL-4D), social engagement (measured with the MOSES-withdrawal subscale), and ADLs
(measured with the PSMS) were investigated using a structured approach.
First, residents who passed away, revoked consent, or relocated to another facility or the community were removed from the analyses.
Second, the analyses were split into residents who were taking antipsychotics at baseline (a chlorpromazine daily dose equivalent of >0 mg at baseline) and those who were taking benzodiazepines at baseline (a diazepam daily dose equivalent of >0 mg at baseline).
Residents taking both an antipsychotic and benzodiazepine at baseline were included in both sets of analyses.
Third, change in the outcome variable between baseline and four months was assessed in these drug groups using paired t-tests.
141 | Page Daniel J. Hoyle Fourth, multiple linear regression models controlling for potential confounders, where appropriate, were created to investigate associations between change in the outcome variable
(psychometric outcomes at four months were subtracted from the baseline scores) and per cent change in the chlorpromazine (for antipsychotic users) or diazepam daily dose equivalents (for benzodiazepine users) between baseline and four months.
The assumption of a normal distribution of the residuals for the model was investigated using quantile-quantile plots (Q-Q plots). A Q-Q plot is a scatterplot that plots quantiles of the data versus quantiles of a distribution to determine how the data is distributed. In the case of minor violations to this assumption, confidence intervals were bootstrapped. Bootstrapping uses the variability within a sample to estimate the sampling distribution.596 The assumption of linearity was also investigated by inspecting plots of the measures and antipsychotic and benzodiazepine doses at baseline and four months for non-linear distribution of the data.
Lastly, the data were restricted to residents who had completely ceased their antipsychotic or benzodiazepine. Paired t-tests were then used to investigate changes in psychometric measures between baseline and four months in these residents.
A p-value of 0.05 was set as the cut-off for statistical significance for all analyses.
5.10.1.4 Falls, general practitioner consultations, hospitalisations, behavioural events
Data on the number of falls, behavioural events, GP consultations, and hospitalisations collected over the four-month study period were converted to a standard unit of ‘events per 30 days’, considering slight differences in follow-up time.
As per Chapter 5.10.1.3, the analyses were split into those taking a regular antipsychotic and those taking a regular benzodiazepine at baseline. Correlations between change (%) in the chlorpromazine daily dose equivalent (for residents initially using antipsychotics) or diazepam daily dose equivalent (for residents initially using benzodiazepines) and the rate of events per
142 | Page Daniel J. Hoyle 30 days was investigated using Spearman’s correlation tests (rs) and interpreted according to
Swinscow.597
5.10.1.5 Influence of residents who had their dose increased
We investigated the effects of increased antipsychotic and/or benzodiazepine doses on the outcomes described in Chapter 5.10.1.3 and 5.10.1.4 by re-running the analyses after removing the small number of residents who had their dose increased. The results of these analyses were virtually unchanged by the removal of these residents, suggesting that the small number of residents who had their dose increased had little influence on the results.
5.10.1.6 Staff-related outcomes
5.10.1.6.1 Job satisfaction
Information on job satisfaction was collected anonymously from RACF staff at baseline and four months using the adjusted-MJS. The adjusted-MJS is scored across five factors (personal satisfaction, workload, team spirit, training and professional support) by averaging the item scores within each factor.592 Furthermore a total score is calculated by summing the score attained for all 22-items.
Considering that survey data could not be matched between baseline and four months
(i.e., potentially different people could have completed the surveys at each time point), independent t-tests were used to compare the factor and total adjusted-MJS scores at baseline and four months.
5.10.1.6.2 Occupational disruptiveness score
As discussed previously, ‘occupational disruptiveness’ was measured using the NPI-NH.
Associations between per cent change in the antipsychotic or benzodiazepine dose and change in the NPI-NH occupational disruptiveness scores were investigated using multiple linear regression models while controlling for potentially confounding variables, where appropriate.
143 | Page Daniel J. Hoyle 5.11 Cost analysis
A basic costing analysis was undertaken to gain an estimate of the cost-benefit of antipsychotic and benzodiazepine dose reduction. A comprehensive economic (cost-effectiveness) analysis could not be carried out due to the absence of a control group. However, by using the total cost of PBS subsidised medications, GP consultations on Medicare, and hospitalisation costs, we were able to calculate an estimate of the costs associated with antipsychotic and benzodiazepine dose reduction. The methodology utilised is described further in Chapter 6.9.
5.12 Feasibility of clinical evaluation
Consideration was given to the selection of clinical outcome measures that could feasibly be assessed within the RACF setting. We expected many residents to have underlying cognitive deficits. Thus, outcome measures that had been validated for proxy assessment by nursing staff were utilised where possible. Only QoL was assessed in the person responsible for the resident, the resident themselves (if cognitive capacity was deemed sufficient to provide informed consent), as well as the nursing staff.
The feasibility of measuring the clinical outcomes within this thesis using the aforementioned tools (Chapter 5.7.1) was assessed on the basis of completeness of data collection (i.e., number of residents who had an assessment performed relative to the total number of residents). These results are reported in Chapter 6.3 through 6.7, and Appendix K.2 and Appendix K.3.
5.13 Ethics and study registration
The study protocol was approved by the Tasmanian Health and Medical Human Research Ethics
Committee (approval number: H0013999) on the 26th of May 2014 (Appendix I).
The study was also registered as an observational study with the Australian and New
Zealand Clinical Trials Registry (ACTRN12617001262392).
144 | Page Daniel J. Hoyle Chapter 6 Results and discussion
6.1 Representativeness of residential aged care facilities within the clinical outcomes study
6.1.1 Introduction
RACFs participating in the clinical outcomes research described in this thesis were recruited from the RedUSe project via convenience sampling. Convenience sampling is a non-probability sampling method that is widely used in clinical research.598 This method involves the enrolment of participants according to their availability and accessibility.598 In our study, convenience sampling was used to overcome budgetary constraints (described in Chapter 5.5.1).
Whilst convenience sampling techniques are widely used in research on older populations,540 this technique may introduce sampling bias where the study sample is not representative of the target population. Therefore, it was necessary to investigate the representativeness of the residents participating in the clinical outcomes study.
Differences in the design of the clinical outcomes study and RedUSe meant that it was not possible to make in-depth comparisons based on resident characteristics. For example, residents were excluded from the clinical outcomes study if they had a documented diagnosis of a severe psychiatric condition whereas these residents were included in RedUSe. Second, the data collection period for resident-level data in the clinical outcomes study was four months compared to six months in RedUSe. Lastly, RedUSe was conducted at the RACF level and did not collect detailed resident information, such as demographics and medical conditions, for all residents.
We were, however, able to investigate similarities and differences between the 28
RACFs participating in the clinical outcomes and the 150 RACFs involved in RedUSe. Additionally, we were able to compare RACFs participating in the clinical outcomes study with national data
145 | Page Daniel J. Hoyle from the AIHW, and residents participating in the clinical outcomes study with the ‘average’
Australian resident based on data from the AIHW.595 Thus, as illustrated in Figure 19, this section of the thesis is dedicated to determining the representativeness of the sample in the clinical outcomes study by comparing RACFs involved in the clinical outcomes study with RACFs involved in the overall RedUSe project and AIHW data, and residents involved in the clinical outcomes study with the ‘average’ Australian RACF resident.
Figure 19: Flow of the analysis for the clinical outcomes study. This sub-chapter is dedicated to investigating the representativeness of the participants involved in the clinical outcomes study. GP=General Practitioner. 6.1.2 Method
6.1.2.1 Data extraction
RACF-level data from RedUSe (including RACF-level data on those RACFs involved in the clinical outcomes study) were extracted from the online database used in the national expansion of
RedUSe.217 This custom-built database housed information extracted from the pharmacy medication packing programs during the project. From this database, the author extracted the following information:
• the location (state) of the RACF;
146 | Page Daniel J. Hoyle • the rurality of the RACF;
• the number of residents per RACF;
• the proportion of residents using antipsychotics and/or benzodiazepines regularly at
baseline;
• the chlorpromazine and diazepam daily dose equivalent at baseline; and,
• the change in chlorpromazine and diazepam daily dose equivalent between baseline
and six months.
Where possible, the data extracted from the RACFs involved in the clinical outcomes study were compared to aged care demographic data from the AIHW collected in 2015.595
Resident-level data from the clinical outcomes study were extracted from a custom- built Microsoft Access database used to store the data. Extracted data were collected from information in the RACF records. This included:
• the age of residents;
• the gender;
• documented diagnoses of dementia at baseline of the study;
• time in care;
• number of medications (regular and when required);
• Charlson Comorbidity Index score;
• the baseline chlorpromazine and diazepam daily dose equivalents; and,
• the change in chlorpromazine and diazepam daily dose equivalents.
Where possible, the data were compared to aged care demographic data from the
AIHW collected in 2015.595
147 | Page Daniel J. Hoyle 6.1.2.2 Statistical analyses
Statistical analyses were split in three. The first compared RACF-level data between RACFs recruited into the clinical outcomes study and RedUSe. The second compared RACF-level data between RACFs recruited into the clinical outcomes study and data from the AIHW.595 The third compared resident-level data between residents involved in the clinical outcomes research and data from the AIHW.595
6.1.2.2.1 Residential aged care facility-level data
Chi-squared tests were used to compare categorical information between RACFs participating in the clinical outcomes study versus all RACFs involved in RedUSe. Specifically, these analyses compared the proportions of residents taking antipsychotics and/or benzodiazepines regularly at baseline, and the location (state) and rurality of the RACFs.
Chi-squared tests were also used to compare categorical information between RACFs participating in the clinical outcomes study versus demographic data from the AIHW.595
Specifically, these analyses compared the rurality, location and size of the RACFs.
Independent t-tests were used to compare the mean number of residents per RACF, the mean baseline chlorpromazine daily dose equivalent, the mean baseline diazepam daily dose equivalent, and the change in chlorpromazine and diazepam daily dose equivalents over six months in RACFs involved in the clinical outcomes study against all RACFs involved in RedUSe.
6.1.2.2.2 Resident-level data
Chi-squared tests were used to compare resident-level data from the clinical outcomes study to the 2015 aged care data from the AIHW.595 Specifically, these analyses compared the distribution of resident ages, gender, and proportion of residents with a documented diagnosis of dementia.
148 | Page Daniel J. Hoyle Paired t-tests were used to investigate changes in the chlorpromazine (for residents initially taking antipsychotics who were present in the clinical outcomes study at baseline and four months) and diazepam daily dose equivalents (for residents initially taking benzodiazepines who were present in the clinical outcomes study at baseline and four months) between baseline and four months.
6.1.3 Results
6.1.3.1 Comparison of residential aged care facility-level data in facilities involved in the clinical outcomes study versus all facilities involved in RedUSe
Overall, 28 RACFs took part in the clinical outcomes study from the 150 RACFs involved in
RedUSe. Comparisons between the RACFs participating in the clinical outcomes study and all
RACFs in RedUSe are presented in Table 24.
The number of residents taking an antipsychotic and/or benzodiazepine who resided in
RACFs participating in the clinical outcomes study (n=921) accounted for approximately 20% of antipsychotic and/or benzodiazepine users in RedUSe at baseline (n=4551).
Whilst the rurality of the RACFs were similar, a statistically significant difference in the proportion of RACFs in each state was noted. This is explained by the recruitment of a large proportion of South Australian RACFs into the clinical outcomes study (46.4%) compared to
RedUSe (21.3%). There was also a lower proportion of RACFs from New South Wales involved in the clinical outcomes study (10.7%) compared to RedUSe (19.3%) and no RACFs in the clinical outcomes study were from Western Australia or the Australian Capital Territory.
There were no statistically significant differences in the mean number of residents per
RACF or proportion of antipsychotic users in RACFs involved in the clinical outcomes study and
RedUSe. However, the proportion of residents using benzodiazepines regularly at baseline was slightly greater in RACFs involved in the clinical outcomes study (23.8% versus 22%, p=0.017).
149 | Page Daniel J. Hoyle No significant differences in the baseline mean chlorpromazine or diazepam daily dose equivalents or changes in the chlorpromazine and diazepam daily dose equivalents between baseline and six months were identified in RACFs involved in the clinical outcomes study compared to the overall RedUSe project.
150 | Page Daniel J. Hoyle Table 24: Comparison between the residential aged care facilities (RACFs) involved in the clinical outcomes study and RACFs involved in the Reducing Use of Sedatives (RedUSe) project.
Outcome RACFs involved in the clinical RACFs involved in RedUSe (n=150) p-value outcomes study (n=28) Rurality; percentage of RACFs (number of RACFs) 0.327 Major city 78.6 (n=22) 70.7 (n=106) Inner regional 21.4 (n=6) 24.0 (n=36) Outer regional 0 (n=0) 5.3 (n=8)
State/territory; percentage of RACFs (number of RACFs) 0.022* Australian Capital Territory 0.0 (n=0) 4.0 (n=6) New South Wales 10.7 (n=3) 19.3 (n=29) South Australia 46.4 (n=13) 21.3 (n=32) Tasmania 7.1 (n=2) 6.7 (n=10) Victoria 17.9 (n=5) 22.0 (n=33) Queensland 17.9 (n=5) 23.3 (n=35) Western Australia 0.0 (n=0) 3.3 (n=5)
Mean number of residents at baseline per RACF (95%CI; 86.4 (95%CI=72.9, 99.9; n=2,419) 81.1 (95%CI=75.9, 86.2; n=12,160) 0.328 total number of residents)
Percentage of residents taking an antipsychotic 21.7 (n=526) 21.9 (n=2,663) 0.837 regularly at baseline (number of residents)
Percentage of residents taking a benzodiazepine 23.8 (n=575) 22.0 (n=2,672) 0.017* regularly at baseline (number of residents)
Percentage of residents taking an antipsychotic and/or 38.1 (n=921) 37.4 (n=4,551) 0.462 benzodiazepine regularly at baseline (number of residents)
Mean chlorpromazine daily dose equivalent per 101.91 (95%CI=88.01, 115.81) 99.20 (95%CI=93.14, 105.25) 0.663 resident taking an antipsychotic at baseline (95%CI); mg
151 | Page Daniel J. Hoyle Outcome RACFs involved in the clinical RACFs involved in RedUSe (n=150) p-value outcomes study (n=28) Mean diazepam daily dose equivalent per resident 5.85 (95%CI=5.18, 6.52) 6.08 (95%CI=5.82, 6.35) 0.373 taking a benzodiazepine at baseline (95%CI); mg
Mean change in chlorpromazine daily dose equivalent -18.68 (95%CI=-25.53, -11.83) -16.13 (95%CI=-18.61, -13.65) 0.313 per resident taking an antipsychotic at baseline (95%CI); mg†
Mean change in diazepam daily dose equivalent per -1.42 (95%CI=-1.76, -1.08) -1.68 (95%CI=-1.85, -1.51) 0.119 resident taking a benzodiazepine at baseline (95%CI); mg† *p<0.05 †For residents present at baseline and six months.
152 | Page Daniel J. Hoyle 6.1.3.2 Comparison of residential aged care facility-level data in facilities involved in the clinical outcomes study versus information from the Australian Institute of Health and
Welfare
RACFs involved in the clinical outcomes study (n=28) were compared to data from the Australian
Institute of Health and Welfare (n=2,681).595 Whilst the rurality and size of RACFs were similar, a statistically significant difference in the proportion of RACFs in each state and territory was noted (Table 25).
Again, the difference in RACF location can be mainly attributed to the large proportion of South Australian RACFs involved in the clinical outcomes study (46.4%) compared to the proportion of South Australian RACFs reported by the AIHW (9.5%). There was also a lower proportion of RACFs from New South Wales involved in the clinical outcomes study (10.7% versus 32.6%).
153 | Page Daniel J. Hoyle Table 25: Comparison between the residential aged care facilities (RACFs) involved in the clinical outcomes study and RACFs data from the Australian Institute of Health and Welfare.595
Outcome RACFs involved in the clinical outcomes RACF data from the Australian Institute of p-value study (n=28) Health and Welfare (n=2,681) Rurality; percentage of RACFs (number of RACFs) 0.230 Major city Inner regional 78.6 (n=22) 60.7 (n=1,628) Outer regional 21.4 (n=6) 24.9 (n=667) Remote 0.0 (n=0) 12.5 (n=334) Very remote 0.0 (n=0) 1.3 (n=36) 0.0 (n=0) 0.6 (n=16) State/territory; percentage of RACFs (number of <0.001 RACFs) Australian Capital Territory 0.0 (n=0) 0.9 (n=25) New South Wales 10.7 (n=3) 32.6 (n=875) Northern Territory 0.0 (n=0) 0.4 (n=12) South Australia 46.4 (n=13) 9.5 (n=255) Tasmania 7.1 (n=2) 2.9 (n=78) Victoria 17.9 (n=5) 28.1 (n=754) Queensland 17.9 (n=5) 16.6 (n=444) Western Australia 0.0 (n=0) 8.9 (n=238)
Size of RACF; percentage of RACFs (number of 0.069 RACFs) 1-20 beds 0.0 (n=0) 5.4 (n=145) 21-40 beds 14.3 (n=4) 18.0 (n=482) 41-60 beds 10.7 (n=3) 24.4 (n=655) 61-80 beds 14.3 (n=4) 17.5 (n=470) 81-100 beds 28.6 (n=8) 13.5 (n=361) More than 100 beds 32.1 (n=9) 21.2 (n=568)
154 | Page Daniel J. Hoyle 6.1.3.3 Comparison of resident-level data for residents involved in the clinical outcomes study versus information from the Australian Institute of Health and Welfare
Two-hundred and six residents consented to take part in the clinical outcomes study out of 921 residents taking daily doses of antipsychotic and/or benzodiazepines at baseline in recruited
RACFs (response rate of 22%). This includes 99 antipsychotic users (19%) who agreed to participate in the study out of 525 residents taking antipsychotics regularly at baseline, and 134 benzodiazepine users (23%) who agreed to participate in the study out of 573 residents taking benzodiazepines regularly at baseline.
It should be noted, however, that some of the residents involved in RedUSe would have had a severe psychiatric illness (e.g., schizophrenia, bipolar disorder) or would have been receiving end-stage palliative care and would not have been included in the clinical outcomes study. It was not possible to identify and remove these residents from the RedUSe dataset, therefore, the true recruitment rate is likely to be higher. Fifty-six residents self-consented and
150 residents consented through the ‘person responsible’ for them (e.g., family, appointed guardian).
Residents participating in the clinical outcomes study were compared to data on permanent residents of Australian RACFs collected on the 30th of June 2015 by the AIHW (Table
26).595 No statistically significant differences in resident characteristics were identified.
However, our sample was slightly older (84% were ≥80 years) than that reported by the AIHW
(76.7% were ≥80 years) and slightly more residents had a documented diagnosis of dementia
(54.4% versus 51.9%).
It was not possible to compare the mean time in care, median number of medication
(regular and when required), mean Charlson Comorbidity Index score, baseline chlorpromazine and diazepam daily dose equivalent, or changes in the chlorpromazine and diazepam daily dose equivalents to data from the AIHW.
155 | Page Daniel J. Hoyle Table 26: Demographic details for residents at baseline of the clinical outcomes study and permanent resident information from the Australian Institute of Health and Welfare, where available.595
Variable Residents participating in the Resident data from the Australian p-value clinical outcomes study (n=206) Institute of Health and Welfare (n=172,045) % Age group* 0.135 <50 years 0.5% (n=1) 0.3% (n=555) 50-59 years 0.5% (n=1) 1.4% (n=2,353) 60-69 years 2.4% (n=5) 5.6% (n=9,623) 70-79 years 12.6% (n=26) 16.1% (n=27,733) 80-89 years 52.4% (n=108) 44.5% (n=76,497) 90-99 years 30.6% (n=63) 30.8% (n=52,939) ≥100 years 1.0% (n=2) 1.4% (n=2,344)
% Female 70.9% (n=146) 68.4% (n=117,633) 0.441
% documented diagnosis of dementia 54.4% (n=112) 51.9% (n=88,572)† 0.480
Mean time in care (95%CI); months 35.43 (95%CI=31.16, 39.70) -
Median number of regular medications at baseline per resident (range) 10 (2-22) -
Median number of ‘when required’ medications at baseline per resident (range) 4 (0-13) -
Median Charlson Comorbidity Index score (range) 2 (0-7) -
Mean baseline chlorpromazine daily dose equivalent for antipsychotic users 47.66 (95%CI=35.23, 60.09) - (95%CI); mg
Mean baseline diazepam daily dose equivalent for benzodiazepine users 5.30 (95%CI=4.57, 6.03) - (95%CI); mg
Mean change in chlorpromazine daily dose equivalent for antipsychotic users -6.88 (-11.59, -2.17) - between baseline and four months (95%CI); mgꬹ
156 | Page Daniel J. Hoyle Variable Residents participating in the Resident data from the Australian p-value clinical outcomes study (n=206) Institute of Health and Welfare (n=172,045) Mean change in diazepam daily dose equivalent between baseline and four -0.84 (-1.36, -0.32) - months (95%CI); mg‡ *age unknown for one resident in the Australian Institute of Health and Welfare data. †Out of 170,628 residents who had an Aged Care Funding Instrument assessment as of 30 June 2015. ꬹFor 83 antipsychotic users present at baseline and four months. ‡For 118 benzodiazepine users present at baseline and four months.
157 | Page Daniel J. Hoyle Follow-up data were available for 179 residents. The flow of participants through the study is presented in Figure 20. Of the total sample (n=179 residents), 83 (46%) were taking an antipsychotic and 118 (66%) were taking a benzodiazepine, with 22 (12%) taking agents from both classes, concurrently.
Figure 20: Diagram showing the flow of participants through the study. RACF=Residential Aged Care Facility.
For residents initially taking antipsychotics who were present at baseline and four months in the clinical outcomes research (n=83), the mean chlorpromazine daily dose equivalent decreased from 42.1 mg/resident/day (95%CI=33.7, 50.6) at baseline to
35.3 mg/resident/day (95%CI=25.8, 44.7) by four months, representing a 16.2% decrease in the mean dose (t(82)=2.905, p=0.005) (Figure 21).
For residents initially taking benzodiazepines who were present at baseline and four months in the clinical outcomes research (n=118), the mean diazepam equivalent dose declined from 5.1 mg/resident/day (95%CI: 4.3, 5.9) at baseline to 4.3 mg/resident/day (95%CI: 3.4, 5.2) by four months; a 15.7% decrease in the mean dose (t(117)=3.20, p=0.002) (Figure 21). As shown in Table 27, most antipsychotic and benzodiazepine dose changes were cessations.
158 | Page Daniel J. Hoyle However, it should be acknowledged that most residents remained on the same dose of antipsychotic (59%) or benzodiazepine medication (59%) between baseline and four months.
Figure 21: Beeswarm plots of the (A) chlorpromazine and (B) diazepam daily dose equivalents (mg) at baseline and four months. The blue dot represents mean and error bars indicate the 95% confidence intervals.
Table 27: Variation in the use of antipsychotics and benzodiazepines within a sample of residents involved in the clinical outcomes research between baseline and four months.
Dose change Antipsychotic agents Benzodiazepine agents Baseline to four months N (%) Baseline to four months N (%) Drug ceased 18 (22) 25 (21) Dose increased 4 (5) 7 (6) Dose decreased 12 (14) 17 (14) Same dose 49 (59) 69 (59) Total 83 (100) 118 (100)
6.1.4 Discussion
Our comparison of RACFs involved in the clinical outcomes study and RedUSe demonstrated similarities between the RACFs in terms of rurality, the number of residents per RACF, the proportion of residents taking antipsychotics, baseline chlorpromazine and diazepam daily dose equivalents and changes in chlorpromazine and diazepam daily dose equivalents between baseline and six months. However, statistically significant differences in the location (by state)
159 | Page Daniel J. Hoyle and a greater proportion of residents taking a benzodiazepine at baseline were identified. No statistically significant differences in resident-level data were found between the residents participating in the clinical outcomes study and 2015 data from the AIHW.
The higher proportion of baseline benzodiazepine users among RACFs involved in the clinical outcomes study (23.8% versus 22.0%) may be attributed to the location of RACFs. A large proportion of RACFs recruited into the clinical outcomes study were from South Australia
(46.4%). Baseline prevalence rates from RedUSe identified significantly greater benzodiazepine use among South Australian RACFs compared to other states and territories.10 Whilst it is not completely clear why South Australian RACFs tended to use benzodiazepines more, lower rates of antipsychotic use suggests possible substitution of antipsychotics with benzodiazepines.10
Supporting this theory, local guidelines recommend benzodiazepines as an alternative to antipsychotics for neuropsychiatric symptoms associated with dementia.484
Overall, 22% of all residents taking an antipsychotic and/or benzodiazepine at baseline were recruited from the 28 RACFs involved in the clinical outcomes study. It should be acknowledged that this recruitment rate is an underestimation considering that some of the residents taking antipsychotics and/or benzodiazepines from the RACFs may not have met the inclusion criteria for the clinical outcomes study, due to having a diagnosis of a severe psychiatric condition (e.g., bipolar disorder, schizophrenia) or receiving end-stage palliative care.
Furthermore, given the ethical requirement for arms-length recruitment, it was not always clear whether the champion nurse at each RACF had attempted to gain consent for all eligible residents.
Low uptake has been reported in many studies based in aged care.599, 600 Recently, the
MEDREV study reported a recruitment rate of only 31.5% (31 of 108 eligible residents).601 Whilst this recruitment rate was slightly higher than reported in this thesis, research assistants were utilised to assist with recruitment in MEDREV.496 This was not feasible in our study considering
160 | Page Daniel J. Hoyle the vast distances between RACFs and the ethical requirement for arm-length recruitment. The participants also differed to our study, with MEDREV recruiting residents with dementia taking a wide range of psychotropic medications (e.g., antidepressants, anti-dementia drugs, antipsychotics, benzodiazepines).601
A major issue with recruitment in the aged care setting is that participants often lack the capacity to provide informed consent.600, 602 In our study, 56 residents (27%) had the capacity to provide informed consent. Consent was acquired from the person responsible for the resident in the remaining 150 residents (73%). Our low recruitment rate may reflect difficulties acquiring consent from the person responsible for residents without the capacity to self-consent.600, 602 Barriers to obtaining consent from the person responsible for the resident include difficulties contacting the relative and concerns regarding the resident’s privacy, lack of benefit derived from participation, the topic of the study being inappropriate, and the absence of financial compensation.602, 603 To our knowledge, earlier psychotropic reduction studies have not reported on the proportion of residents who are recruited through self-consent or proxy- consent. However, a study investigating proxy assent within a randomised controlled trial of vaccination testing in UK RACFs had a lower proportion of residents recruited through the proxy
(43%) compared to our study (73%).602 This reflects the higher proportion of cognitively impaired residents included in our study. Future research, particularly observational studies, may overcome difficulties recruiting participants without the capacity to self-consent by utilising an opt-out approach.599
The distribution of resident age groups, gender and documented diagnosis of dementia were similar between residents involved in the clinical outcomes study and data from the AIHW.
Furthermore, the age and proportion of females in our sample is similar to that reported in earlier interventions.514, 522, 531 Although not statistically significant, our resident cohort had a greater proportion of people with a documented diagnosis of dementia. This is not surprising
161 | Page Daniel J. Hoyle considering entry into the clinical outcomes study required residents to be taking an antipsychotic and/or benzodiazepine at baseline and these medications are often used to mitigate neuropsychiatric symptoms associated with dementia.25, 27 Rather, the proportion of residents with a documented diagnosis of dementia was expected to be much higher considering this inclusion criterion. However, it is possible that a lack of documentation may account for the lower than expected proportion of people with a documented diagnosis of dementia.80
It was not possible to compare the mean time in care, median number of medications
(regular and when required), mean Charlson Comorbidity Index score, baseline chlorpromazine and diazepam daily dose equivalents, and change in chlorpromazine and diazepam daily dose equivalents to data from the AIHW. However, the time spent in care and Charlson Comorbidity
Index scores in our sample were comparable to earlier papers in similar patient groups.530, 604
On the other hand, the number of medications taken per resident was slightly higher in our study compared to an earlier retrospective audit of RMMRs.605 This is not unexpected though considering that our cohort was slightly older.605
The average chlorpromazine daily dose equivalent at baseline among residents recruited into the clinical outcomes research (47.66 mg, n=99 residents taking antipsychotics) was less than half the average chlorpromazine daily dose equivalent at baseline across the
RACFs involved in the clinical outcomes study (101.91mg, n=525 residents taking antipsychotics) and across all RACFs involved in RedUSe (99.20mg, n=2,663 residents taking antipsychotics).
Differences between the residents recruited into the clinical outcomes research and RedUSe likely account for these differences. For example, analyses in RedUSe were performed at the
RACF-level and involved all residents. On the other hand, the clinical outcomes research was performed at the resident-level and excluded residents with severe psychiatric conditions (e.g., schizophrenia, bipolar disorder). Residents with severe psychiatric conditions are often treated
162 | Page Daniel J. Hoyle with higher doses of antipsychotics. For example, risperidone is recommended at daily doses of up to 6 mg for schizophrenia and bipolar disorder18, 19 but only 2 mg daily for neuropsychiatric symptoms associated with dementia.28, 31
The changes in the chlorpromazine and diazepam daily dose equivalents between baseline and six months were also similar between RACFs involved in the clinical outcomes research (chlorpromazine daily dose equivalent reduced by 18.68 mg and diazepam daily dose equivalent reduced by 1.42 mg) and RedUSe (chlorpromazine daily dose equivalent reduced by
16.13 mg and diazepam daily dose equivalent reduced by 1.68 mg).
Resident-level data in the clinical outcomes research was only collected between baseline and four months and reported a reduction in the chlorpromazine daily dose equivalent by 6.88 mg and diazepam daily dose equivalent by 0.84 mg. The differences in the change in chlorpromazine daily dose equivalent may be partially explained by the lower baseline antipsychotic doses used by residents recruited into the clinical outcomes research. The use of lower doses would reduce the extent to which the dose could be reduced.
Secondly, the shorter follow-up time for resident-level data in the clinical outcomes research may have limited the opportunity for residents taking antipsychotics and/or benzodiazepines to have their dose reduced relative to the longer, six-month follow-up period in RedUSe. Although, the proportion of residents who had their antipsychotic and/or benzodiazepine dose reduced (14% for both medications) or ceased (22% for antipsychotic users and 21% for benzodiazepine users) was similar to that reported in RedUSe, which found that 15% of residents had their antipsychotic and/or benzodiazepine dose reduced and 24% had their medication ceased.50 Furthermore, the proportion of residents who had no dose change in our study (59% for both medications) was similar to that reported in RedUSe (52% for antipsychotic users and 54% for benzodiazepine users).50 The lack of dose change among most
163 | Page Daniel J. Hoyle residents emphasises that reducing the use of these medications is challenging even within robust, multifaceted interventions to improve their use.50, 531, 601
6.1.4.1 Limitations
Although best efforts have been made to reduce potential bias within the data, and the results presented above do indicate the sample is largely representative of the population, due to the nature of the research there are some limitations which may need to be considered when interpreting the results.
First, there were inherent differences in the residents recruited into the clinical outcomes study compared to those in RedUSe. One example is that residents with severe psychiatric conditions (e.g., bipolar disorder, schizophrenia) were excluded from the clinical outcomes study but not RedUSe. Thus, generalisation of the results from the clinical outcomes study should be limited to permanent residents without a diagnosis of a severe psychiatric condition (e.g., bipolar disorder, schizophrenia) and who are not receiving end-stage palliative care.
Second, on average, residents involved in the clinical outcome study took lower doses of antipsychotics compared to the average antipsychotic dose in RedUSe. Therefore, the results of our study may be more applicable to residents taking lower doses of antipsychotics (i.e., those taking low-dose antipsychotics for neuropsychiatric symptoms associated with dementia).
6.1.4.2 Conclusion
Overall, RACFs involved in the clinical outcomes study were relatively similar to RedUSe.
However, the over-representation of RACFs from South Australia and a greater proportion of benzodiazepine users at baseline should be considered when interpreting the results of the study.
164 | Page Daniel J. Hoyle Residents participating in the clinical outcomes study were not substantially dissimilar to the average Australian RACF resident (as per 2015 AIHW data). However, it should be acknowledged that it may not be appropriate to generalise resident-level outcomes reported in this thesis to the broader population within the RedUSe project without acknowledging inherent differences in resident characteristics associated with the recruitment process.
165 | Page Daniel J. Hoyle 6.2 Identification of potentially confounding variables
6.2.1 Introduction
This chapter describes an observational, prospective cohort sub-study of a sample of residents involved with the national expansion of RedUSe. The purpose of this sub-chapter was to identify differences in resident characteristics in order to control for them as potentially confounding variables in upcoming analyses. Funding rules prohibited clinical trial research methodology, including randomisation or the use of control groups. Ultimately, it was the prescriber’s decision whether or not to reduce the dose of the antipsychotic and/or benzodiazepine medication within RedUSe.217 Several resident and facility characteristics may have influenced these prescribing decisions. As just one example, there is published evidence of worsened neuropsychiatric symptoms upon antipsychotic reduction in residents with an initial NPI score >14.492, 514 Consequently, antipsychotic dose reduction may have been avoided in these residents.
The differences in resident characteristics influencing psycholeptic prescribing may have also affected the clinical outcomes reported in our study. Using the previous example, if antipsychotic dose reduction was preferentially attempted in residents with less severe neuropsychiatric symptoms at baseline, the likelihood of worsened neuropsychiatric symptoms with dose reduction may be mitigated.
In addition to resident characteristics, the prescribing decision may have been influenced by the organisational culture of a RACF.43, 143, 187, 489, 490 Organisational culture has been defined in various ways.43 These definitions often encompass descriptions of ‘shared beliefs, attitudes, values, and norms of members of an organisation’.606
Considering that resident characteristics may have differential effects on prescribing and these characteristics may contribute to the clinical outcomes seen upon antipsychotic
166 | Page Daniel J. Hoyle and/or benzodiazepine reduction, it was necessary to attempt to identify these baseline characteristics and statistically control for their effect, where appropriate, within analyses investigating the clinical outcomes associated with dose changes (Figure 22). Additionally, it was important to identify and control for the influence of unknown (or unmeasured) factors, such as organisational culture, on prescribing in order to account for the extent to which these resident characteristics may have influenced prescribing. Thus, the aim of this sub-study was to identify fixed (known) baseline resident characteristics that may have influenced the prescribing of psycholeptic medication whilst controlling for the random (unknown/unmeasured) effects of the RACF ‘environment’ (e.g., the influence of organisational culture, prescribers) and residents.
Figure 22: Flow of the analysis for the clinical outcomes study. This sub-chapter is dedicated to identifying characteristics related to changes in antipsychotic and benzodiazepine doses. GP=General Practitioner. 6.2.2 Methods
6.2.2.1 Statistical analysis
Several analyses were investigated prior to the design of the model described below. However, insufficient data, violation of distributional assumptions, and variances that could not be accounted for with traditional approaches (e.g., differences in organisational culture within the
167 | Page Daniel J. Hoyle RACFs, the beliefs/practices of the prescribers) led to the development of the statistical models described within this sub-chapter.
Initially, we considered characterising residents as antipsychotic and/or benzodiazepine reducers and non-reducers based on changes in their chlorpromazine or diazepam daily dose equivalents. We then compared changes in the clinical outcome measures between the reducer and non-reducer groups using independent t-tests or Mann-Whitney U tests. Older studies investigating clinical outcomes associated with interventions to reduce antipsychotic use had typically utilised these analyses.522, 524 However, being unable to control for clinical differences at baseline in our study meant that it would not have been appropriate to compare the reducer and non-reducer groups. Furthermore, this type of analysis did not account for the initial antipsychotic and/or benzodiazepine dose or consider the magnitude of dose change.
Next, multiple linear regression was considered to control for resident differences at baseline between the reducer and non-reducer groups. This initially required logistic regression modelling to evaluate potential covariates. The advantage of logistic regression was that it does not require residuals to be normally distributed, thus resolving issues with the distribution of our data. However, valid inference from logistic regression requires large sample sizes, and perhaps more importantly, classifying residents as reducers or non-reducers based on prescribing changes over the intervention ignores model uncertainty including variability in prescribing at both time points (baseline and four months post-intervention) and the magnitude in any reduction (or increase) in prescribing. Thus, with advice from a statistician, the decision was made to use linear mixed effects modelling with transformations, treating antipsychotic and benzodiazepine doses at both time points as continuous variables.
6.2.2.2 Covariate selection
Initially, baseline variables that may contribute to antipsychotic and/or benzodiazepine prescribing, based on the literature, were selected as fixed effects (Table 28).299, 492, 524, 607, 608
168 | Page Daniel J. Hoyle Additionally, the RACF and resident identification numbers were included as random effects to account for unknown factors (e.g., the influence of organisational culture, prescriber attitudes).
Broadly, fixed effects are variables that are expected to influence the response variable (in this case antipsychotic or benzodiazepine prescribing). Random effects are grouping factors that we are trying to control so that we can detect the true relationship between the fixed effects and the response variable.609 For example, whilst RACFs may have an effect on antipsychotic prescribing, we wanted to control for this effect so we could determine whether the fixed effects had an impact on prescribing.
Table 28: Baseline resident characteristics that may impact antipsychotic and/or benzodiazepine use.299, 492, 524, 607, 608
Antipsychotic use Benzodiazepine use Age of the resident Age of the resident Diazepam daily dose equivalent Chlorpromazine daily dose equivalent CMAI factor 1 (aggressive behaviour) CMAI factor 1 (aggressive behaviour) CMAI factor 2 (physically nonaggressive CMAI factor 2 (physically nonaggressive behaviour) behaviour) CMAI factor 3 (verbally agitated behaviour) CMAI factor 3 (verbally agitated behaviour) Documented diagnosis of agitation Documented diagnosis of agitation Documented diagnosis of anxiety Documented diagnosis of anxiety Documented diagnosis of dementia Documented diagnosis of dementia Documented diagnosis of psychosis Documented diagnosis of psychosis Documented diagnosis of sleep disorders Documented diagnosis of sleep disorders Gender Gender Length of time admitted to RACF Length of time admitted to RACF NPI-NH factor 1 (agitation) NPI-NH factor 1 (agitation) NPI-NH factor 2 (mood) NPI-NH factor 2 (mood) NPI-NH factor 3 (psychosis) NPI-NH factor 3 (psychosis) NPI-NH factor 4 (sleep) NPI-NH factor 4 (sleep) NPI-NH factor 5 (elated behaviour) NPI-NH factor 5 (elated behaviour) Total NPI-NH score Total NPI-NH score CMAI=Cohen-Mansfield Agitation Inventory, NPI-NH=Neuropsychiatric Inventory-Nursing Home version
Pair plots and consideration of domain-specific knowledge, for example, the similarities in the NPI-NH and CMAI scales, were used to assess multicollinearity among covariates. The variables ‘NPI-NH total score’, ‘NPI-NH factor 1 (agitation)’, ‘NPI-NH factor 2 (mood)’, ‘NPI-NH factor 3 (psychosis)’, ‘NPI-NH factor 4 (sleep)’, and ‘NPI-NH factor 5 (elated behaviour)’ were removed due to multicollinearity. Multicollinearity is where one or more independent variables
(e.g., NPI-NH total and CMAI factor 3 score) are highly correlated.610
169 | Page Daniel J. Hoyle A scatterplot was used to assess the degree of correlation between changes in the prescribing of antipsychotic and benzodiazepine doses over four months of RedUSe (Figure 23).
There was no evidence that changes in prescribing of antipsychotics had any correlation
(positive or negative) with changes in prescribing of benzodiazepines at the resident level.
Figure 23: Change in the chlorpromazine and diazepam daily dose equivalent between baseline and four months.
Linear mixed effects models were fitted with the ‘lme4’611 package in R.594 Inspection of
Q-Q plots revealed large violations of the assumption of normality. Doses were shifted by
0.5 mg and then loge transformed to improve the normality of residuals. In order to account for non-independence of residuals, a random intercept was fitted for the resident identification number and a random intercept and slope for the RACF identification number. Backwards stepwise selection with α=0.10 provided evidence for retaining variables or not.
Interactions between all variables and the intervention were tested in order to investigate the possibility that some variables might have had differential time-dependent effects due to the effect of the RedUSe project.
170 | Page Daniel J. Hoyle 6.2.2.3 Parameter estimation
Due to violation of Gaussian assumptions and small sample sizes, Markov Chain Monte Carlo
(MCMC) techniques were used to estimate model parameters as posterior distributions using the Stan No-U-Turns sampler612 via the ‘brms’ R package.613 Trace plots and R-hat statistics were used to assess chain convergence, and posterior predictive checks and leave-one-out cross- validation were used to assess model fit.614, 615 Variability not explained by fixed effects (intra- class correlation, ICC) was computed using draws from the posterior predictive distribution with the ‘icc’ function in the ‘sjstats’ library.616 We report 95% Highest Density Intervals (HDI).
Dosages were standardised to z-scores and can be interpreted in units of standard deviations.
An advantage of this approach was that rather than estimating the likelihood of the data given the parameters, it allows us to estimate a joint posterior density for the parameters given the data – which was more aligned with the question of interest.
6.2.3 Results
6.2.3.1 Antipsychotic users
There was evidence (using backwards stepwise regression with the criteria α=0.10) that the baseline variables ‘CMAI factor 3 (verbally agitated behaviour)’ and ‘CMAI factor 2 (physically nonaggressive behaviour)’ partially explained the variation in antipsychotic doses at baseline and four months, as well as the change in antipsychotic dose. Considering this, these covariates and their interaction with the change in antipsychotic dose were retained as covariates in models investigating the clinical impact that antipsychotic reduction had within RedUSe (Table
29). CMAI factor 3 (verbally agitated behaviour) included questions rating the frequency of the following symptoms: complaining, constant requests for attention, negativism, repetitious sentences or questions, and screaming.31, 98 CMAI factor 2 (physically nonaggressive behaviour) included questions rating the frequency of the following symptoms: pacing, inappropriate
171 | Page Daniel J. Hoyle robing/disrobing, trying to get to a different place, handling things inappropriately, general restlessness and repetitious mannerisms.31, 98, 563
Overall, residents with a higher baseline CMAI factor 3 score were more likely to be taking higher doses of antipsychotics at baseline and four months (by 0.04 standard deviations per unit increase in CMAI factor 3). Furthermore, each unit increase in CMAI factor 3 was associated with an increase in antipsychotic dose by 0.01 standard deviations between baseline and four months (Figure 24). Residents with a high baseline CMAI factor 2 score were more likely to take lower doses of antipsychotics (by 0.02 standard deviations per unit increase in
CMAI factor 2). Reduction in antipsychotic dose during RedUSe was 0.02 standard deviations greater per unit increase in the CMAI factor score at baseline (Figure 25). For Figure 24-Figure
27, if the lines representing baseline (green) and four months (purple) are parallel, the dose is unlikely to change between baseline and four months based on the investigated characteristic.
Convergence or divergence of these lines represents possible dose changes related to the studied characteristic.
The model was improved by the inclusion of a random slope and intercept for the RACF, which indicated that some of the changes in prescribing could be attributed to the RACF environment (e.g., influence of organisational culture or prescriber), independent of resident characteristics. The total variance explained by the model was 77.8%, with 30% (95%HDI 0.00,
0.67) of the variance not explained by the fixed effects and the sample mean being explained by the RACF (Table 29).
172 | Page Daniel J. Hoyle Table 29: Estimated standardised changes in prescribed antipsychotic dosages (in standard deviation units) and 95% highest density intervals (HDI).
Estimate Lower 95% HDI Upper 95% HDI Baseline (intercept) -0.05 -0.55 0.43 Four months post-RedUSe 0.02 -0.19 0.24 CMAI factor 3 0.04 -0.01 0.08 CMAI factor 3*intervention 0.01 -0.01 0.03 CMAI factor 2 -0.02 -0.06 0.02 CMAI factor 2*intervention -0.02 -0.04 0.00 LOO-adjusted R2 0.778 ICC - RACF 0.30 0.00 0.67 ICC - Resident 0.58 0.25 0.90 LOO=Leave-one-out cross-validation, ICC=Intra-class correlation, RedUSe=Reducing Use of Sedatives, CMAI=Cohen-Mansfield Agitation Inventory, RACF=Residential Aged Care Facility
Figure 24: Scatterplot of the chlorpromazine daily dose equivalent at baseline (green) and four months (purple) versus the baseline Cohen-Mansfield Agitation Inventory (CMAI) factor 3 score. Green and purple lines indicate means. CMAI factor 3 scores range from 5 to 35 (higher scores=more frequent behaviour).
173 | Page Daniel J. Hoyle
Figure 25: Scatterplot of the chlorpromazine daily dose equivalent at baseline (green) and four months (purple) versus the baseline Cohen-Mansfield Agitation Inventory (CMAI) factor 2 score. Green and purple lines indicate means. CMAI factor 2 scores range from 6 to 42 (higher scores=more frequent behaviour).
6.2.3.2 Benzodiazepine users
There was evidence (using backwards stepwise regression with the criteria α=0.10) that the variables ‘Documented diagnosis of dementia’ and baseline ‘CMAI factor 2 (physically nonaggressive behaviour)’ partially explained the variation in benzodiazepine doses at baseline and four months. There was also an interaction between the intervention (RedUSe) and the variables ‘Documented diagnosis of dementia’ and ‘Chlorpromazine daily dose equivalent’, which indicates that these variables partially explained the variation in change in benzodiazepine doses between baseline and four months. Whilst the ‘Chlorpromazine daily dose equivalent*intervention’ interaction met the criteria for the backwards stepwise regression (α=0.10), the Bayesian estimate for the effect on change in benzodiazepine dose was zero. This indicates that the effect of this interaction was so small it was unlikely to be clinically or practically meaningful. Additionally, the main effect of the variable ‘Chlorpromazine daily dose equivalent’ on benzodiazepine prescribing at baseline and four months did not meet the cut-off criteria in the backwards stepwise regression. Considering that the ‘Chlorpromazine daily dose equivalent*intervention’ interaction does not seem to be clinically or practically
174 | Page Daniel J. Hoyle meaningful and the ‘Chlorpromazine daily dose equivalent’ variable did not meet the cut-off criteria in the backwards stepwise regression, these covariates were not included in the assessment of clinical outcomes associated with RedUSe. The other covariates had an effect, as per the Bayesian estimate, thus, were subsequently retained in models investigating changes in clinical outcomes associated with benzodiazepine dose change (Table 30).
Overall, residents with a documented diagnosis of dementia took lower doses of benzodiazepines at baseline and four months (by 0.26 standard deviations) (Figure 26).
Furthermore, an interaction between a documented diagnosis of dementia and the intervention was found suggesting that the reduction in benzodiazepine dose between baseline and four months was 0.05 standard deviations greater in those with a documented diagnosis of dementia.
Residents with a higher baseline CMAI factor 2 score were more likely to be taking lower doses of benzodiazepine at baseline and four months (by 0.02 standard deviations per unit increase in the CMAI factor 2 score) (Figure 27). Whilst the lines in Figure 27 seem to diverge
(representing a decrease in benzodiazepine dose between baseline and four months associated with greater CMAI factor 2 scores), the interaction ‘CMAI factor 2*intervention’ did not reach the cut-off criteria (α=0.10) for inclusion in the model.
The model was improved by the inclusion of a random slope and intercept for the RACF, which indicated that some of the changes in prescribing could be attributed to the RACF environment (e.g., influence of organisational culture or prescriber), independent of resident characteristics. The total variance explained by the model was 70.6%, with 6% (95%HDI 0.00,
0.18) of the variance not explained by the fixed effects and the sample mean being explained by the RACF.
175 | Page Daniel J. Hoyle Table 30: Estimated standardised changes in prescribed benzodiazepine dosages (in standard deviation units) and 95% highest density intervals (HDI).
Estimate Lower 95% HDI Upper 95% HDI Baseline (intercept) 0.49 0.14 0.85 Four months post-RedUSe -0.17 -0.34 -0.01 Documented diagnosis of dementia -0.26 -0.62 0.10 Intervention*documented diagnosis of -0.05 -0.28 0.17 dementia CMAI factor 2 -0.02 -0.05 0.00 Chlorpromazine daily dose equivalent -0.00 -0.01 0.01 Intervention*chlorpromazine daily dose 0.00 -0.00 0.01 equivalent LOO-adjusted R2 0.706 ICC - RACF 0.06 0.00 0.18 ICC - resident 0.75 0.62 0.85 LOO=Leave-one-out cross-validation, ICC=Intra-class correlation, RedUSe=Reducing Use of Sedatives, CMAI=Cohen-Mansfield Agitation Inventory, RACF=Residential Aged Care Facility
Figure 26: Beeswarm plot of the diazepam daily dose (mg) at baseline and four months based on whether the resident had a documented diagnosis of dementia (red) or not (black). Red and black crosses indicate means.
176 | Page Daniel J. Hoyle
Figure 27: Scatterplot of the diazepam daily dose equivalent at baseline (green) and four months (purple) versus the baseline Cohen-Mansfield Agitation Inventory (CMAI) factor 2 score. Green and purple lines indicate means. CMAI factor 2 scores range from 6 to 42 (higher scores=more frequent behaviour). 6.2.4 Discussion
The intention of the current analysis was to detect differences in resident characteristics related to changes in antipsychotic and benzodiazepine dosages between baseline and four months using linear mixed effects modelling. The purpose of this was to identify potentially confounding variables to control for when investigating associations between changes in clinical outcomes and antipsychotic and/or benzodiazepine dose changes in subsequent chapters. Whilst we have briefly attempted to rationalise reasons for differences in antipsychotic and benzodiazepine prescribing related to these characteristics, it was beyond the scope of this thesis to investigate, in detail, the reasons for these differences.
For antipsychotic users, more frequent verbally agitated behaviour was related to higher antipsychotic doses at baseline and four months, and an increase in antipsychotic dose between these time points. On the other hand, more frequent physically nonaggressive behaviour was associated with the use of lower antipsychotic dosages at baseline and four months and a greater reduction in antipsychotic dose between these time points.
177 | Page Daniel J. Hoyle For benzodiazepine users, having a documented diagnosis of dementia was associated with the use of lower benzodiazepine dosages at baseline and four months and a greater reduction in benzodiazepine doses between these time points. Frequent physically nonaggressive behaviour was associated with the use of lower benzodiazepine dosages at baseline and four months but the effect on the change in benzodiazepine dose did not reach the cut-off criteria (α=0.10) for inclusion in the model. Whilst the interaction between the baseline chlorpromazine daily dose equivalent and the intervention met the cut-off criteria
(α=0.10) for inclusion in the model, Bayesian estimates were too small to indicate a practically or clinically meaningful effect on change in benzodiazepine dosages.
Verbal behaviour has been shown to cause concern, frustration, anxiety, anger and/or complaints from RACF staff, visitors and other residents.98 Whilst there is limited evidence for the effectiveness of antipsychotics in the management of verbally disruptive behaviours,98 a
2015 US study found verbal behaviour to be one of the most common reasons for the use of antipsychotic medication among 204 residents from 26 RACFs.617 Considering that verbally disruptive behaviours may be distressing for RACF staff, visitors and other residents, prescribers may have been less inclined to reduce the antipsychotic dose in residents showing frequent verbally disruptive behaviour. Furthermore, higher doses of antipsychotics may have been required to sedate and, thereby, manage residents with this type of disruptive behaviour.
On the other hand, more frequent physically nonaggressive behaviours were associated with the use of lower doses of antipsychotics and benzodiazepines. Furthermore, residents with more frequent symptoms were more likely to have their antipsychotic dose reduced. Current clinical guidelines do not recommend the use of antipsychotics or benzodiazepines for physically nonaggressive behaviour considering their low efficacy and increased risk of adverse events.98, 106 Whilst further research is required to explain why lower doses of antipsychotics and benzodiazepines were used and why there was a greater reduction in antipsychotic doses
178 | Page Daniel J. Hoyle for residents with more frequent physically nonaggressive behaviour, we have provided several theories below that may partially explain these phenomena.
First, lower antipsychotic and benzodiazepine doses may have been used at baseline and four months because these symptoms are relatively less disruptive. A US study of 191 residents from 11 RACFs comparing the disruptiveness of different types of agitation found that physically nonaggressive behaviour was the least disruptive to RACF nursing staff.618
Additionally, the RedUSe project educated staff on inappropriate antipsychotic use, including using antipsychotics for physically nonaggressive symptoms. Considering that these symptoms are relatively less disruptive and antipsychotic use for these symptoms is considered inappropriate, residents displaying physically nonaggressive symptoms may have been considered ‘easier targets’ for antipsychotic reduction. However, this theory does not explain why residents with more frequent symptoms are prescribed lower doses of antipsychotics and benzodiazepines.
Second, considering that neither antipsychotics nor benzodiazepines are recommended for physically nonaggressive symptoms and they may increase the risk of falls, particularly in residents who wander frequently,619 prescribers may have tried to limit and reduce the doses of antipsychotics and benzodiazepines to minimise the risk of falls in those with more frequent physically nonaggressive behaviour.
Lastly, the greater decrease in antipsychotic dose over the study period among those with more frequent physically nonaggressive behaviour may be explained by the concern that more frequent physically nonaggressive behaviour could be a sign of movement disorders (e.g., akathisia) secondary to antipsychotic usage.620 This concern may have been amplified by the educational information delivered as part of the RedUSe intervention.50
Similar to this study, two other studies have reported lower utilisation of benzodiazepines among residents with dementia.297, 621 There are several possible reasons that
179 | Page Daniel J. Hoyle could explain the use of lower doses of benzodiazepines in residents with a documented diagnosis of dementia.
First, benzodiazepines can contribute to cognitive impairment.373, 408, 409 Prescribers may have been cognisant of this effect and tried to use lower doses to minimise the risk of worsening cognition in residents with dementia.
Second, residents with dementia have an elevated risk of falls.622 Benzodiazepine use may increase this risk further, particularly when used at higher doses.267, 422, 623 Therefore, prescribers may have chosen to use lower doses of benzodiazepines in the residents with a documented diagnosis of dementia to minimise the risk of falls.
Third, concomitant use of an antipsychotic may have allowed for lower doses of a benzodiazepine to be used in residents with a documented diagnosis of dementia. In our study, residents with a documented diagnosis of dementia were more likely to be taking a concomitant antipsychotic (30.9%) compared to those without a documented diagnosis of dementia (9.1%).
Benzodiazepines are often utilised to promote sleep and manage anxiety.26, 624 Antipsychotics, particularly quetiapine and olanzapine, also have sleep promoting and anxiolytic properties.625,
626 Consequently, the use of an antipsychotic may have enabled lower benzodiazepine doses to be used. Additionally, antipsychotics may also increase the risk of adverse effects, such as excessive sedation and falls.205, 422 Considering that use of a benzodiazepine may increase these adverse effects, lower benzodiazepine doses may have been utilised to minimise these risks in those taking a concomitant antipsychotic.
Fourth, it should be noted that dementia is often not formally diagnosed in the RACF setting.80, 627 A recent Australian study reported that up to one-fifth of residents may have dementia despite not being formally diagnosed.627 Furthermore, the diagnosis of dementia is often delayed.628 This may be related to perceptions that there is little value gained by formally diagnosing dementia in someone who is older and living in a RACF.628 It is highly likely that some
180 | Page Daniel J. Hoyle of the residents without a documented diagnosis of dementia that were involved in this study, in fact, had dementia. Given the concern regarding further cognitive impairment with benzodiazepine use, particularly in those with severe impairment, differences in levels of cognitive impairment may explain the variation in the benzodiazepine doses used when comparing those with a documented diagnosis of dementia versus those without. However, considering that we did not evaluate cognitive function in our assessments it was not possible to determine the accuracy of this theory.
Fifth, a 2012 Belgian study found that benzodiazepine use is less likely among people with dementia.297 Whilst the authors acknowledged they had expected to see an increase in use with dementia considering associations with nocturnally disturbed sleep,629 they recognised that residents with severe dementia may have lost the ability to complain about their sleeping problems. Subsequently, benzodiazepine prescriptions for these residents may not have been renewed. It is possible that we have identified a similar situation in our research where those with a documented diagnosis of dementia have more severe dementia and are unable to communicate sleeping problems. Additionally, alteration of benzodiazepine use within RedUSe may have not been objected to by these residents.
Neither having a documented diagnosis of dementia nor psychosis met the cut-off for inclusion in the model for antipsychotic users. It was expected that those with a documented diagnosis of dementia would be taking lower doses of antipsychotics considering that guidelines recommend lower maximum doses (e.g., risperidone is recommended at a maximum of 2 mg per day when used for neuropsychiatric symptoms in people with dementia).28, 31 Considering that this sub-study excluded residents with a severe psychiatric condition (e.g., bipolar disorder, schizophrenia), almost all (n=68, 82%) antipsychotic users had a documented diagnosis of dementia. The indication for antipsychotic use in the remaining residents (n=15, 18%) is unclear.
Given that dementia is often not formally diagnosed in the RACF setting, there is potential that
181 | Page Daniel J. Hoyle these residents may also have dementia which may explain why no significant differences in antipsychotic doses were detected.41, 627 Additionally, it was expected those with a documented diagnosis of psychosis may be taking higher doses of antipsychotics. However, with very few residents having a documented diagnosis of psychosis (n=7, 8%), this analysis may not have had sufficient power to detect a difference.
In addition to the tested resident characteristics (fixed effects), unknown RACF and resident characteristics (random effects) accounted for a substantial amount of variation in antipsychotic and benzodiazepine dosages. This suggests that unknown differences in the RACFs and residents (e.g., differences in organisational culture, prescriber attitudes) played a major role in the prescribing of these medications. A study on the variation in antipsychotic prescribing in US RACFs found that resident characteristics, such as age, diagnosis of dementia or comorbidities, accounted for only 36% of prescribing, which suggests that the majority of the prescribing was caused by unknown influences relating to the RACF.489 In our study, our models for antipsychotic users and benzodiazepine users explained 77.8% and 70.6% of the variance in the data, respectively. For antipsychotic users, unknown RACF characteristics explained 30% of variance not explained by fixed effects or sample mean. For benzodiazepine users, unknown
RACF characteristics explained 6% of variance not explained by fixed effects or sample mean.
Ideally, variance in psychotropic prescribing should be almost completely accounted for by individual clinical factors (e.g., severity of symptoms, comorbidities). However, the substantial proportion of variance in psychotropic prescribing accounted for by differences between RACFs is suggestive of inappropriate prescribing.43
Unknown RACF characteristics explained a larger proportion of variance in antipsychotic prescribing than benzodiazepine prescribing. This suggests that RACF characteristics, such as organisational culture and prescriber attitudes, have a greater influence on antipsychotic prescribing. Whilst previous research has investigated the effect of
182 | Page Daniel J. Hoyle organisational culture on overall psychotropic prescribing,43, 490, 630 further studies are required to compare the magnitude of effect that organisational culture has on different classes of psychotropic medication.
6.2.5 Implications for clinical outcome assessments
The fixed effects (resident characteristics) accounting for variability in antipsychotic and benzodiazepine prescribing may impact the clinical outcomes of residents when reducing the dose of these medications. Therefore, the fixed effects identified in this chapter were controlled, where appropriate, using multiple linear regression models when assessing changes in the clinical outcomes for the residents.
For antipsychotic users, the fixed effects included the baseline variables ‘CMAI factor 3
(verbally agitated behaviour)’ and ‘CMAI factor 2 (physically nonaggressive behaviour)’. For benzodiazepine users, the fixed effects included the variables ‘documented diagnosis of dementia’ and baseline ‘CMAI factor 2 (physically nonaggressive behaviour)’.
183 | Page Daniel J. Hoyle 6.3 Neuropsychiatric symptoms
6.3.1 Introduction
Despite evidence that neuropsychiatric symptoms are unlikely to worsen upon antipsychotic and benzodiazepine dose reduction for the majority of residents (see Chapter 3 and Chapter
4),47, 104 a common barrier to dose reduction among nursing staff and prescribers is the perception that the original symptoms may return or worsen (discussed in Chapter 3).44, 45
To address this concern, Chapter 6.3 of this thesis aimed to investigate the association between antipsychotic and benzodiazepine dose reduction and changes in neuropsychiatric symptoms in a sample of residents in RedUSe. Figure 28 illustrates where these findings fit in this overall thesis.
Figure 28: Flowchart of the analysis for the clinical outcomes study. This sub-chapter is dedicated to investigating the association between antipsychotic and benzodiazepine dose reduction and change in neuropsychiatric symptoms. GP=General Practitioner. 6.3.2 Methods
6.3.2.1 Outcome measures
Neuropsychiatric symptoms were assessed at baseline and four months, through psychometric testing of RACF nursing and care staff, using the NPI-NH (Chapter 5.7.1.4.1) and the CMAI
184 | Page Daniel J. Hoyle (Chapter 5.7.1.4.2) psychometric tools. Total and factor scores were calculated for the NPI-NH and CMAI (method of calculation outlined in Chapter 5.7.1.4.1 and Chapter 5.7.1.4.2, respectively). Greater NPI-NH scores indicated more frequent and/or severe neuropsychiatric symptoms. Greater CMAI scores indicated more frequent agitated/aggressive symptoms.
6.3.2.2 Statistical analysis
Given RedUSe aimed to reduce inappropriate prescribing of antipsychotic and benzodiazepine medication, two separate sets of analyses were run; one for antipsychotic users (those with a chlorpromazine daily dose equivalent >0 mg at baseline) and one for benzodiazepine users
(those with a diazepam daily dose equivalent >0 mg at baseline). The distribution of the NPI-NH and CMAI scores were initially investigated using beeswarm plots (scatterplots with would-be overlapping data points separated so each is visible) and a paired t-test was utilised to detect statistically significant changes in scores between baseline and four months.
To investigate the relationship between the magnitude of antipsychotic and benzodiazepine dose changes and changes in the NPI-NH and CMAI total and factor scores, linear regression models were fitted with the ‘lm’ function in R.594 Potentially confounding variables, identified in Chapter 6.2, were controlled for within the models, where appropriate.
In the antipsychotic users’ model, it was not considered appropriate to control for any of the identified potentially confounding variables considering multicollinearity between the NPI-NH and CMAI total and factor scores. In the benzodiazepine users’ model, the potentially confounding variable ‘documented diagnosis of dementia’ was included. However, the baseline
CMAI factor 2 score was not included, considering multicollinearity between the NPI-NH and
CMAI total and factor scores. From the output of these models, we were able to estimate the unit change in the NPI-NH and CMAI total and factor scores associated with the percentage change in the antipsychotic and benzodiazepine doses between baseline and four months.
185 | Page Daniel J. Hoyle Inspection of Q-Q plots of the residuals revealed approximately normal distributions of the residuals with small violations of this assumption in the tails. To account for this violation, confidence intervals were bootstrapped. Plots of the NPI-NH and CMAI total and factor scores, and antipsychotic and benzodiazepine doses at baseline and four months were inspected for non-linear distribution of the data.
The change in the NPI-NH and CMAI total score was assessed for residents who completely ceased their antipsychotic or benzodiazepine between baseline and four months, using paired t-tests.
Lastly, considering that a variety of other medications can have positive and negative effects on neuropsychiatric symptoms, Wilcoxon Signed Rank tests were used to investigate changes in the number of cholinesterase inhibitors, memantine, anticonvulsants, adrenergic agents, non-benzodiazepine anxiolytics, z-drugs, antidepressants, opioids, non-benzodiazepine sedatives, antihistamines and anticholinergics used between baseline and four months (ATC codes used for each medication class listed in Appendix J). Additionally, changes in the proportion of residents taking these medications at baseline and four months were tested using
χ2 tests (or Fisher’s exact tests where the expected frequencies were less than five).631
6.3.3 Results
6.3.3.1 Feasibility of measuring neuropsychiatric symptoms associated with antipsychotic and/or benzodiazepine dose change
Overall, it appeared feasible to measure changes in neuropsychiatric symptoms associated with antipsychotic and/or benzodiazepine dose change using the NPI-NH and CMAI; data were available for all participants who were present in the study at baseline and four months.
186 | Page Daniel J. Hoyle 6.3.3.2 Antipsychotic users
6.3.3.2.1 Neuropsychiatric Inventory- Nursing Home version
A total of 83 residents initially using an antipsychotic were assessed with the NPI-NH psychometric tool by a RACF staff member at baseline and four months. Approximately 36% of these residents (30/83) had their antipsychotic dose reduced or ceased, 5% (4/83) had their dose increased and 59% (49/83) had no change in their dose. Whilst small improvements were seen in the NPI-NH total and factor scores between baseline and four months, paired t-tests indicated that the results were not statistically significant (Table 31, Figure 29).
Table 31: Mean Neuropsychiatric Inventory-Nursing Home version total and factor scores at baseline and four months.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Total 26.2 (20.4, 31.9) 24.0 (19.4, 28.6) 0.82 82 0.413 Factor 1 7.1 (5.4, 8.7) 6.2 (4.7, 7.7) 1.15 82 0.253 Factor 2 8.3 (6.5, 10.1) 8.0 (6.5, 9.5) 0.30 82 0.765 Factor 3 3.3 (2.0, 4.6) 2.9 (1.8, 3.9) 0.75 82 0.458 Factor 4 4.4 (3.1, 5.6) 3.5 (2.4, 4.6) 1.28 82 0.206 Factor 5 2.0 (1.2, 2.9) 1.9 (1.1, 2.7) 0.29 82 0.773 df=degrees of freedom, 95%CI=95% confidence interval
187 | Page Daniel J. Hoyle
Figure 29: Beeswarm plot of the NPI-NH (A) total, (B) factor 1, (C) factor 2, (D) factor 3, (E) factor 4 and (F) factor 5 scores at baseline and four months for all residents taking an antipsychotic at baseline. The blue dot represents mean and the error bars indicate the 95% confidence interval. NPI-NH=Neuropsychiatric Inventory-Nursing Home version.
Using linear regression, we found a slight, albeit non-significant, decrease in overall neuropsychiatric symptoms (measured with the NPI-NH total score) associated with the magnitude of antipsychotic dose reduction (β=0.013; i.e., total NPI-NH score was, on average,
0.13 points lower for each 10% reduction in the chlorpromazine daily dose equivalent, p=0.782,
95%CI=-0.078, 0.104; Table 32 and Figure 30). Figure 30 illustrated a tendency towards improvements in agitation (factor 1), mood (factor 2), psychosis (factor 3), and sleep/motor activity (factor 4) associated with antipsychotic dose reduction. However, estimates from the linear regression models (Table 32) suggested that these effects were relatively weak and not
188 | Page Daniel J. Hoyle statistically significant. It should be acknowledged that whilst trendlines have been added to the scatterplots to visualise the change in outcome with the change in dose, the line representing the change in the outcome with an increase in the dose is influenced by very few data points (i.e., few residents who had their dose increased).
There was a non-significant 2.5-point decrease in the NPI-NH total score between baseline (mean=25.1, 95%CI=13.9, 36.3) and four months (mean=22.6, 95%CI=12.6, 32.6) in those 18 residents who completely ceased their antipsychotic (t(17)=0.394, p=0.699).
Table 32: Estimated changes in the NPI-NH total and factor scores associated with changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value Intercept -1.912 -7.557, 3.733 -0.674 0.502 Antipsychotic dose change (%) 0.013 -0.078, 0.104 0.278 0.782 2 NH total
- Adjusted R -0.011 2
NPI Multiple R 0.001 Intercept -0.648 -2.269, 0.973 -0.796 0.429 Antipsychotic dose change (%) 0.011 -0.015, 0.037 0.835 0.406
1 2
NH factor NH factor Adjusted R -0.004 - 2
NPI Multiple R 0.009 Intercept -0.213 -2.178, 1.753 -0.215 0.830 Antipsychotic dose change (%) 0.003 -0.029, 0.035 0.192 0.848
2 2
NH factor NH factor Adjusted R -0.012 - 2
NPI Multiple R 0.001 Intercept -0.388 -1.728, 0.951 -0.577 0.565 Antipsychotic dose change (%) 0.004 -0.018, 0.025 0.356 0.723
3 2
NH factor NH factor Adjusted R -0.011 - 2
NPI Multiple R 0.002 Intercept -0.843 -2.270, 0.584 -1.175 0.243
4 Antipsychotic dose change (%) 0.001 -0.022, 0.024 0.050 0.960 NH - Adjusted R2 -0.012 NPI factor Multiple R2 0.000 Intercept -0.253 -1.224, 0.717 -0.519 0.605 Antipsychotic dose change (%) -0.006 -0.021, 0.010 -0.726 0.470
5 2
NH factor NH factor Adjusted R -0.006 - 2
NPI Multiple R 0.007 NPI-NH=Neuropsychiatric Inventory-Nursing Home version, 95%CI=95% confidence interval
189 | Page Daniel J. Hoyle
Figure 30: Scatterplots showing the relationship between change (%) in the chlorpromazine daily dose equivalent and changes in the NPI-NH (A) total, (B) factor 1, (C) factor 2, (D) factor 3, (E) factor 4 and (F) factor 5 scores. NPI- NH=Neuropsychiatric Inventory-Nursing Home version, CPZ=chlorpromazine, DDE=daily dose equivalent.
190 | Page Daniel J. Hoyle 6.3.3.2.2 Cohen-Mansfield Agitation Inventory
A total of 83 residents initially using an antipsychotic were assessed with the CMAI psychometric tool at baseline and four months. Overall, the CMAI total and factor 2 (physically nonaggressive behaviour) and 3 (verbally agitated behaviour) scores increased (higher scores equal more frequent agitation), albeit not by a statistically or clinically significant amount, 3 between baseline and four months. On the other hand, factor 1 (physically aggressive behaviour) remained relatively stable between baseline and four months (Table 33, Figure 31).
Table 33: Mean Cohen-Mansfield Agitation Inventory total and factor scores at baseline and four months for residents using an antipsychotic at baseline.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Total 55.9 (49.3, 62.5) 59.3 (53.7, 65.0) -1.05 82 0.318 Factor 1 16.1 (13.8, 18.4) 15.7 (13.7, 17.7) 0.37 82 0.713 Factor 2 13.7 (11.7, 15.7) 15.5 (13.7, 17.2) -1.60 82 0.113 Factor 3 12.2 (10.5, 14.0) 14.1 (12.4, 15.8) -1.77 82 0.081 df=degrees of freedom, 95%CI=95% confidence interval
3 Clinically significant changes in the CMAI total and factor scores are defined as: Change in total≥8 Change in factor 1≥3 Change in factor 2≥6 Change in factor 3≥4
191 | Page Daniel J. Hoyle
Figure 31: Beeswarm plot of the CMAI (A) total, (B) factor 1, (C) factor 2 and (D) factor 3 scores at baseline and four months for all residents taking an antipsychotic at baseline. The blue dot represents mean and the error bars indicate the 95% confidence interval. CMAI=Cohen-Mansfield Agitation Inventory.
Despite the overall tendency for agitation/aggression to slightly increase among residents initially taking an antipsychotic, we found no evidence that the magnitude of antipsychotic dose reduction was associated with an increase in the frequency of symptoms
(Figure 32). Instead, antipsychotic dose reduction was associated, non-significantly, with a small reduction in the frequency of agitated/aggressive symptoms. In particular, the strongest association was between the magnitude of antipsychotic dose reduction and improvements in physically nonaggressive symptoms (β=0.037; i.e., CMAI factor 2 score was 0.37 points lower for each 10% reduction in the chlorpromazine daily dose equivalent, p=0.052, 95%CI=-0.000,
0.075; Table 34).
192 | Page Daniel J. Hoyle There was a non-significant 2.5-point decrease in the CMAI total score between baseline (mean=53.2, 95%CI=38.6, 67.9) and four months (mean=50.7, 95%CI=41.8, 59.5) in the
18 residents who had their antipsychotic completely ceased (t(17)=0.328, p=0.747).
Table 34: Estimated changes in the CMAI total and factor scores associated with changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value
Intercept 4.975 -2.205, 12.156 1.379 0.172 Antipsychotic dose change 0.073 -0.042, 0.189 1.264 0.210 (%) Adjusted R2 0.007 CMAI totalCMAI Multiple R2 0.019 Intercept -0.082 -2.437, 2.274 -0.069 0.945 Antipsychotic dose change 0.016 -0.022, 0.053 0.814 0.418 (%) Adjusted R2 -0.004 2 CMAI factorCMAI 1 Multiple R 0.008 Intercept 2.581 0.252, 4.911 2.205 0.030 Antipsychotic dose change 0.037 -0.000, 0.075 1.972 0.052 (%) Adjusted R2 0.034 2 CMAI factorCMAI 2 Multiple R 0.046 Intercept 2.193 -0.028, 4.415 1.964 0.053 Antipsychotic dose change 0.016 -0.020, 0.052 0.889 0.377 (%) Adjusted R2 -0.003 2 CMAI factorCMAI 3 Multiple R 0.010 CMAI=Cohen-Mansfield Agitation Inventory, 95%CI=95% confidence interval
193 | Page Daniel J. Hoyle
Figure 32: Scatterplots showing the relationship between change (%) in the chlorpromazine daily dose equivalent and changes in the CMAI (A) total, (B) factor 1, (C) factor 2 and (D) factor 3 scores. CMAI=Cohen-Mansfield Agitation Inventory, CPZ=chlorpromazine, DDE=daily dose equivalent.
6.3.3.3 Benzodiazepine users
6.3.3.3.1 Neuropsychiatric Inventory- Nursing Home version
One hundred and eighteen residents initially using a benzodiazepine were assessed with the
NPI-NH psychometric tool at baseline and four months. Approximately 35% (42/118) of these residents had their benzodiazepine dose reduced or ceased, 6% (7/118) had their dose increased and 59% (69/118) had no change in their dose. Overall, there were small, albeit non- significant, improvements in the NPI-NH total score and four of five of the factor scores (Table
35, Figure 33). Factor 5 remained constant between baseline and four months.
194 | Page Daniel J. Hoyle Table 35: Mean Neuropsychiatric Inventory-Nursing Home version total and factor scores at baseline and four months.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Total 18.1 (14.3, 22.0) 16.2 (12.8, 19.6) 1.05 117 0.296 Factor 1 5.0 (3.9, 6.0) 4.2 (3.2, 5.2) 1.38 117 0.170 Factor 2 6.3 (4.9, 7.7) 5.4 (4.2, 6.6) 1.20 117 0.231 Factor 3 1.8 (1.0, 2.5) 1.2 (0.7, 1.7) 1.41 117 0.160 Factor 4 2.9 (2.1, 3.6) 2.7 (1.9, 3.6) 0.26 117 0.797 Factor 5 1.5 (0.9, 2.1) 1.5 (0.9, 2.1) -0.19 117 0.849 df=degrees of freedom, 95%CI=95% confidence interval
Figure 33: Beeswarm plot of the NPI-NH (A) total, (B) factor 1, (C) factor 2, (D) factor 3, (E) factor 4 and (F) factor 5 scores at baseline and four months for all residents taking a benzodiazepine at baseline. The blue dot represents mean and the error bars indicate the 95% confidence interval. NPI-NH=Neuropsychiatric Inventory-Nursing Home version.
Small improvements in the NPI-NH total and factor scores were seen in proportion to the magnitude of benzodiazepine dose reduction (Figure 34). Using multiple linear regression models to control for a documented diagnosis of dementia, small, albeit non-significant,
195 | Page Daniel J. Hoyle improvements in the NPI-NH total and factor scores were seen with benzodiazepine dose reduction (Table 36). Our estimates indicated that a 10% reduction in the benzodiazepine dose was, on average, associated with a 0.38-point (95%CI=-0.14, 0.90) improvement in the NPI-NH total score (p=0.153). Furthermore, a 10% reduction in benzodiazepine dose between baseline and four months was associated, non-significantly, with a 0.09-point (95%CI=-0.04, 0.21) and
0.10-point (95%CI=-0.12, 0.31) decrease in NPI-NH factor scores related to sleep (factor 4) and mood (factor 2), respectively.
Residents who had their benzodiazepine dose ceased (n=25) had, on average, a non- significant 6.5-point improvement in the NPI-NH total score between baseline (mean=20.9,
95%CI=9.9, 31.8) and four months (mean=14.4, 95%CI=7.8, 20.9; t(24)=1.355, p=0.188).
Table 36: Estimated changes in the NPI-NH total and factor scores associated with changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between baseline and four months while controlling for a documented diagnosis of dementia.
Outcome Estimate (β) 95%CI t-statistic p-value Intercept -0.076 -4.917, 4.765 -0.031 0.975 Benzodiazepine dose change (%) 0.038 -0.014, 0.090 1.439 0.153 Documented diagnosis of -2.632 -9.831, 4.567 -0.724 0.470 NH total
- dementia Adjusted R2 0.006 NPI Multiple R2 0.023
Intercept -0.219 -1.656, 1.219 -0.301 0.764 Benzodiazepine dose change (%) 0.007 -0.009, 0.022 0.840 0.403 Documented diagnosis of -0.917 -3.055, 1.221 -0.850 0.397 dementia NH factor 1 NH factor - Adjusted R2 -0.004 NPI Multiple R2 0.013
Intercept -0.911 -2.904, 1.08 -0.905 0.367 Benzodiazepine dose change (%) 0.010 -0.012, 0.031 0.910 0.365 Documented diagnosis of 0.439 -2.525, 3.403 0.293 0.770 dementia NH factor 2 NH factor - Adjusted R2 -0.010 NPI Multiple R2 0.008
Intercept -0.171 -1.205, 0.864 -0.327 0.744 Benzodiazepine dose change (%) 0.002 -0.009, 0.013 0.308 0.758 Documented diagnosis of -0.775 -2.313, 0.763 -0.998 0.320 dementia NH factor 3 NH factor - Adjusted R2 -0.008 NPI Multiple R2 0.010
196 | Page Daniel J. Hoyle Intercept 0.705 -0.425, 1.834 1.236 0.219 Benzodiazepine dose change (%) 0.009 -0.004, 0.021 1.382 0.170 Documented diagnosis of -1.512 -3.192, 1.67 -1.784 0.077 dementia NH factor 4 NH factor - Adjusted R2 0.027 NPI Multiple R2 0.044
Intercept 0.039 -0.791, 0.868 0.092 0.927 Benzodiazepine dose change (%) 0.008 -0.001, 0.016 1.669 0.098 Documented diagnosis of 0.346 -0.888, 1.580 0.555 0.580 dementia NH factor 5 NH factor - Adjusted R2 0.009 NPI Multiple R2 0.026 NPI-NH=Neuropsychiatric Inventory-Nursing Home version, 95%CI=95% confidence interval
Figure 34: Scatterplots showing the relationship between change (%) in the diazepam daily dose equivalent and changes in the NPI-NH (A) total, (B) factor 1, (C) factor 2, (D) factor 3, (E) factor 4 and (F) factor 5 scores. NPI- NH=Neuropsychiatric Inventory-Nursing Home version, DDE=daily dose equivalent.
197 | Page Daniel J. Hoyle 6.3.3.3.2 Cohen-Mansfield Agitation Inventory
A total of 118 residents initially using a benzodiazepine at baseline were assessed with the CMAI psychometric tool at baseline and four months. There were no appreciable changes in scores between baseline and four months (Table 37).
Table 37: Mean Cohen-Mansfield Agitation Inventory total and factor scores at baseline and four months for residents using a benzodiazepine at baseline.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Total 46.9 (43.1, 50.6) 47.4 (43.8, 51.0) -0.27 117 0.791 Factor 1 12.1 (11.0, 13.3) 11.6 (10.7, 12.5) 0.94 117 0.349 Factor 2 11.2 (9.9, 12.4) 11.4 (10.1, 12.6) -0.33 117 0.741 Factor 3 11.3 (10.2, 12.5) 11.9 (10.6, 13.2) -0.89 117 0.375 df=degrees of freedom, 95%CI=95% confidence interval
Figure 35: Beeswarm plot of the CMAI (A) total, (B) factor 1, (C) factor 2 and (D) factor 3 scores at baseline and four months for all residents taking a benzodiazepine at baseline. The blue dot represents mean and the error bars indicate the 95% confidence interval. CMAI=Cohen-Mansfield Agitation Inventory.
198 | Page Daniel J. Hoyle Improvements in the CMAI total and factor scores were associated with the magnitude of benzodiazepine dose reduction (Figure 36). Using multiple linear regression models to control for residents with a documented diagnosis of dementia, trends towards improvements in the
CMAI total and factor scores were detected with greater benzodiazepine dose reduction (Table
38). In particular, a 10% reduction in benzodiazepine dose between baseline and four months was associated with a 0.49-point (95%CI=-0.06, 1.03; p=0.078) decrease in the CMAI total score, a 0.17-point (95%CI=-0.01, 0.34; p=0.062) decrease in the CMAI factor 2 score (physically nonaggressive behaviour), and a 0.16-point (95%CI=-0.03, 0.34; p=0.095) decrease in the CMAI factor 3 score (verbally agitated behaviour).
Among residents who had their benzodiazepine ceased (n=25), there was a non- significant 7.8-point decrease in the mean CMAI total score between baseline (mean=54.2,
95%CI=43.9, 64.5) and four months (mean=46.4, 95%CI=38.7, 54.1) (t(24)=1.546, p=0.135).
Table 38: Estimated changes in the CMAI total and factor scores associated with changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value Intercept 1.434 -3.587, 6.456 0.566 0.573 Benzodiazepine dose change (%) 0.049 -0.006, 0.103 1.779 0.078 Documented diagnosis of -0.194 -7.662, 7.273 -0.051 0.959 dementia 2
CMAI totalCMAI Adjusted R 0.010 Multiple R2 0.027
Intercept 0.036 -1.394, 1.467 0.050 0.960 Benzodiazepine dose change (%) 0.007 -0.008, 0.023 0.947 0.346 Documented diagnosis of -0.924 -3.052, 1.203 -0.861 0.391 dementia Adjusted R2 -0.003 CMAI factorCMAI 1 Multiple R2 0.015
Intercept 0.285 -1.353, 1.923 0.345 0.731 Benzodiazepine dose change (%) 0.017 -0.001, 0.034 1.885 0.062 Documented diagnosis of 0.480 -1.956, 2.917 0.390 0.697 dementia Adjusted R2 0.014 CMAI factorCMAI 2 Multiple R2 0.031
199 | Page Daniel J. Hoyle Intercept 0.437 -1.291, 2.165 0.501 0.617 Benzodiazepine dose change (%) 0.016 -0.003, 0.034 1.683 0.095 Documented diagnosis of 0.943 -1.626, 3.512 0.727 0.469 dementia Adjusted R2 0.011 CMAI factorCMAI 3 Multiple R2 0.028 CMAI=Cohen-Mansfield Agitation Inventory, 95%CI=95% confidence interval
Figure 36: Scatterplots showing the relationship between the change (%) in the diazepam daily dose equivalent and changes in the CMAI (A) total, (B) factor 1, (C) factor 2 and (D) factor 3 scores. CMAI=Cohen-Mansfield Agitation Inventory, DDE=daily dose equivalent.
6.3.3.4 Change in other medications
As indicated in Table 39, there was no appreciable change in the number of medications from other medication classes that may have a beneficial or negative effect on neuropsychiatric symptoms (e.g., cholinesterase inhibitors, memantine, anticonvulsants).
200 | Page Daniel J. Hoyle Table 39: Utilisation of other medication classes that may have beneficial or negative effects on neuropsychiatric symptoms, at baseline and four months.
Medication class/groups Median number Median number (range) p-value (range) of medications of medications used at used at baseline four months Cholinesterase inhibitors 0 (0, 1) 0 (0, 1) 0.346 Memantine 0 (0, 1) 0 (0, 1) NA Anticonvulsants 0 (0, 3) 0 (0, 3) 0.120 Adrenergic agents 0 (0, 1) 0 (0, 1) 0.773 Non-benzodiazepine anxiolytics 0 (0, 0) 0 (0, 0) NA Z-drugs 0 (0, 1) 0 (0, 1) 1.000 Antidepressants 1 (0, 2) 1 (0, 2) 0.802 Tricyclic antidepressants 0 (0, 1) 0 (0, 1) NA Mirtazapine 0 (0, 2) 0 (0, 2) 0.777 Opioids 0 (0, 3) 0 (0, 3) 0.152 Non-benzodiazepine sedatives 0 (0, 1) 0 (0, 1) NA Antihistamines 0 (0, 1) 0 (0, 1) NA Anticholinergics 0 (0, 2) 0 (0, 2) NA
Similarly, there was no appreciable change in the proportion of residents taking these other medication classes between baseline and four months (Table 40).
Table 40: Proportion of residents using other medication classes that may have beneficial or negative effects on neuropsychiatric symptoms, at baseline and four months.
Medication class/groups Proportion (number) Proportion (number) of p-value of residents using the residents using the medication at baseline medication at four months Cholinesterase inhibitors 6.7% (12) 5.6% (10) 0.826 Memantine 1.1% (2) 1.1% (2) 1.000 Anticonvulsants 14.5% (26) 12.3% (22) 0.642 Adrenergic agents 31.8% (57) 31.3% (56) 1.000 Non-benzodiazepine anxiolytics 0% (0) 0% (0) NA Z-drugs 0.6% (1) 0.6% (1) 1.000 Antidepressants 54.2% (97) 54.8% (98) 1.000 Tricyclic antidepressants 5.6% (10) 5.6% (10) 1.000 Mirtazapine 16.8% (30) 18.4% (33) 0.781 Opioids 17.9% (32) 20.1% (36) 0.686 Non-benzodiazepine sedatives 0.6% (1) 0.6% (1) 1.000 Antihistamines 1.7% (3) 1.7% (3) 1.000 Anticholinergics 8.4% (15) 8.4% (15) 1.000
6.3.4 Discussion
The key overall finding is that there was no evidence of worsened neuropsychiatric symptoms with antipsychotic and/or benzodiazepine dose reduction within a sample of residents involved in RedUSe. Moreover, there were trends towards less frequent physically nonaggressive
201 | Page Daniel J. Hoyle behaviour (CMAI factor 2) with antipsychotic and benzodiazepine dose reduction and less frequent verbally agitated behaviour (CMAI factor 3) with benzodiazepine dose reduction.
Generally, the greatest improvements in the NPI-NH and CMAI total and factor scores were seen with antipsychotic and benzodiazepine cessation.
The absence of evidence to suggest worsening in neuropsychiatric symptoms with antipsychotic dose reduction aligns with previous intervention studies that have utilised similar strategies.47, 531 In particular, the possible positive effect of antipsychotic dose reduction on neuropsychiatric symptoms is consistent with a smaller antipsychotic deprescribing intervention performed in Australia over the same timeframe, called the ‘Halting Antipsychotic use in Long Term care’ (HALT) project.530 This multicomponent intervention recruited residents taking antipsychotics from a sample of 23 RACFs (n=133) who consented to have their antipsychotic withdrawn. Residents were excluded from this study if they had a primary psychiatric illness (e.g., schizophrenia, bipolar disorder) or severe neuropsychiatric symptoms
(e.g., a total NPI-NH score ≥50, NPI-NH).530 HALT consisted of an education/training component for health care staff and individualised antipsychotic deprescribing protocols. The intervention resulted in dose reduction for 87.3% of residents at six months post-intervention.530 In comparison, RedUSe was a facility-based intervention where multiple quality improvement strategies were implemented for staff and health practitioners to ensure appropriate psychotropic use, along with a structured interdisciplinary psychotropic review process.50
Residents were not individually recruited to participate. Over 38% of all residents prescribed antipsychotics at baseline in RedUSe (n=835) had their dose reduced or ceased at six months.50
The estimated effect of antipsychotic cessation (i.e., 100% dose reduction) on the NPI-
NH total score in our study (1.3-point improvement compared to no dose change) is similar to that reported by Brodaty et al. (1.0-point improvement).530 However, the estimated impact of antipsychotic cessation on the CMAI total score was greater in our study (7.3-point
202 | Page Daniel J. Hoyle improvement compared to no dose change) than reported in HALT (1.7-point improvement).530
The statistical analysis methodology may account for differences in the CMAI. Whilst HALT compared the same resident before and after, our study compared residents who had completed antipsychotic cessation against those who had no change in their dose. When limiting our study to those who had their antipsychotic ceased (n=18), the reduction in the total
NPI-NH (2.5 points) and CMAI score (2.5 points) were similar to that reported by HALT.
Nonetheless, the findings from both interventions add to the weight of evidence that reducing antipsychotics does not generally worsen neuropsychiatric symptoms. Instead, the evidence from HALT and our study indicates that reduction may modestly improve some behavioural symptoms. One possible explanation could be that continued use of antipsychotics may be exacerbating neuropsychiatric symptoms, or, alternatively, that the use of non-pharmacological strategies and/or resident-centred care has increased in the absence of pharmacological treatment.
Earlier RCTs by Ballard et al. have reported an increase in neuropsychiatric symptoms upon antipsychotic withdrawal in residents with an initial NPI-NH score greater than 14.492, 632
However, in direct contrast, evidence in our study reported improvements in symptoms upon antipsychotic and benzodiazepine dose reduction in residents with higher initial NPI-NH scores.
In our sample, 19 of 30 (63%) residents who had their antipsychotic dose reduced and
18 of 42 (43%) residents who had their benzodiazepine dose reduced had an initial total NPI-
NH score greater than 14. Among these residents, the total NPI-NH score decreased by an average of 18.5 points (95%CI=-33.6, -3.4) in residents who had their antipsychotic dose reduced and an average of 14.6 points (95%CI=-25.8, -3.3) in residents who had their benzodiazepine dose reduced.
Whilst it was beyond the scope of this study to determine reasons for these differences, it should be acknowledged that Ballard et al.’s earlier studies492, 632 had a definitive date of
203 | Page Daniel J. Hoyle antipsychotic withdrawal, whereas residents in the current study had their antipsychotic dose reduced at various time points between baseline and four months. Therefore, improvements in their neuropsychiatric symptoms may have been associated with changes in the underlying condition prior to antipsychotic dose reduction. Additionally, the studies by Ballard et al.492, 632 were placebo-controlled trials with no extra training provided to RACF staff on non- pharmacological strategies. The training delivered in RedUSe50 may have led to increased use of non-pharmacological strategies and/or resident-centred care in the absence of pharmacological treatment. In any case, future, adequately powered studies are needed to conclusively determine factors related to change in neuropsychiatric symptoms upon antipsychotic and/or benzodiazepine dose reduction.
In our study, benzodiazepine dose reduction was not associated with worsened agitation (measured with the CMAI), anxiety (measured with NPI-NH factor 2) or sleep disturbances (measured with NPI-NH factor 4). Instead, there were trends toward improved agitation scores with benzodiazepine dose reduction.
In our study, improvement in agitation was driven predominantly by reduced frequency of physically nonaggressive behaviour (e.g., wandering, restlessness) associated with antipsychotic and benzodiazepine dose reduction. Benzodiazepine dose reduction was also associated with reduced verbally agitated behaviour (e.g., constant requests for attention). It is possible that confusion and disorientation associated with ongoing use of antipsychotics and/or benzodiazepines may have exacerbated symptoms such as constant requests for attention and/or wandering.205, 242, 537 Additionally, long-term antipsychotic use is associated with extrapyramidal side effects, such as akathisia, which consists of motor restlessness, feelings of inner tension, discomfort and wandering620, 633; symptoms that may be mistaken for physically nonaggressive behaviour.231 Further prospective, controlled studies are necessary to investigate these effects.
204 | Page Daniel J. Hoyle The lack of evidence to suggest worsened sleep disturbances with benzodiazepine dose reduction aligns with another smaller Australian intervention.518 Gilbert et al. recruited 69 residents from two geographically separated RACFs (intervention RACF=30 residents, control
RACF=39 residents). Residents who were acutely unwell, bed-bound or had a diagnosis of dementia were excluded from the study. The intervention involved educating nursing staff, sending prompts for dose reduction to the prescribers and a relaxation course for the residents.
Whilst the proportion of residents taking benzodiazepines in the intervention setting decreased significantly (70% to 35% over 12 weeks, p<0.001), there were no statistically significant differences between the mean sleep satisfaction scores for the intervention or control groups over time.518 Combined with our results, these findings provide support that, generally, benzodiazepine dose reduction in RACF residents does not worsen sleep disturbances within interventions that provide education for RACF staff.
6.3.4.1 Limitations
The outcomes of this study should be interpreted with consideration of several limitations. First,
RedUSe was funded as an implementation project, with experimental research not permitted.
This precluded the use of a control group which made it difficult to determine whether changes in the outcome were a result of dose changes or other time-dependent effects (e.g., changes in neuropsychiatric symptoms with time). Future prospective studies should utilise a controlled design to overcome this limitation.
Second, whilst trends toward improved agitation with antipsychotic and benzodiazepine dose reduction were identified in this study, our results are derived from a small sample which impacts our ability to detect statistically significant associations. In Chapter 5.6.3, we calculated a required sample size of 184 residents taking antipsychotics and benzodiazepines. However, resident recruitment from many RACFs was lower than expected.
Whilst efforts were made to increase the sample size by recruiting more RACFs through
205 | Page Daniel J. Hoyle subsequent waves of the project, unexpected time-delays and budget constraints limited our ability to recruit more facilities. A major unexpected time-delay included the need to reapply for ethical approval for the sub-study through separate RACF ethics committees despite having
National Ethics Application Form (NEAF) ethical approval provided by the University.
Unfortunately, the RedUSe intervention had often started before ethical approval for the sub- study was provided, thus baseline data could not be sought. Secondly, there were vast distances between RACFs and limited funding available to facilitate travel to extra facilities. It is important that future studies are sufficiently powered and funded to investigate these trends further.
Third, considering the ethical requirement for arms-length recruitment, we relied on the champion nurses to identify all eligible residents and gain consent from them or the ‘person responsible’ for them. These nurses were not trained researchers and were required to perform this task on top of their usual nursing duties. Efforts were made to confirm that the nurse had attempted to recruit all eligible residents. For example, each champion nurse was provided with a form to indicate the number of eligible residents identified and the number of consent forms sent. This form was rarely completed; therefore, it was not considered a reliable measure for recruitment. Considering the reliance on the champion nurses to recruit residents, there was possibly an element of non-random recruitment of residents who were stable and perceived to no longer have a need for their sedating medication. This may have affected the representativeness of the sample to the broader antipsychotic and/or benzodiazepine taking population. Future studies should investigate other approaches to ensure that all eligible residents are approached, such as using trained impartial research assistants to recruit residents.
Fourth, as a national implementation project, it was not considered logistically viable to collect data outside of the baseline and four-month data collection periods. Consequently, we could not accurately determine the duration of antipsychotic and/or benzodiazepine use prior to the intervention or whether residents had multiple dose changes between baseline and
206 | Page Daniel J. Hoyle four months. Future interventions should ideally review antipsychotic and/or benzodiazepine use in the months prior to the intervention to investigate differences in outcomes based on the duration of use. Furthermore, information on antipsychotic and/or benzodiazepine dose changes should be collected regularly and/or prescribing records assessed to provide a more detailed picture of the impact that the intervention had on medication use.
Fifth, some residents may have experienced co-occurring delirium. Therefore, it is possible that some of the trends towards improvement in symptoms at four months with antipsychotic dose reduction may be attributed, in part, to the resolution of delirium. A total of eight residents had a history of delirium but it was unclear whether this was a current diagnosis.
Lastly, neither RACF staff nor assessors were blind to medication use at baseline and four months. Whilst this may have introduced some reporting bias, it is unlikely, as we did not detect any statistically significant changes in the neuropsychiatric symptoms related to dose reduction.
6.3.5 Conclusion
Our results found no evidence that neuropsychiatric symptoms were worsened with antipsychotic and/or benzodiazepine dose reduction. Interestingly, there were small potential improvements and/or possible mitigation of deterioration in symptoms with antipsychotic and benzodiazepine reduction. In the future, large scale randomised-controlled trials would be required to determine whether there are indeed positive effects of dose reduction over long periods of time.
207 | Page Daniel J. Hoyle 6.4 Quality of life
6.4.1 Introduction
Maximisation of quality of life (QoL) is a major goal within residential aged care.634, 635 The World
Health Organization defines QoL as “an individual’s perception of their position in life in the context of the culture and value systems in which they live and in relation to their goals, expectations, standards and concerns. It is a broad ranging concept affected in a complex way by the person’s physical health, psychological state, personal beliefs, social relationships and their relationship to salient features of their environment”.636 In Australia, enhancement of resident QoL is referred to several times within the Quality of Care Principles 2014 for residential aged care,637 and is expected to be included as a RACF quality indicator in the near future.638, 639 Despite the focus on preserving and promoting QoL, residents of RACFs are more likely to report a lower QoL compared to their community-dwelling counterparts.500
Several studies have identified neuropsychiatric symptoms228-231 and psychotropic drug use500, 640, 641 as risk factors for lower QoL. However, despite QoL measures being increasingly used in research within RACFs,642 studies investigating relationships between psychotropic dose reduction and QoL are still lacking. Through our systematic review, in Chapter 4, only three studies reported on changes in QoL associated with a reduction in antipsychotic use.515, 517, 520
Two of these studies found no change in QoL,517, 520 whereas the other reported a significant decline in QoL associated with antipsychotic cessation.515 None of the studies investigated the impact that benzodiazepine dose reduction has on QoL.
A barrier to psychotropic reduction among nursing staff and prescribers is the perception that the resident’s QoL may deteriorate.44, 162 Considering uncertainty in the relationship between psychotropic reduction and QoL and the goal to maximise QoL within residential aged care, the aim of Chapter 6.4 is to investigate changes in QoL associated with
208 | Page Daniel J. Hoyle antipsychotic and benzodiazepine dose reduction. Figure 37 illustrates where these findings fit in this thesis.
Figure 37: Flow of the analysis for the clinical outcomes. This sub-chapter is dedicated to investigating the association between antipsychotic and benzodiazepine dose reduction and change in quality of life. GP=General Practitioner. 6.4.2 Methods
QoL was assessed at baseline and four months in a sample of residents involved in RedUSe, through psychometric testing of the RACF nursing and care staff, the residents (if they were able to provide informed consent) and a mail-out to the ‘person responsible’ for the residents (e.g., family member, legal guardian), using the AQoL-4D tool (Chapter 5.7.1.5).541 Whilst the responses from the RACF nursing and care staff were the primary measure of QoL, responses from the residents and ‘person responsible’ for the residents were sought as well to facilitate comparisons.
Differences between self-reported and proxy-reported QoL have been documented in the literature.643-645 Generally, proxies rate QoL lower than the patient themselves.643, 644
However, a recent study assessing the QoL, using the EuroQoL-5D (EQ-5D), in 556 RACF UK residents found that care staff overestimated the EQ-5D utility scores for residents with no cognitive impairment and underestimated scores for those with severe cognitive impairment.645
209 | Page Daniel J. Hoyle Inter-rater reliability in the current study was investigated by calculating intraclass correlation (ICC) coefficients using a one-way random-effects model within the ‘irr’ package in
R.594 Most of the ICC coefficients indicated poor to moderate inter-rater reliability across the utility and dimension scores (Appendix K).646 Considering these results and earlier findings of substantial inter-rater variability, the decision was made to analyse the data separately based on the rater (nursing staff, ‘person responsible’ for the residents, and the residents themselves).
RACF staff rated QoL at baseline and four months for all but one resident (one resident initially taking a benzodiazepine was erroneously missed at baseline). In contrast, the ‘person responsible’ for the resident assessed the resident’s QoL at baseline and four months in approximately one-third of cases. Similarly, the resident assessed their own QoL at baseline and four months in approximately one-third of cases. However, resident-rated QoL among antipsychotic users was collected in just four residents (Table 41). Considering that relatively few residents had their QoL measured by the ‘person responsible’ for the resident or the residents themselves, analysis of these data is reported in Appendix K, while RACF staff-rated
QoL is presented here.
Table 41: Count of residents with an Assessment of Quality of Life-4D score at baseline and four months, stratified by the rater.
Nursing staff Person responsible Resident AP users: 83 AP users: 32 AP users: 4 BZ users: 117 BZ users: 31 BZ users: 38 AP users=Residents prescribed antipsychotics at baseline, BZ users=Residents prescribed benzodiazepines at baseline 6.4.2.1 Assessment of Quality of Life scoring
As per Chapter 5.7.1.5.1, the AQoL-4D utility and dimension scores ranged from -0.04 (QoL worse than death) to 0 (QoL equivalent to death) to 1 (best QoL).
6.4.2.2 Statistical analysis
Similar to Chapter 6.3, two separate sets of analyses were run; one for antipsychotic users
(those with a chlorpromazine daily dose equivalent >0 mg at baseline) and one for
210 | Page Daniel J. Hoyle benzodiazepine users (those with a diazepam daily dose equivalent >0 mg at baseline). The distribution of the AQoL-4D utility and dimension scores were initially investigated using beeswarm plots. A paired t-test was utilised to detect statically significant changes in scores between baseline and four months.
To investigate the relationship between the magnitude of antipsychotic and benzodiazepine dose changes and the AQoL-4D utility and dimension scores, linear regression models were fitted with the ‘lm’ function in R.594 Potentially confounding variables, identified in Chapter 6.2, were controlled for within the models where appropriate. In the antipsychotic users’ model, this included CMAI factor 2 and 3 baseline scores. In the benzodiazepine users’ model, we controlled for the CMAI factor 2 baseline score and having a ‘documented diagnosis of dementia’. From the output of these models, we were able to estimate the unit change in the AQoL-4D utility and dimension scores associated with the percentage change in the antipsychotic and benzodiazepine doses between baseline and four months.
Inspection of Q-Q plots of the residuals revealed approximately normal distributions of the residuals with small violations of this assumption in the tails. To account for this violation, confidence intervals were bootstrapped. Plots of the AQoL-4D utility and dimension scores, and antipsychotic and benzodiazepine doses at baseline and four months were inspected for non- linear distribution of the data.
Lastly, the change in the AQoL-4D utility and dimension scores was assessed for residents who completely ceased their antipsychotic or benzodiazepine between baseline and four months, using paired t-tests.
211 | Page Daniel J. Hoyle 6.4.3 Results
6.4.3.1 Feasibility of measuring quality of life associated with antipsychotic and/or benzodiazepine dose change
When proxy-rated by nursing staff, it appeared feasible to measure QoL using the AQoL-4D; data were available for all participants present in the study at baseline and four months with the exception of one resident who was erroneously omitted at baseline.
6.4.3.2 Antipsychotic users
A total of 83 residents initially taking an antipsychotic had their QoL assessed by RACF nursing or care staff at baseline and four months. The AQoL-4D utility and the relationship, physical senses, and psychological wellbeing dimension scores remained relatively stable between baseline and four months. However, there was a clinically and statistically significant increase in the independent living dimension score (Table 42, Figure 38).4
4 A change in the AQoL-4D score by ≥0.06 points is considered clinically significant.
212 | Page Daniel J. Hoyle Table 42: Mean Assessment of Quality of Life-4D utility and dimension scores at baseline and four months.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Utility 0.15 (0.10, 0.21) 0.17 (0.11, 0.23) -0.559 82 0.578 Independent living 0.25 (0.17, 0.33) 0.38 (0.29, 0.46) -2.654 82 0.010 Relationships 0.47 (0.40, 0.54) 0.45 (0.39, 0.52) 0.384 82 0.702 Physical senses 0.74 (0.69, 0.79) 0.72 (0.67, 0.77) 0.860 82 0.393 Psychological wellbeing 0.82 (0.78, 0.86) 0.82 (0.78, 0.85) 0.038 82 0.970 df=degrees of freedom, 95%CI=95% confidence interval
Figure 38: Beeswarm plot of the AQoL-4D (A) utility and (B) independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing dimension scores at baseline and four months for all residents taking an antipsychotic at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval. AQoL-4D=Assessment of Quality of Life-4D tool.
The relationship between the magnitude of antipsychotic dose change and change in
AQoL-4D utility and dimension scores were investigated using regression modelling. After controlling for the CMAI factor 2 and 3 baseline scores, we identified a slight, albeit non- significant, increase in the AQoL-4D utility score associated with the magnitude of antipsychotic dose reduction (β=-0.001; i.e., AQoL-4D utility score was 0.01 higher, on average, for each 10%
213 | Page Daniel J. Hoyle reduction in the chlorpromazine daily dose equivalent, p=0.124, 95%CI=-0.002, 0.000; Table 43 and Figure 39).
As per Table 43, the magnitude of antipsychotic dose reduction was associated with a similar increase in the AQoL-4D dimension scores, albeit not significantly, after controlling for the CMAI factor 2 and 3 baseline scores.
There was, however, a clinically and statistically significant increase in the AQoL-4D utility score between baseline (mean=0.094, 95%CI=-0.010, 0.199) and four months
(mean=0.206, 95%CI=0.059, 0.353) in the 18 residents who had their antipsychotic ceased
(t(17)=-2.259, p=0.037). The increase in the utility score was primarily driven by an improvement in the independent living dimension, which increased from an average of 0.141
(95%CI=-0.001, 0.283) at baseline to 0.382 (95%CI=0.169, 0.594) at four months (t(17)=-2.404, p=0.028). There were no statistically or clinically significant changes in other AQoL-4D dimension scores.
Table 43: Estimated changes in the AQoL-4D utility and dimension scores associated with changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value Intercept -0.087 -0.209, 0.035 -1.418 0.160 Antipsychotic dose change (%) -0.001 -0.002, 0.000 -1.553 0.124 Baseline CMAI factor 2 score 0.006 -0.005, 0.017 1.041 0.301 Baseline CMAI factor 3 score 0.001 -0.012, 0.013 0.123 0.902 Utility Adjusted R2 0.037 Multiple R2 0.072 Intercept 0.064 -0.120, 0.248 0.696 0.489 Antipsychotic dose change (%) -0.001 -0.003, 0.000 -1.485 0.142 Baseline CMAI factor 2 score 0.004 -0.012, 0.020 0.482 0.631
living Baseline CMAI factor 3 score -0.002 -0.020, 0.017 -0.157 0.875 dimension Independent Independent Adjusted R2 -0.001 Multiple R2 0.035 Intercept -0.055 -0.230, 0.121 -0.619 0.538 Antipsychotic dose change (%) -0.001 -0.002, 0.001 -0.671 0.504 Baseline CMAI factor 2 score -0.002 -0.017, 0.014 -0.195 0.846 Baseline CMAI factor 3 score 0.004 -0.014, 0.022 0.433 0.666 dimension 2 Relationships Adjusted R -0.030 Multiple R2 0.008
214 | Page Daniel J. Hoyle Intercept -0.117 -0.214, -0.020 -2.404 0.019 Antipsychotic dose change (%) -0.001 -0.001, 0.000 -1.023 0.310 Baseline CMAI factor 2 score 0.008 -0.001, 0.016 1.745 0.085 Baseline CMAI factor 3 score -0.002 -0.011, 0.008 -0.297 0.767 dimension 2
Physical senses Adjusted R 0.051 Multiple R2 0.086 Intercept -0.061 -0.142, 0.021 -1.480 0.143 Antipsychotic dose change (%) -0.001 -0.001, 0.000 -1.513 0.134 Baseline CMAI factor 2 score 0.001 -0.006, 0.008 0.317 0.752 Baseline CMAI factor 3 score 0.003 -0.006, 0.011 0.638 0.525 wellbeing dimension 2 Psychological Adjusted R 0.017 Multiple R2 0.053 AQoL-4D=Assessment of Quality of Life-4D tool, 95%CI=95% confidence interval
Figure 39: Scatterplots showing the relationship between change (%) in the chlorpromazine daily dose equivalent and changes in the AQoL-4D (A) utility and (B) independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing dimension scores. AQoL-4D=Assessment of Quality of Life-4D tool, CPZ=chlorpromazine, DDE=daily dose equivalent.
215 | Page Daniel J. Hoyle 6.4.3.3 Benzodiazepine users
A total of 117 residents initially taking a benzodiazepine had their QoL assessed by RACF nursing or care staff at baseline and four months. Overall, the AQoL-4D utility and dimension scores remained relatively stable between baseline and four months (Table 44, Figure 40).
Table 44: Mean Assessment of Quality of Life-4D utility and dimension scores at baseline and four months.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Utility 0.20 (0.15, 0.24) 0.22 (0.17, 0.27) -0.917 116 0.361 Independent living 0.32 (0.25, 0.38) 0.37 (0.30, 0.44) -1.500 116 0.136 Relationships 0.58 (0.52, 0.64) 0.61 (0.55, 0.67) -0.890 116 0.375 Physical senses 0.78 (0.74, 0.82) 0.76 (0.72, 0.80) 1.294 116 0.198 Psychological wellbeing 0.80 (0.76, 0.83) 0.78 (0.74, 0.82) 0.751 116 0.454 df=degrees of freedom, 95%CI=95% confidence interval
Figure 40: Beeswarm plot of the AQoL-4D (A) utility and (B) independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing dimension scores at baseline and four months for all residents taking a benzodiazepine at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval. AQoL- 4D=Assessment of Quality of Life-4D tool.
216 | Page Daniel J. Hoyle Linear regression was used to examine the relationship between the magnitude of benzodiazepine dose reduction and changes in the AQoL-4D utility and dimension scores. After controlling for the baseline CMAI factor 2 score and having a documented diagnosis of dementia, we found no evidence of deterioration (or improvement) in the AQoL-4D utility or dimension scores associated with the magnitude of benzodiazepine dose reduction (Table 45 and Figure
41).
The AQoL-4D utility score remained relatively constant between baseline (mean=0.156,
95%CI=0.066, 0.247) and four months (mean=0.200, 95%CI=0.106, 0.295) in the 25 residents who had their benzodiazepine ceased (t(24)=-0.765, p=0.452). On the other hand, the relationships dimension score increased by a clinically but not statistically significant amount between baseline (mean=0.542, 95%CI=0.410, 0.674) and four months (mean=0.603,
95%CI=0.465, 0.742) in these residents (t(24)=-0.894, p=0.380). There were no clinically or statistically significant changes in other AQoL-4D dimension scores.
Table 45: Estimated changes in the AQoL-4D utility and dimension scores associated with changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value Intercept -0.114 -0.218, -0.010 -2.166 0.032 Benzodiazepine dose change (%) 0.000 -0.001, 0.001 0.003 0.997 Documented diagnosis of 0.054 -0.059, 0.167 0.946 0.346 dementia Utility Baseline CMAI factor 2 score 0.010 0.002, 0.019 2.482 0.015 Adjusted R2 0.054 Multiple R2 0.078 Intercept -0.141 -0.279, -0.004 -2.036 0.044 Benzodiazepine dose change (%) 0.000 -0.001, 0.001 0.202 0.841 Documented diagnosis of 0.083 -0.067, 0.232 1.099 0.274 dementia Baseline CMAI factor 2 score 0.015 0.004, 0.025 2.632 0.010 Independent Independent 2
living dimension Adjusted R 0.066 Multiple R2 0.090 Intercept -0.030 -0.152, 0.092 -0.492 0.624 Benzodiazepine dose change (%) -0.001 -0.001, 0.001 -0.161 0.872 Documented diagnosis of 0.012 -0.120, 0.144 0.179 0.858 dementia
dimension Baseline CMAI factor 2 score 0.005 -0.005, 0.014 0.944 0.347 Relationships Adjusted R2 -0.016 Multiple R2 0.011
217 | Page Daniel J. Hoyle Intercept -0.096 -0.167, -0.025 -2.693 0.008 Benzodiazepine dose change (%) 0.001 -0.000, 0.001 0.376 0.708 Documented diagnosis of 0.046 -0.030, 0.123 1.198 0.233 dementia
dimension Baseline CMAI factor 2 score 0.005 -0.001, 0.010 1.702 0.092 2 Physical senses Adjusted R 0.028 Multiple R2 0.053 Intercept -0.041 -0.124, 0.043 -0.956 0.341 Benzodiazepine dose change (%) 0.000 -0.000, 0.001 0.975 0.332 Documented diagnosis of 0.014 -0.077, 0.105 0.307 0.759 dementia wellbeing
dimension Baseline CMAI factor 2 score 0.002 -0.005, 0.009 0.630 0.530 Psychological Adjusted R2 -0.013 Multiple R2 0.013 AQoL-4D=Assessment of Quality of Life-4D tool, 95%CI=95% confidence interval
Figure 41: Scatterplots showing the relationship between change (%) in the diazepam daily dose equivalent (DDE) and changes in the AQoL-4D (A) utility and (B) independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing dimension scores. AQoL-4D=Assessment of Quality of Life-4D tool.
218 | Page Daniel J. Hoyle 6.4.4 Discussion
The key overall finding is that there was no evidence of worsened QoL with antipsychotic and/or benzodiazepine dose reduction. Importantly, however, there was a clinically and statistically significant improvement in QoL among residents who had their antipsychotic ceased. This improvement was primarily driven by improvements in the independent living dimension.
Additionally, there was a clinically but not statistically significant improvement in the AQoL-4D relationships dimension score associated with benzodiazepine withdrawal.
Based on our systematic review and updated literature search in Chapter 4, only five studies (from three interventions) had reported on the relationship between antipsychotic dose reduction and change in QoL.515, 517, 520, 531 The lack of deterioration in QoL with antipsychotic dose reduction in our study is consistent with the two oldest studies.517, 520 However, unlike these studies, which used dementia care mapping,517, 520 we used a multi-attribute utility tool to measure QoL. The use of a multi-attribute utility tool takes significantly less time to administer and allows for the calculation of quality-adjusted life years (QALYs), which are commonly used in economic evaluations of health programs.647 The calculation of cost per QALY gained indicates the explicit value of improvements in QoL over time lived.566
The other three studies presented QoL outcomes associated with the WHELD antipsychotic reduction program.515, 531, 532 In 2016, Ballard et al. performed a proof-of-concept study which included a person-centred care intervention with an antipsychotic review and/or a social interaction intervention and/or an exercise intervention among people with dementia residing in 16 UK RACFs.515 Residents who received the antipsychotic review intervention alone exhibited a 4.54-point (95%CI=-9.25, 0.19) deterioration (p=0.06) in the DEMQOL-proxy score compared to residents not receiving the antipsychotic review intervention.515 However, there was no evidence of deterioration in QoL in residents who received both an antipsychotic review and social interaction intervention.515 It is noteworthy that residents receiving the antipsychotic
219 | Page Daniel J. Hoyle review intervention alone also exhibited deterioration in neuropsychiatric symptoms, which were mitigated by the addition of a social interaction intervention.514, 516 It is possible that the deterioration in neuropsychiatric symptoms may have worsened QoL228-231 or may have contributed to staff distress leading to the reporting of lower QoL.648 Ballard et al. postulated that the social interaction intervention enabled staff to better understand and integrate person- centred care into their practice resulting in improved QoL.515
The WHELD intervention has since been optimised to incorporate both the antipsychotic review and social interaction interventions and has been implemented across 69
UK RACFs.531 The nine-month WHELD intervention led to a statistically significant 2.54-point improvement in the DEMQOL-proxy score (95%CI=0.81, 4.28) compared to treatment-as- usual.531 Additionally, there was a 1.96-point improvement in the DEMQOL-proxy score
(95%CI=0.38, 3.24) compared to treatment-as-usual among residents with clinically significant agitation (CMAI total score >40).532
Differences between WHELD and RedUSe may, in part, account for the differences in
QoL results between the interventions. In particular, WHELD was performed only in those with dementia and did not result in reductions to antipsychotic usage (antipsychotic prevalence was minimal at baseline).531 However, our study resulted in a 16.2% mean decrease in antipsychotic dose among residents with and without a diagnosis of dementia. It should be acknowledged, though, that WHELD was a person-centred care and psychological intervention designed to holistically improve the management of neuropsychiatric symptoms in residents with dementia.514 Furthermore, WHELD was implemented in an environment with relatively low antipsychotic use (approximately 9% at baseline), which indicates that most antipsychotic use may have been appropriate in this intervention and not have required reduction.531 RedUSe, on the other hand, was more focussed on the review and reduction of inappropriate antipsychotic and benzodiazepine use regardless of an underlying diagnosis of dementia.50 Furthermore,
220 | Page Daniel J. Hoyle antipsychotic use at baseline was higher (21.6%), which may have provided greater opportunity for dose reduction.50
Earlier studies have reported reductions in proxy-rated QoL for people with dementia over time.649, 650 On the other hand, a recent study investigating longitudinal changes in QoL in residents with (n=158) and without dementia (n=254) found no significant changes in either group.651 The authors acknowledged that the inclusion of a more heterogenic population, in terms of age, and cognitive and functional limitations compared to previous studies, may have accounted for the differences in outcomes.651 Our study was also composed of a highly heterogeneous sample, with and without dementia. Therefore, like Ydstebø et al.,651 the lack of statistically significant changes in QoL with dose reduction may reflect the heterogeneity in our sample.
Unlike earlier studies, we reported a clinically and statistically significant improvement in overall QoL with antipsychotic cessation. This change was driven by improvements in items related to the independent living dimension. Questions included in this dimension assessed the resident’s ability to perform self-care tasks and ADLs, and mobilise.541 The ability to perform self-care tasks and ADLs, and mobilise is dependent on cognitive (e.g., planning), motor (e.g., balance), and perceptual abilities.652 These abilities may be impaired by antipsychotic use (see
Chapter 3). Thus, cessation of antipsychotic use may have resulted in improvements related to the independent living dimension.
To our knowledge, there have been no similar multicomponent interventions in RACFs to explore the effects of benzodiazepine dose reduction on QoL (Chapter 4). However, our finding that QoL was unaffected by benzodiazepine dose reduction is consistent with a pilot study, albeit in a different clinical group of residents, on the feasibility of benzodiazepine discontinuation in five Belgian RACFs.494 Bourgeois et al. assessed the change in QoL for 38 cognitively competent residents taking benzodiazepines (long term) for insomnia who,
221 | Page Daniel J. Hoyle alongside their prescriber, had agreed to initiate discontinuation of their benzodiazepine. No differences in the change in QoL were detected among these residents between baseline and eight months.494 The inclusion of residents using benzodiazepines for other conditions (e.g., anxiety, agitation) in our study provides support that benzodiazepine dose reduction may be achieved in a broader range of residents, with and without cognitive impairment, without negatively affecting QoL, in most cases.
6.4.4.1 Limitations
The outcomes of this study should be interpreted with consideration of several limitations.
Some of the limitations discussed in Chapter 6.3 also affected our analysis in Chapter 6.4, such as:
• absence of a control group;
• small sample size;
• possible recruitment bias; and,
• lack of rater blinding.
Additionally, comparison with and pooling of results from other studies is challenging due to differences in the data collection tools and follow-up time period. In 2017, Millar et al. recommended a core outcome set for effectiveness trials aimed at optimising prescribing in older adults in RACFs.506 Adoption of this core outcome set may help to improve comparisons between interventions as well as pooling of the results.
Furthermore, the results presented in this sub-chapter are based on proxy-rated QoL, which are known to underestimate QoL.643, 644 Whilst we have provided the results for resident- rated and ‘person responsible’ for the resident-rated QoL in Appendix K, few residents in our study had the cognitive capability to participate in an interview. Unfortunately, none of the residents who had their antipsychotic ceased were able to be interviewed. Therefore, we are
222 | Page Daniel J. Hoyle unable to comment on whether residents rated their own QoL higher upon antipsychotic discontinuation.
6.4.5 Conclusion
In conclusion, RACF staff-rated QoL did not deteriorate with antipsychotic or benzodiazepine dose reduction. Moreover, there was evidence of an improvement in QoL related to independent living associated with antipsychotic cessation. These results suggest that antipsychotic and/or benzodiazepine medications can be safely reduced in the RACF setting without negatively affecting the resident’s QoL and, in the case of antipsychotic withdrawal, potentially improve QoL. Larger, controlled, prospective studies are needed to further investigate changes in QoL related to antipsychotic dose reduction.
223 | Page Daniel J. Hoyle 6.5 Social engagement
6.5.1 Introduction
Social engagement is the “ability to initiate social interaction and be receptive to social overtures from others”.653 Deterioration of physical and cognitive abilities in many older people can limit social engagement in the community and lead to RACF placement.653 RACF placement may further disrupt relationships with family and friends and, combined with physical and cognitive impairments, lead to social withdrawal.653 A 2002 study found that residents of US
RACFs spent 65% of their time doing little to nothing and only 12% of their time in social activities.654 Whilst little research has investigated the engagement of Australian RACF residents, recent qualitative research of the lived experiences of RACF residents in Victoria and
Queensland identified themes of social withdrawal.655 Residents in this study felt that staff provided good clinical care but were often too time-poor to engage in meaningful social interactions. Furthermore, residents indicated that there is little on offer at RACFs to keep them socially active.655
Higher levels of social engagement may be associated with several benefits to residents of RACFs. In particular, social engagement has been associated with reductions in depression656 and mortality,653 and improvements in function,657 cognition657, 658 and QoL.658 Additionally, the addition of a social interaction intervention to an antipsychotic review intervention has been shown to mitigate negative effects that the antipsychotic review intervention alone had on neuropsychiatric symptoms and QoL.514-516 This further strengthens a positive relationship between social engagement and QoL and supports an association between social engagement and reduced neuropsychiatric symptoms.
The effect that antipsychotic use has on social engagement is debatable. It is argued that, when prescribed cautiously, antipsychotics can improve the physical and psychological wellbeing of older people, and, potentially, result in increased levels of social engagement.635
224 | Page Daniel J. Hoyle Conversely, adverse effects associated with antipsychotic use, such as sedation, may negatively affect a resident’s ability to engage socially in the RACF.635
No studies within our systematic review (Chapter 4) had reported on the effects of antipsychotic and/or benzodiazepine dose reduction on social engagement. However, recently an Australian antipsychotic deprescription study, HALT, found that antipsychotic deprescription was not associated with changes in social engagement, as measured with the MOSES- withdrawal subscale.530
Considering that social engagement is important to the psychological and physical wellbeing of RACF residents and there is a concern that an increased sedative load may reduce levels of social engagement, Chapter 6.5 aimed to investigate changes in resident social engagement associated with antipsychotic and/or benzodiazepine dose reduction within the
RedUSe program. Figure 42 illustrates where these findings fit in this overall thesis.
Figure 42: Flowchart of the analysis for the clinical outcomes study. This sub-chapter is dedicated to investigating the association between antipsychotic and benzodiazepine dose reduction and change in social engagement. GP=General Practitioner.
225 | Page Daniel J. Hoyle 6.5.2 Methods
Social engagement was assessed at baseline and four months, through psychometric testing of
RACF nursing and care staff, using the MOSES-withdrawal subscale (Chapter 5.7.1.6).541 The total score was calculated by summing the scores from each item. Scores varied from 8 to 32, where higher scores indicated higher levels of social withdrawal (i.e., less social engagement).
6.5.2.1 Statistical analyses
Similar to the preceding sub-chapter (Chapter 6.4), two separate analyses were run; one for antipsychotic users (those with a chlorpromazine daily dose equivalent >0 mg at baseline) and one for benzodiazepine users (those with a diazepam daily dose equivalent >0 mg at baseline).
The distribution of the MOSES-withdrawal subscale scores was initially investigated using beeswarm plots. A paired t-test was also utilised to detect statically significant changes in scores between baseline and four months.
To investigate the relationship between the magnitude of antipsychotic and/or benzodiazepine dose changes and changes in the MOSES-withdrawal subscale score, linear regression models were fitted with the ‘lm’ function in R.594 Potentially confounding variables, identified in Chapter 6.2, were controlled for within the models where appropriate. In the antipsychotic users’ model, this included CMAI factor 2 and 3 baseline sores. In the benzodiazepine users’ model, we controlled for the potentially confounding variables
‘documented diagnosis of dementia’ and the CMAI factor 2 baseline score. From the output of these models, we were able to estimate the unit change in the MOSES-withdrawal subscale score associated with the percentage change in the antipsychotic and benzodiazepine doses between baseline and four months.
Inspection of Q-Q plots of the residuals of the regression models indicated that they were normally distributed. Additionally, plots of the MOSES-withdrawal subscale scores, and
226 | Page Daniel J. Hoyle antipsychotic and benzodiazepine doses at baseline and four months were inspected for non- linear distribution of the data.
Lastly, the change in the MOSES-withdrawal subscale score was assessed for residents who completely ceased their antipsychotic or benzodiazepine between baseline and four months, using paired t-tests.
6.5.3 Results
6.5.3.1 Feasibility of measuring social withdrawal associated with antipsychotic and/or benzodiazepine dose change
Overall, it appeared feasible to measure changes in social withdrawal associated with antipsychotic and/or benzodiazepine dose change using the MOSES-withdrawal subscale; data were available for all participants who were present in the study at baseline and four months.
6.5.3.2 Antipsychotic users
A total of 83 residents initially using an antipsychotic had their social engagement assessed by
RACF nursing or care staff with the MOSES-withdrawal subscale at baseline and four months.
On average, MOSES-withdrawal scores increased between baseline (mean=19.22,
95%CI=17.90, 20.53) and four months (mean=20.71, 95%CI=19.45, 21.97), signifying that the residents had become less engaged (Figure 43). The results of a paired t-test indicated that this difference was statistically significant (t(82)=-2.188, p=0.032).
227 | Page Daniel J. Hoyle
Figure 43: Beeswarm plot of the MOSES-withdrawal subscale scores at baseline and four months for all residents taking an antipsychotic at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval. MOSES=Multidimensional Observation Scale for Elderly Subjects.
The relationship between the magnitude of antipsychotic dose change and change in social engagement was investigated using regression modelling. After controlling for CMAI factor 2 and 3 baseline scores, we found no significant association between change in MOSES- withdrawal subscale and the magnitude of antipsychotic dose change (β=0.016; i.e., MOSES- withdrawal subscale was 0.16 lower for each 10% reduction in the chlorpromazine daily dose equivalent, p=0.192; 95%CI=-0.008, 0.039; Table 46, Figure 44).
On average, the MOSES-withdrawal subscale score did not change between baseline
(mean=20.7, 95%CI=17.8, 23.7) and four months (mean=20.7, 95%CI=17.1, 24.2) in the 18 residents who had their antipsychotic ceased (t(17)=0.034, p=0.974).
228 | Page Daniel J. Hoyle Table 46: Estimated changes in the MOSES-withdrawal subscale score associated with changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value Intercept 2.823 0.198, 5.447 2.141 0.035
Antipsychotic dose change (%) 0.016 -0.008, 0.039 1.316 0.192 - Baseline CMAI factor 2 score -0.100 -0.332, 0.133 -0.851 0.398 Baseline CMAI factor 3 score 0.030 -0.235, 0.294 0.224 0.823 MOSES subscale 2 withdrawal Adjusted R 0.004 Multiple R2 0.041 MOSES=Multidimensional Observation Scale for Elderly Subjects, 95%CI=95% confidence interval
Figure 44: Scatterplot of the change in the MOSES-withdrawal subscale score versus change (%) in the chlorpromazine daily dose equivalent. The blue line represents the predicted change in the MOSES-withdrawal subscale score based on the change (%) in the chlorpromazine daily dose equivalent. MOSES=Multidimensional Observation Scale for Elderly Subjects.
6.5.3.3 Benzodiazepine users
A total of 118 residents initially using a benzodiazepine had their social engagement assessed by RACF nursing or care staff with the MOSES-withdrawal subscale at baseline and four months.
On average, the MOSES-withdrawal subscale score increased between baseline
(mean=15.98, 95%CI=14.92, 17.04) and four months (mean=17.15, 95%CI=15.95, 18.35) signifying that residents became less engaged (Figure 45). The results of a paired t-test indicated that this difference was statistically significant (t(117)=-2.402, p=0.018).
229 | Page Daniel J. Hoyle
Figure 45: Beeswarm plot of the MOSES-withdrawal subscale scores at baseline and four months for all residents taking a benzodiazepine at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval. MOSES=Multidimensional Observation Scale for Elderly Subjects.
The relationship between the magnitude of benzodiazepine dose change and change in the MOSES-withdrawal subscale score was investigated using regression modelling. After controlling for the CMAI factor 2 baseline score and documented diagnosis of dementia, we found no evidence of a relationship between change in the MOSES-withdrawal subscale and magnitude of benzodiazepine dose change (β=-0.004; i.e., the MOSES-withdrawal subscale score increased by 0.04 points, on average, for every 10% reduction in the diazepam daily dose equivalent, p=0.590; 95%CI=-0.018, 0.010; Table 47, Figure 46).
On average, the MOSES-withdrawal subscale score did not increase substantially between baseline (mean=15.8, 95%CI=13.6, 17.9) and four months (mean=17.0, 95%CI=14.1,
19.8) in the 25 residents who had their benzodiazepine ceased (t(24)=-0.964, p=0.345).
230 | Page Daniel J. Hoyle Table 47: Estimated changes in the MOSES-withdrawal subscale score associated with changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value Intercept 1.980 0.071, 3.889 2.055 0.042
Benzodiazepine dose change (%) -0.004 -0.018, 0.010 -0.541 0.590 - Documented diagnosis of dementia 0.139 -1.919, 2.198 0.134 0.894 Baseline CMAI factor 2 score -0.084 -0.235, 0.066 -1.108 0.270 MOSES subscale 2 withdrawal Adjusted R -0.013 Multiple R2 0.013 MOSES=Multidimensional Observation Scale for Elderly Subjects, 95%CI=95% confidence interval
Figure 46: Scatterplot of the change in the MOSES-withdrawal subscale score versus change (%) in the diazepam daily dose equivalent. The blue line represents the predicted change in the MOSES-withdrawal subscale score based on the change (%) in the diazepam daily dose equivalent. MOSES=Multidimensional Observation Scale for Elderly Subjects. 6.5.4 Discussion
Overall, there were statistically significant increases in social withdrawal between baseline and four months across all residents initially using antipsychotic and/or benzodiazepines. However, there was no indication that antipsychotic and/or benzodiazepine dose reduction was associated with social withdrawal to a greater extent than those who continued or increased
231 | Page Daniel J. Hoyle the dose of their sedative. Instead, there was a slight, albeit, non-significant improvement in social withdrawal among residents who had their antipsychotic dose reduced.
As mentioned in the introduction to this sub-chapter, none of the antipsychotic or benzodiazepine dose reduction interventions included in our systematic review (Chapter 4) investigated the impact on social engagement. However, our results align with a recent
Australian antipsychotic deprescription project, called the HALT project, which found no significant change in social engagement upon antipsychotic deprescription.530
Saleh et al. have previously argued that cautious prescribing of antipsychotics may improve the physical and psychological wellbeing of older people, potentially with positive effects on social engagement.635 Conversely, the sedative effects and risk of other adverse effects may negatively affect a resident’s ability to socially engage in the RACF.635 Considering that social withdrawal increased for all residents between baseline and four months but this effect was not mediated by changes in the antipsychotic and/or benzodiazepine doses (i.e., dose changes were not related to changes in social withdrawal), our results indicate that factors other than antipsychotic and/or benzodiazepine use (e.g., changes in underlying conditions) may have a greater impact on social engagement. For example, the progression of underlying dementia may have accounted for an increase in social withdrawal between baseline and four months.659 Social withdrawal is one of the earliest psychiatric symptoms of Alzheimer’s disease and worsens as the disease progresses.659
6.5.4.1 Strengths and limitations
To our knowledge, this is the first RACF-based intervention to report on the association between benzodiazepine dose reduction and resident social engagement. Additionally, unlike the HALT study,530 which treated antipsychotic deprescription as a binary outcome (i.e., deprescribed or did not deprescribe), our study was able to estimate the association between the magnitude of dose change and change in social engagement.
232 | Page Daniel J. Hoyle However, the outcomes of this study should be interpreted with consideration of several limitations. Some of these limitations are discussed in detail in Chapter 6.3, including:
• absence of a control group;
• small sample size;
• possible recruitment bias; and,
• lack of rater blinding.
Furthermore, whilst we used the validated MOSES-withdrawal subscale to monitor social engagement, a cut-off for a clinically important change in the subscale has not been determined. In this study, we utilised the 0.5 of a standard deviation rule, using baseline scores to estimate a minimum clinically important change in score (minimum clinically important change of 3.1 points with this method). However, future research is needed to determine cut- offs for a clinically significant change in the MOSES-withdrawal subscale.
6.5.5 Conclusion
Our results indicate that antipsychotic and benzodiazepine dose reduction, in a sample of residents taking part in the RedUSe project, was not associated with a change in social engagement. Whilst there was an increase in social withdrawal across the overall sample, this likely relates to other factors, such as progression in underlying conditions (e.g., progression in dementia). Future research utilising larger sample sizes may be useful to further investigate associations between antipsychotic and/or benzodiazepine dose reduction and change in social engagement.
233 | Page Daniel J. Hoyle 6.6 Activities of daily living
6.6.1 Introduction
ADLs are practical everyday tasks that are considered necessary for self-care (e.g., personal hygiene, dressing, toileting/continence, ambulating).587, 589, 652 These activities are also known as basic ADLs. ADLs related to independent living in the community (e.g., managing finances, medications) are called instrumental ADLs.652, 660 The ability to perform basic and instrumental
ADLs depends on cognitive (e.g., planning), motor (e.g., balance), and perceptual abilities.652
Given that this thesis describes a study analysing the effects of antipsychotic and/or benzodiazepine dose reduction in residents of RACFs, reporting of the instrumental ADLs, which are not required for fundamental function, was considered unnecessary. Therefore, the term
“ADL” throughout this thesis refers solely to basic ADLs, such as dressing, toileting and ambulating.
Adequate performance of ADLs is necessary for the independence of older people in the community and the inability to perform these ADLs has been linked to increased RACF placement.661 A 2007 meta-analysis found that older people with dependencies in ≥3 ADLs had a 3.25-fold increased risk of being admitted to a RACF over the next two to six years.661 In 2015, a 36-month follow-up of older people (n=1,001) receiving domiciliary care found that transfer to RACF care was associated with worsening in ADL performance.589 Considering that many residents entering RACFs experience difficulties performing ADLs, it is unsurprising that a substantial proportion of direct care is directed toward assistance with ADLs.662, 663 A 2019 study investigating the use of direct care in 47 Norwegian RACFs (n=537) every six months for three years found that assistance with ADLs constituted the biggest proportion of direct care time
(50-60%) and a decrease in the ability to perform ADLs was associated with an increase in direct care time.662
234 | Page Daniel J. Hoyle Two earlier interventions in our systematic review (Chapter 4) investigated associations between antipsychotic dose reduction and ADL dependency.520, 524 Whilst both studies reported no changes in the ADL dependency levels and antipsychotic dose reduction, the ADL tool used in the study by Ballard et al.520 and the method of scoring used by Thapa et al.524 may have not been sensitive enough to detect small changes in function. We did not identify any studies investigating associations between benzodiazepine dose reduction and ADL dependency.47
Ballard et al.520 measured dependency for assistance with ADLs using the Barthel
Index.664 The Barthel Index has been criticised for insensitivity to change, especially in people performing at the extremes of the scale.665, 666 Considering that RACF residents generally require high levels of care, the floor effects of the Barthel Index may have masked the measurement of small changes in ADL dependency.
Thapa et al.524 monitored changes in ADL performance using the PSMS.587 However, the researchers used a scoring method by which residents performing ADLs completely independently received a score of 1 and residents requiring any assistance with ADLs, regardless of magnitude, received a score of 0 (total score falls between 0 and 6, where higher scores indicate greater independence with ADLs). Given that most residents would be unlikely to perform many, if any, of the tasks independently, this method of scoring is insensitive to small improvements in ADL performance and, therefore, is not ideal for use in studies based in the
RACF setting.
Considering that improvements in the functioning of residents may reduce the time required by care staff to assist with ADLs and, subsequently, allow for better resource allocation,
Chapter 6.6 aimed to investigate changes in ADL dependency levels associated with antipsychotic and/or benzodiazepine dose reduction in a sample of residents within RedUSe.
Figure 47 illustrates where these findings fit in this thesis.
235 | Page Daniel J. Hoyle
Figure 47: Flowchart of the analysis for the clinical outcomes study. This sub-chapter is dedicated to investigating the association between antipsychotic and benzodiazepine dose reduction and change in activities of daily living. GP=General Practitioner. 6.6.2 Methods
6.6.2.1 Physical Self-Maintenance Scale scoring
As per Chapter 5.7.1.7, the PSMS scores presented in this chapter are calculated by summing all item scores. The overall score ranges from 6 to 30, where higher scores represent greater dependency on others to perform ADLs.
6.6.2.2 Statistical analysis
As per the preceding sub-chapter (Chapter 6.5), two separate sets of analyses were run; one for antipsychotic users (those with a chlorpromazine daily dose equivalent >0 mg at baseline) and one for benzodiazepine users (those with a diazepam daily dose equivalent >0 mg at baseline).
The distribution of the PSMS was initially investigated using beeswarm plots. A paired t-test was also utilised to detect statistically significant changes in scores between baseline and four months for all residents initially prescribed an antipsychotic or benzodiazepine.
To investigate the relationship between the magnitude of antipsychotic and benzodiazepine dose changes and the PSMS score, linear regression models were fitted with
236 | Page Daniel J. Hoyle the ‘lm’ function in R.594 Potentially confounding variables, identified in Chapter 6.2, were controlled for within the models where appropriate. In the antipsychotic users’ model, this included CMAI factor 2 and 3 baseline sores. In the benzodiazepine users’ model, we controlled for the potentially confounding variables ‘documented diagnosis of dementia’ and the CMAI factor 2 baseline score. From the output of these models, we were able to estimate the unit change in the PSMS score associated with the percentage change in the antipsychotic and benzodiazepine doses between baseline and four months.
Inspection of Q-Q plots of the residuals revealed approximately normal distributions of the residuals with small violations of this assumption in the tails. To account for this violation, confidence intervals were bootstrapped. Plots of the PSMS scores and antipsychotic and benzodiazepine doses at baseline and four months were inspected for non-linear distribution of the data.
Lastly, the change in the PSMS score was assessed for residents who completely ceased their antipsychotic or benzodiazepine between baseline and four months, using paired t-tests.
6.6.3 Results
6.6.3.1 Feasibility of measuring activities of daily living associated with antipsychotic and/or benzodiazepine dose change
Overall, it appeared feasible to measure changes in activities of daily living associated with antipsychotic and/or benzodiazepine dose change using the PSMS; data were available for all participants who were present in the study at baseline and four months.
6.6.3.2 Antipsychotic users
A total of 83 residents initially using an antipsychotic had their ability to perform ADLs assessed with the PSMS at baseline and four months by RACF nursing or care staff.
237 | Page Daniel J. Hoyle On average, residents became slightly, albeit non-significantly, more dependent on staff for their ADLs between baseline (mean=19.82, 95%CI=18.40, 21.24) and four months
(mean=20.57, 95%CI=19.16, 21.98) (t(82)=-1.448, p=0.152; Figure 48).
Figure 48: Beeswarm plot of the Physical Self-Maintenance Scale scores at baseline and four months for residents taking an antipsychotic at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval.
The relationship between the magnitude of antipsychotic dose change and change in the total PSMS score was investigated using regression modelling. After controlling for baseline
CMAI factor 2 and 3 scores, we found a small, albeit non-significant, association between an increase in the PSMS score and magnitude of antipsychotic dose reduction (β=-0.013; i.e., total
PSMS score was 0.13 points higher, on average, for each 10% reduction in the chlorpromazine daily dose equivalent, p=0.154, 95%CI=-0.031, 0.005; Table 48, Figure 49).
On average, the PSMS score increased, non-significantly, by one point between baseline
(mean=20.8, 95%CI=17.7, 24.0) and four months (mean=21.8, 95%CI=18.5, 25.2) in the 18 residents who had their antipsychotic ceased (t(17)=-0.802, p=0.434).
238 | Page Daniel J. Hoyle Table 48: Estimated changes in the Physical Self-Maintenance Scale score associated with changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value Intercept 1.528 -0.453, 3.509 1.536 0.129 - Antipsychotic dose change (%) -0.013 -0.031, 0.005 -1.438 0.154 Baseline CMAI factor 2 score -0.057 -0.232, 0.119 -0.639 0.524
Scale Baseline CMAI factor 3 score -0.023 -0.223, 0.176 -0.230 0.819 Adjusted R2 0.007 Physical Self Maintenance Maintenance Multiple R2 0.043 95%CI=95% confidence interval
Figure 49: Scatterplot of the change in the Physical Self-Maintenance Scale score versus change (%) in the chlorpromazine daily dose equivalent. The blue line represents the predicted change in Physical Self-Maintenance Scale score based on the change (%) in the chlorpromazine daily dose equivalent.
6.6.3.3 Benzodiazepine users
Overall, 118 residents initially using a benzodiazepine had their ability to perform ADLs assessed with the PSMS by RACF nursing and care staff at baseline and four months.
On average, the residents’ ability to perform ADLs remained relatively stable between baseline (mean=16.40, 95%CI=15.33, 17.47) and four months (mean=16.58, 95%CI=15.46, 17.70)
(t(117)=-0.497, p=0.620; Figure 48).
239 | Page Daniel J. Hoyle
Figure 50: Beeswarm plot of the Physical Self-Maintenance Scale scores at baseline and four months for residents taking a benzodiazepine at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval.
The relationship between the magnitude of benzodiazepine dose change and change in the PSMS total score was investigated using regression modelling. After controlling for the CMAI factor 2 baseline score and documented diagnosis of dementia, a slight, albeit non-significant, decline in the total PSMS score was associated with a greater magnitude of benzodiazepine dose reduction (β=0.007; i.e., PSMS total score was 0.07 points lower, on average, for each 10% reduction in the diazepam daily dose equivalent, p=0.177, 95%CI=-0.003, 0.017; Table 49, Figure
51).
On average, the PSMS score decreased slightly, albeit non-significantly, between baseline (mean=17.4, 95%CI=15.2, 19.5) and four months (mean=15.9, 95%CI=13.6, 18.2) in the
25 residents who had their benzodiazepine ceased (t(24)=1.482, p=0.151).
240 | Page Daniel J. Hoyle Table 49: Estimated changes in the Physical Self-Maintenance Scale score associated with changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value Intercept 1.659 0.298, 3.020 2.414 0.017
Benzodiazepine dose change (%) 0.007 -0.003, 0.017 1.358 0.177 - Documented diagnosis of 0.369 -1.099, 1.837 0.498 0.620 dementia Baseline CMAI factor 2 score -0.137 -0.244, -0.029 -2.516 0.013
Physical Self Adjusted R2 0.047 Maintenance Scale Multiple R2 0.072 95%CI=95% confidence interval
Figure 51: Scatterplots of the change in the Physical Self-Maintenance Scale score versus change (%) in the diazepam daily dose equivalent. The blue line represents the predicted change in Physical Self-Maintenance Scale score based on the change (%) in the diazepam daily dose equivalent. 6.6.4 Discussion
To our knowledge, this is the first study to investigate the impact benzodiazepine dose reduction has on ADL dependency levels. Overall, we found no evidence that reducing the antipsychotic or benzodiazepine dose was related to significant deteriorations on ADL dependency levels between baseline and four months. However, regression modelling indicated tendencies toward reduced ADL dependency levels with benzodiazepine dose
241 | Page Daniel J. Hoyle reduction and cessation and increased ADL dependency levels with antipsychotic dose reduction and cessation, although statistical significance was not reached considering the sample size.
The absence of evidence to suggest a significant increase in ADL dependency levels with antipsychotic dose reduction aligns with findings from two earlier interventions..520, 524 As mentioned in the introduction to this chapter (Chapter 6.6.1), there was concern that the tools or scoring method used in these interventions may have limited ability to detect changes in ADL dependency. First, Ballard et al.520 measured dependency for assistance with ADL using the
Barthel Index,664 which has been criticised for insensitivity to change, especially in people performing at the extremes of the scale.665, 666 Second, similar to our study, Thapa et al.524 monitored changes in ADL performance using the PSMS.587 However, Thapa et al. used a different scoring method which may have been less sensitive to the small changes in ADL dependency levels.524 The lack of association between antipsychotic dose change and changes in ADL dependency levels despite differences in tools and scoring methods used supports the hypothesis that antipsychotic dose reduction does not significantly affect ADL dependency levels.
Originally, we postulated that a reduction in the CNS depressant effects of the antipsychotic may have improved cognition and, subsequently, the ability to perform ADLs.
However, there was a slight tendency toward increased ADL dependency levels with antipsychotic cessation. Most residents in our sample were likely using antipsychotics to treat neuropsychiatric symptoms associated with dementia, in particular agitation, aggression and psychosis. Considering dementia is a progressive neurodegenerative disease, neuropsychiatric symptoms change with damage to the CNS. It is possible that residents who have had their antipsychotic dose reduced may be those who no longer exhibit these symptoms as their
242 | Page Daniel J. Hoyle dementia has progressed. In these cases, advancing dementia may also be increasing the need for further assistance with ADLs.
Non-significant reductions in ADL dependency levels associated with benzodiazepine dose reduction suggest that a reduction in the CNS depressant effects of the benzodiazepines may have allowed for a slight improvement in the ability to perform ADLs independently. Whilst changes in ADLs have not been reported from comparable interventions, there is a small amount of evidence from an eight-week benzodiazepine deprescribing study of 30 residents from one RACF that suggests discontinuation of benzodiazepines may lead to improvements in cognitive function and stability of the body.667 Potential improvements in stability and cognitive function in our study may have improved the confidence in residents performing ADLs.
6.6.4.1 Strength and limitations
To our knowledge, this is the only study aimed at reducing benzodiazepine use in RACFs which has monitored for changes in ADL dependency levels. Furthermore, the scoring method used in our study overcomes the issues of insensitivity seen with the tools and scoring methods used in previous studies investigating associations between antipsychotic dose change and change in
ADL dependency levels.520, 524
However, the outcomes of this study should be interpreted with consideration of several limitations. Some of these limitations are discussed in detail in Chapter 6.3, including:
• absence of a control group;
• small sample size;
• possible recruitment bias; and,
• lack of rater blinding.
Furthermore, the ADL dependency levels were measured with psychometric testing of the RACF staff rather than interviewing the resident themselves. It is possible that the ADL
243 | Page Daniel J. Hoyle dependency levels may be underestimated by the RACF staff in cases where the residents perform the ADLs without the knowledge of the staff. However, it was not considered feasible to include these questions in resident interviews considering that most residents did not have the cognitive levels sufficient to provide informed responses.
Last, the PSMS tool is not designed specifically for use in the RACF setting. Consequently, some of the ADLs assessed in this tool are not possible to perform independently considering the setting rather than the residents’ ability to complete them (e.g., the resident may not shower themselves independently because RACF staff may be required to assist with this process due to facility policies).
6.6.5 Conclusion
The small amount of time that RACF staff have available to assist residents with ADLs is of great concern in Australian RACFs.668 Therefore, it is important that interventions to improve antipsychotic and benzodiazepine use do not lead to an increase in the amount of assistance residents require with ADLs. This study found no evidence that suggests antipsychotic and benzodiazepine dose reduction within a sample of residents involved in RedUSe significantly increases ADL dependency levels. Whilst there were tendencies for ADL dependency levels to increase with antipsychotic dose reduction and decrease with benzodiazepine dose reduction, the small change in the PSMS score is unlikely to be of clinical significance.
244 | Page Daniel J. Hoyle 6.7 Risk assessment
6.7.1 Introduction
Residents involved in our study were continuously monitored by RACF staff to investigate the relationship between antipsychotic and/or benzodiazepine dose change and falls, behavioural events, and usage of medical resources (such as GP consultations and hospitalisations). Figure
52 illustrates where these findings fit in this thesis. It needs to be noted that limitations of this study, most notably the absence of a control group and pre-intervention data, prevent a more comprehensive analysis from being carried out.
Figure 52: Flowchart of the analysis for the clinical outcomes study. This sub-chapter is dedicated to investigating the association between antipsychotic and benzodiazepine dose reduction and the rate of falls, behavioural events, General Practitioner (GP) visits and hospitalisations.
6.7.1.1 Falls
A fall is defined by the World Health Organization as “an event which results in a person coming to rest inadvertently on the ground or floor or other lower level”.266 Fall rates vary considerably in RACFs depending on case mix. Between 30 and 70% of RACF residents fall at least once per year.267 Falls in older people are associated with potentially serious physical consequences, including fracture, head trauma, soft-tissue injuries and severe lacerations.268 In developed
245 | Page Daniel J. Hoyle countries, falls account for 85% of injury-related hospital admissions in people over 65 years.269
Hip fractures account for the majority of injuries.270 Over 70% of people report mobility issues after 120 days following a hip fracture.271 Consequently, this can lead to a disability, loss of independence, fear of falling and decreased QoL.272, 669 Importantly, there is evidence of increased mortality with falls. In a prospective study of 300 Norwegian women, the relative risk of death among people experiencing at least two falls during one year was 1.6 (95%CI=1.1, 2.4) compared to those who did not have a fall.272
The risk factors for falls are multifactorial.268 The most common risk factors include gait and balance instability, visual impairment, cognitive and functional impairment and psychotropic medication.268 As discussed in Chapter 3, systematic reviews and meta-analyses consistently demonstrate an increased risk of falls with the use of antipsychotics and benzodiazepines in older populations.195, 278-281 Consequently, recommendations have been made to reduce antipsychotic and benzodiazepine prescribing to minimise the risk of falls.195
Whilst there is evidence of an increased risk of falls with the use of antipsychotics and benzodiazepines in older populations,195, 278-281 it is unclear whether reducing these medications reduces the rate of falls. It is possible that the underlying condition (e.g., dementia) may increase the risk of falls. Therefore, removal of the psychotropic may not address this underlying risk factor.
Only two antipsychotic reduction interventions, discussed in Chapter 4, have reported on the impact on falls.517, 530 The 2006 RCT performed by Fossey et al. reported no difference in the proportion of residents who fell at least once between the intervention and control groups.517 On the other hand, Brodaty et al. reported a statistically non-significant decrease in the proportion of residents who had at least one fall in the six months prior to the implementation of the HALT intervention compared to six months post-intervention.530 Whilst we did not identify any interventions (similar to RedUSe) to reduce benzodiazepine use that
246 | Page Daniel J. Hoyle reported on the rate of falls,47 a small benzodiazepine withdrawal study published in 2018, involving 31 French RACF residents, found a statistically significant reduction in falls among residents who were completely withdrawn from their benzodiazepine in the six months pre- intervention compared to the six months post-intervention (2.3±0.5 versus 0.5±0.2 falls/resident, p=0.01).670 Furthermore, in the outpatient setting, an RCT of psychotropic medication withdrawal and a home-based exercise program to prevent falls among 93 women and men aged 65 years and over reported a significant reduction in the risk of falls over 44 weeks.671
On the other hand, an increased prevalence of fractures has been associated with reduced benzodiazepine use following Medicaid funding changes in the US.672 Commentary on this article suggested that reduced benzodiazepine use may have led to increased motivation of previously sedated residents to perform ADLs, which, in turn, increased their risk of falls and fractures.673
Considering the significant impact that falls can have on morbidity, mortality and health care costs, and the lack of investigation on the rate of falls within antipsychotic and/or benzodiazepine dose reduction interventions, we investigated relationships between the rate of falls and magnitude of antipsychotic and/or benzodiazepine dose change.
6.7.1.2 Behavioural events
As discussed in Chapter 3, the perception of worsening in neuropsychiatric symptoms is a barrier to antipsychotic and benzodiazepine dose reduction.44, 45, 170, 491 In Chapter 6.3, we found no evidence of worsening in neuropsychiatric symptoms between baseline and four months.
However, the tools used in Chapter 6.3 assessed the neuropsychiatric symptoms over the previous two weeks. Consequently, these tools may be influenced by the recall of the person rating the neuropsychiatric symptoms at that time point.126
247 | Page Daniel J. Hoyle Considering that a worsening in neuropsychiatry symptoms is a barrier to antipsychotic and benzodiazepine dose reduction,44, 45, 170, 491 we investigated the relationship between the rate of behavioural events reported by RACF staff and the magnitude of antipsychotic and/or benzodiazepine dose change.
6.7.1.3 General practitioner consultations
In our systematic review in Chapter 4,47 only one study reported on the association between reduced antipsychotic use and the number of GP consultations.520 The intervention provided psychiatric support to six RACFs (n=208 residents) resulting in reduced antipsychotic use from
44% to 28% (p=0.07) over nine months.520 Ballard et al. reported a lower number of GP consultations in the intervention group compared to the control group (3.4±3.2 versus 5.4±4.8, p=0.0001).520 Whilst this study highlighted the possibility for reduced GP visits with an antipsychotic reduction program, it is unclear whether a similar outcome would be identified in an antipsychotic dose reduction program that does not include psychiatric support. Additionally, to our knowledge, no previous interventions to reduce benzodiazepine prescribing in RACFs have reported on changes in GP consultations.
It is possible that an antipsychotic and/or benzodiazepine reduction intervention that does not provide ongoing psychiatric support may result in increased GP consultations due to the need for closer monitoring during the dose reduction period to manage withdrawal effects and the potential return or worsening of neuropsychiatric symptoms. Equally, a reduction in adverse effects from these medications may decrease the need for GP consultations.
As discussed in Chapter 2, access to GPs is progressively becoming more difficult in
Australian RACFs. Therefore, it is important that interventions in the RACF setting do not result in increased demand for GP consultations. To address this issue, we investigated the relationship between the rate of GP visits and the magnitude of antipsychotic and/or benzodiazepine dose change.
248 | Page Daniel J. Hoyle 6.7.1.4 Hospitalisations
RACF residents often suffer from multiple conditions, polypharmacy, and have functional and cognitive impairment.7, 674, 675 Therefore, it is not surprising that residents of RACFs have high rates of hospitalisations.676, 677 A systematic review of 21 studies reported hospitalisation rates between 6.8 and 45.7% for various time periods of follow-up ranging from three months to six years.676 Another systematic review (n=27 studies) found that most studies had at least 30 transfers from the RACF to the emergency department for every 100 beds/year.677
Hospitalisations for RACF residents is associated with high costs to the government and adverse health consequences, such as greater cognitive impairment and functional decline or hospital-acquired infections.678, 679 In a study of 277 residents with dementia from 16 UK RACFs,
Romeo et al. reported that hospital admissions account for 50% of costs, after excluding RACF fees.680 Furthermore, the authors reported that the use of psychotropic medication was associated with increased hospital costs.680 Psychotropic medication, such as antipsychotics and benzodiazepines, are associated with an increased risk of falls and fractures, pneumonia, stroke, and delirium, which may lead to hospital admission.195, 379, 681, 682
Given that antipsychotics and benzodiazepines may contribute to an increased risk of hospitalisation, it would follow that a reduction in their use may reduce the hospitalisation rates.
In Chapter 4, we identified three studies that have reported on the impact reduced antipsychotic use has on hospitalisations.519, 520, 530 A 1992 educational intervention to reduce antipsychotic use found no difference in hospitalisation rates between intervention and control
RACFs.519 In 2002, Ballard et al. found that residents of RACFs receiving a psychiatric liaison intervention spent less time, albeit non-significantly, admitted to a psychiatric inpatient facility compared to residents of control RACFs (0.6±8.3 vs 1.5±9.6, p=0.19). However, it is unclear from this study whether antipsychotic dose reduction had any impact on other reasons for hospitalisation, besides a psychiatric admission.520 More recently, Brodaty et al. reported on the
249 | Page Daniel J. Hoyle effect HALT (an antipsychotic discontinuation intervention) had on the hospitalisation rates of
93 residents.530 Over 12 months, the proportion of residents who were hospitalised at least once decreased from 18.8% (in the six months pre-intervention) to 8.8% (in the six months post- intervention).530
Considering the high costs and clinical consequences of hospitalisation on RACF residents, and limited evidence suggesting reduced hospitalisations with antipsychotic dose reduction, we aimed to investigate relationships between the magnitude of antipsychotic and/or benzodiazepine dose change and the rate and duration of hospitalisations.
6.7.2 Methods
6.7.2.1 Data collection
The champion nurse at each RACF was requested to collect data on the falls, behavioural events,
GP consultations and hospitalisations incurred by residents consenting to the clinical outcomes study between baseline and four months. In addition to the date and resident involved in the event, further details were recorded by the champion nurse as per Table 50.
Table 50: Information collected on falls, behavioural events, general practitioner (GP) consultations and hospitalisations.
Event Information collected Falls Severity of fall Behavioural events Severity of behavioural event GP consultations Reason for GP consultation Hospitalisations Reason for hospitalisation Duration of hospitalisation
To indicate the severity of falls and behavioural events the champion nurse was required to tick a box on the data collection form (Appendix F) that most closely described the severity of the event (Table 51).
250 | Page Daniel J. Hoyle Table 51: Falls and behavioural events severity assessment.
Severity Falls Behavioural events Examples Examples Minor Minor injury that does not require Managed in-house – No addition significant medical intervention (e.g., treatment needed (e.g., excessive small skin tear, minor bruise) wandering, calling out) Moderate Injury requiring GP or outpatient GP needed or treatment started (e.g., management and not resulting in self-harm or struck staff/other permanent disability residents) Severe Severe injury resulting in hospitalisation Specialist visit and requiring substantial rehabilitation Hospitalisation required GP=General Practitioner
Progress notes and GP consultation records for all residents were audited at the four- month data collection time point by the thesis author to ensure accuracy and completeness of the data. Discrepancies were discussed with the champion nurse and records were amended where necessary.
As mentioned in Chapter 5, due to issues with data collection (i.e., lack of time for champion nurses to collect information, reliance on filters in electronic records), the majority of data was ultimately collected by the thesis author through comprehensive auditing of the resident’s progress notes and GP consultation records.
As discussed in Chapter 6.1, the follow-up time varied slightly between residents. To standardise the number of falls, behavioural events, GP consultations and hospitalisations, the total number of events (i.e., falls, behavioural events, GP consultations, hospitalisations) recorded were divided by the number of days that the resident had data collected for and multiplied by 30 to give the average number of events residents had over a period of 30 days.
This method of standardisation was repeated for minor, moderate and severe falls and behavioural events. Additionally, the length of hospital admissions was summed to give a total number of days that residents were admitted to hospital for the study period. The proportion
(%) of the data collection period the resident spent in hospital was then calculated.
251 | Page Daniel J. Hoyle 6.7.2.2 Statistical analysis
Given RedUSe aimed to improve prescribing of antipsychotic and benzodiazepine medication, two separate sets of analyses were run; one for antipsychotic users (those with a chlorpromazine daily dose equivalent >0 mg at baseline) and one for benzodiazepine users
(those with a diazepam daily dose equivalent >0 mg at baseline). The distribution of the rate of falls, behavioural events, GP consultations and hospitalisations was investigated using beeswarm plots for residents initially prescribed an antipsychotic or benzodiazepine.
Correlations between change (%) in the chlorpromazine daily dose equivalent (for residents initially taking antipsychotics) or diazepam daily dose equivalent (for residents initially taking benzodiazepines) and the rate of falls, behavioural events, GP consultations and hospitalisations per 30 days were investigated using Spearman’s correlation tests (rs). Similarly, correlations between change (%) in the chlorpromazine daily dose equivalent (for residents initially taking antipsychotics) or diazepam daily dose equivalent (for residents initially taking benzodiazepines) and the proportion of data collection period spent in hospital among residents with at least one hospitalisation was investigated using Spearman’s correlation tests.
Falls data were re-analysed taking into account mobility. Residents who were not able to ambulate without assistance (e.g., bed-bound residents) were excluded from the re-analysis.
6.7.3 Results
6.7.3.1 Feasibility of monitoring falls, behavioural events, GP consultations and hospitalisations associated with antipsychotic and/or benzodiazepine dose change
Overall, it appeared feasible to monitoring the number of falls, behavioural events, GP consultations and hospitalisations associated with antipsychotic and/or benzodiazepine dose change; data were available for all participants who were present in the study at baseline and four months.
252 | Page Daniel J. Hoyle 6.7.3.2 Documented events
Thirty-one of 83 residents (36.2%) initially taking antipsychotics and 41 of 118 residents (34.8%) initially taking benzodiazepines fell during the study period.
Residents taking antipsychotics fell an average of 0.31 times (95%CI=0.17, 0.45) every
30 days. Thirty residents had a minor fall (mean=0.28 falls/30 days, 95%CI=0.14, 0.41), six residents had a moderate fall (mean=0.02 falls/30 days, 95%CI=0.00, 0.04) and two residents had a severe fall over the study period (mean=0.01 falls/30 days, 95%CI=-0.00, 0.02) (Figure 53).
Residents taking benzodiazepines fell an average of 0.23 times (95%CI=0.14, 0.33) every
30 days. Thirty-eight residents had a minor fall (mean=0.22 falls/30 days, 95%CI=0.12, 0.31), five residents had a moderate fall (mean=0.01 falls/30 days, 95%CI=0.00, 0.02) and three residents had a severe fall over the study period (mean=0.01 falls/30 days, 95%CI=-0.00, 0.02)
(Figure 53).
Figure 53: Number of falls per 30 days, stratified by severity, for residents initially taking (A) antipsychotics or (B) benzodiazepines.
Overall, 60 of 83 residents (72.3%) initially taking an antipsychotic and 67 of 118 residents (56.8%) initially taking a benzodiazepine had at least one documented behavioural event.
253 | Page Daniel J. Hoyle Residents taking antipsychotics had an average of 2.66 documented behavioural events
(95%CI=1.66, 3.65) every 30 days. Fifty-eight residents had a minor event (mean=1.99 events/30 days, 95%CI=1.17, 2.80), 24 residents had a moderately-severe behavioural event (mean=0.67 events/30 days, 95%CI=0.21, 1.13) and none of the residents had a documented severe behavioural event (Figure 54).
Residents taking benzodiazepines had an average of 1.22 documented behavioural events (95%CI=0.73, 1.72) every 30 days. Sixty-five residents had a minor event (mean=0.81 events/30 days, 95%CI=0.50, 1.12), 24 residents had a moderately-severe behavioural event
(mean=0.41 events/30 days, 95%CI=0.10, 0.72) and none of the residents had a documented severe behavioural event (Figure 54).
Figure 54: Number of behavioural events per 30 days, stratified by severity, for residents initially taking (A) antipsychotics or (B) benzodiazepines.
Seventy-nine of 83 residents (95.2%) initially taking an antipsychotic and 115 of 118 residents (97.5%) initially prescribed a benzodiazepine had at least one documented GP consultation.
Residents taking an antipsychotic had an average of 1.60 documented GP consultations
(95%CI=1.31, 1.89) every 30 days. Whereas, residents initially taking a benzodiazepine had an average of 1.54 documented GP consultations (95%CI=1.31, 1.77) every 30 days (Figure 55).
254 | Page Daniel J. Hoyle
Figure 55: Number of general practitioner (GP) consultations per 30 days for residents initially taking (A) antipsychotics or (B) benzodiazepines.
Eleven of 83 residents (13.3%) initially taking an antipsychotic had at least one hospitalisation, equating to an average of 0.04 hospitalisations (95%CI=0.02, 0.07) every 30 days
(Figure 56). Residents with at least one hospitalisation spent, on average, 5.11% (95%CI=0.34,
9.89) of the data collection period in hospital (Figure 57).
Similarly, 15 of 118 residents (12.7%) initially taking a benzodiazepine had at least one hospitalisation, equating to an average of 0.04 hospitalisations (95%CI=0.02, 0.06) every 30 days
(Figure 56). Residents with at least one hospitalisation spent, on average, 7.70% (95%CI=3.51,
11.88) of the data collection period in hospital (Figure 57).
255 | Page Daniel J. Hoyle
Figure 56: Number of hospitalisations per 30 days for residents initially taking (A) antipsychotics or (B) benzodiazepines.
Figure 57: Proportion (%) of data collection period spent in hospital for residents initially taking (A) antipsychotics or (B) benzodiazepines who had at least one hospital stay.
6.7.3.3 Relationships between documented events and dose change
No meaningful correlations were detected between the magnitude of antipsychotic and benzodiazepine dose change and the rate of falls, behavioural events, GP consultations and hospitalisations (Table 52).
256 | Page Daniel J. Hoyle Re-analysis of falls data, after excluding residents taking antipsychotics who were unable to mobilise independently (n=26), found no appreciable correlation between the magnitude of antipsychotic dose change and the rate of falls (rs=-0.17, p=0.195). However, after excluding residents taking benzodiazepines who were unable to mobilise independently (n=22), a greater magnitude of benzodiazepine dose reduction was weakly correlated with a statistically significant lower rate of falls (rs=0.20, p=0.049).
Table 52: Correlation (rs) between change (%) in antipsychotic and benzodiazepine dose and the rate of falls, behavioural events, general practitioner (GP) consultations and hospitalisations per 30 days.
Total events per 30 Medication class rs value p-value days Falls Antipsychotic -0.04 0.713 Benzodiazepine 0.13 0.152 Behavioural events Antipsychotic -0.07 0.527 Benzodiazepine 0.10 0.292 GP consultations Antipsychotic -0.16 0.144 Benzodiazepine 0.04 0.693 Hospitalisations Antipsychotic -0.04 0.715 Benzodiazepine 0.14 0.143
Correlations between antipsychotic and benzodiazepine dose reduction and the proportion of data collection period spent in hospital are reported in Table 53. A weak, albeit non-significant, negative correlation was detected between antipsychotic dose change and the proportion of data collection period spent admitted to hospital (i.e., antipsychotic dose reductions were associated with a higher proportion of time spent in hospital) for residents with at least one hospitalisation (n=11). Conversely, a weak, albeit non-significant, positive correlation was detected between benzodiazepine dose change and the proportion of data collection period spent in hospital (i.e., benzodiazepine dose reductions were associated with a smaller proportion of time spent in hospital), for residents with at least one hospitalisation
(n=15).
257 | Page Daniel J. Hoyle Table 53: Correlation (rs) between change (%) in antipsychotic and benzodiazepine dose and proportion of time spent in hospital for residents who had at least one hospital admission.
Outcome Medication class rs value p-value Proportion of data Antipsychotic -0.27 0.417 collection period Benzodiazepine 0.35 0.195 spent in hospital
6.7.4 Discussion
This study aimed to explore relationships between antipsychotic and/or benzodiazepine dose change and the rate of falls, behavioural events, GP consultations and hospitalisations, and the duration of hospitalisations. It was found that antipsychotic and benzodiazepine dose change was not associated with differences in the rate of falls, behavioural events, GP consultations and hospitalisations, or duration of hospitalisations. However, a weak but statistically significant correlation between benzodiazepine dose reduction and a lower rate of falls was detected among mobile residents.
6.7.4.1 Falls
The absence of a correlation between antipsychotic dose reduction and the rate of falls aligns with results from earlier interventions.517, 530 Whilst antipsychotic use is associated with an increased risk of falls,195, 278-281 there are several factors that may explain why the rate of falls was not lower in those who had their antipsychotic dose reduced.
First, falls are multifactorial.268 Gait and balance instability, visual impairment, cognitive and functional impairment, and psychotropic medication are common risk factors for falls.268
Considering that antipsychotics are regularly prescribed to residents with dementia who often have high degrees of cognitive and functional impairment, it is possible that these other factors may have had a greater influence on the rate of falls. Second, insufficient time may have passed to observe an effect of antipsychotic dose reduction on falls. A study (n=594 residents) by Berry et al. found no evidence of a change in acute fall rates following changes in the antipsychotic dose.416 The researchers suggested that long periods of time are required for antipsychotic dose
258 | Page Daniel J. Hoyle changes to affect the rate of falls through their effects on gait and orthostasis.416 Third, residents who had their antipsychotic dose continued or increased may have had more severe neuropsychiatric symptoms than those who had their dose reduced. Consequently, these residents may have received closer staff supervision providing less opportunity for residents to fall.
Our finding of a relationship between a lower rate of falls among mobile residents with a greater magnitude of benzodiazepine dose reduction is consistent with an earlier study.670 In
2018, Javelot et al. reported on the number of falls in the six months before and after implementing a benzodiazepine withdrawal intervention in 31 residents of a French RACF.670
Overall, there was a significant reduction in the number of falls (2.3±0.6 versus 0.5±0.2 falls, p=0.01) in the 11 residents who had their benzodiazepine withdrawn.670 Whilst differences in the analyses make comparison difficult, it should be noted that Javelot et al.’s study was restricted to residents taking a benzodiazepine as an anxiolytic.670 On the other hand, we included residents who were using benzodiazepines for a variety of indications (besides those with a severe psychiatric condition). Our results are also consistent with evidence from Berry et al. who reported a significant reduction in falls risk immediately following benzodiazepine withdrawal.416 Together, these findings provide evidence of a possible reduction in falls with benzodiazepine dose reduction among mobile residents.
6.7.4.2 Behavioural events
Our findings, that antipsychotic and benzodiazepine dose reduction was not associated with an increase in the rate of behavioural events align with findings from earlier antipsychotic dose reduction studies (see Chapter 4) and our results attained from the NPI-NH and CMAI scales in
Chapter 6.3.
This result provides further support that antipsychotic and benzodiazepine dose reduction generally does not negatively affect the incidence of behavioural events.
259 | Page Daniel J. Hoyle 6.7.4.3 General practitioner consultations
Our findings that antipsychotic dose reduction was not associated with differences in the rates of GP visits conflicts with findings from another antipsychotic dose reduction study.520 In 2002,
Ballard et al. reported fewer GP visits for residents residing in RACFs exposed to a psychiatric liaison intervention when compared to residents from control RACFs.520 The psychiatric liaison intervention allowed RACF staff to refer the residents directly to the psychiatric liaison team, reducing the need for close supervision by the resident’s usual GP.520 Conversely, RedUSe aimed to optimise existing QUM practices in Australia and, thus, GPs were responsible for the monitoring of the resident during antipsychotic dose reduction.50 Additionally, the reduction in
GP visits in the comparatively longer study (nine months versus four months) may be associated with a reduction in long-term antipsychotic adverse effects among residents who had their antipsychotic withdrawn.520
To our knowledge, this is the first benzodiazepine dose reduction intervention to investigate the rate of GP visits. Overall, we found no evidence of differences in the rate of GP visits based on benzodiazepine dose change. Whilst a greater rate of GP visits may have been expected among residents who had their benzodiazepine dose reduced to monitor for benzodiazepine withdrawal symptoms, it is possible that if the symptoms did worsen, they may have been managed in-house (e.g., transient increase in sleep disturbances with benzodiazepine dose reduction may have been managed by non-pharmacological strategies).
Equally, one may expect a lower rate of GP visits with benzodiazepine dose reduction due to a reduction in adverse effects. However, our study’s timeframe may have been insufficient to capture these changes.
6.7.4.4 Hospitalisations
Our finding of no appreciable correlation between the magnitude of antipsychotic dose change and rate of hospitalisations aligns with findings from an earlier antipsychotic reduction
260 | Page Daniel J. Hoyle intervention.519 However, in a recent antipsychotic deprescribing intervention, called HALT, the proportion of residents having at least one hospitalisation reduced by over half between the six-months pre-intervention (18.8%) and the six-months post-intervention (8.8%).530 Whilst methodological differences between HALT and our study prevent a direct comparison, our findings together with those reported by Brodaty et al.530 and Avorn et al.519 suggest that antipsychotic dose reduction/withdrawal is not associated with increased hospitalisations.
The magnitude of antipsychotic dose reduction was also not significantly correlated with the proportion of data collection period spent in hospital. This result aligns with an earlier psychiatric liaison intervention which reported no statistically significant differences in the duration of psychiatric hospital admissions compared to a control.520 This addition to the literature further supports the evidence that antipsychotic use can be reduced without an increased need for hospitalisations. Whilst it is possible that a reduction in adverse effects (e.g., reduced falls over the long-term) may lead to reduced hospitalisations, larger longitudinal studies of greater duration are required to investigate this further.
There was no appreciable correlation between benzodiazepine dose reduction and the rate of hospitalisations or proportion of data collection period spent in hospital. Whilst a reduction in the rate of hospitalisation may have been expected considering the statistically significant relationship between a reduced rate of falls and a greater magnitude of benzodiazepine dose reduction among ambulating residents, most falls resulted in little to no harm (i.e., most falls were of minor severity) and, consequently, the number of residents requiring hospitalisation due to a fall-related injury is likely low. Overall, the results suggest that benzodiazepine dose reduction within this study was not related to an increase in hospitalisations. However, larger longitudinal studies of greater duration are needed given the possibility of reduced fall-related injuries.
261 | Page Daniel J. Hoyle 6.7.4.5 Strengths and limitations
This research is one of few to investigate the impact that antipsychotic and/or benzodiazepine dose reduction has on the rate of falls, behavioural events, GP consultations and hospitalisations, and duration of hospital admission. However, major limitations prevent firm conclusions from being made. Without a control group or a definitive ‘before’ and ‘after’ period it was not possible to assess whether there was a change in the number of events in individual residents. Instead, we had to compare between residents. Doing this, however, does not control for resident characteristics which may increase a resident’s propensity for an event regardless of changes in the antipsychotic and/or benzodiazepine dose. Future studies should optimally utilise matched control groups or at least measure the number of falls, behavioural events, GP consultations and hospitalisations in a defined period prior to the intervention and after the intervention to allow for comparison of the change in events rather than the raw number of events or rate of events.
Second, residents had their antipsychotic and/or benzodiazepine dose reduced at variable times throughout the study. Furthermore, some residents had failed dose reduction attempts and were restarted with their original dose. This creates uncertainty as to whether the resident had their dose reduced prior to an event occurrence or whether a resident who was taking the same dose at the end of the study had an event during a dose reduction attempt.
Whilst efforts were made to record dates in which dose changes were made throughout the study period, the data were not consistently available (i.e., due to medication chart archiving and/or missing medication charts). Future studies need to consider methods to account for variability in the date of dose changes when investigating objective outcomes, such as falls and behavioural events, within real-world interventions.
Third, the reporting of events in the ‘risk assessment’ booklet was inconsistent between and within RACFs. Whilst the RACF records were reviewed for the aforementioned events, we
262 | Page Daniel J. Hoyle cannot rule out that these events were not recorded in all cases by the RACF staff or the resident’s GP. To illustrate, a study of 153 patients in a Queensland hospital reported missing falls data from hospital incident reporting systems.683 In this study, only 75.5% of falls were reported in the incident reporting system.683
Fourth, the variation in the reporting of events was extreme. It appears that standard classification for these clinical outcomes does not exist. This may have resulted in under- and over-reporting of these outcomes by nursing staff between RACFs and within the same RACF.
For example, some nursing staff classify a ‘slip’ as a fall, but other staff may not report a fall unless an assessment by a GP is required. Whilst defined criteria for the severity of falls and behavioural events were provided, auditing of RACF notes revealed inconsistencies in reporting by the champion nurses. These inconsistencies reflect the high workload of the champion nurses and the fact that nurses are not trained in research methods, thus, are not aware of the importance of consistency with ranking. It needs to be acknowledged, however, that RedUSe was a pragmatic implementation study designed for reach as opposed to strict research methodology. Thus, we had to make avail of the resources available.
6.7.5 Conclusion
We found no evidence to suggest that antipsychotic and/or benzodiazepine dose reduction is related to an increased rate of falls, behavioural events, GP consultations or hospitalisations.
Moreover, we report on evidence of a lower rate of falls with a greater magnitude of benzodiazepine dose reduction in mobile residents. Larger prospective studies utilising a controlled design and/or standardised pre- and post-intervention data collection periods are needed to further investigate the impact antipsychotic and/or benzodiazepine dose reduction has on falls, behavioural events, GP consultations and hospitalisations.
263 | Page Daniel J. Hoyle 6.8 Staff-related outcomes
6.8.1 Introduction
Caring for residents of RACFs is a demanding occupation that is generally undervalued by society.
RACFs are known to operate in resource-poor conditions, leading to high workloads and time pressures for staff.170 Additionally, poor levels of remuneration have contributed to difficulties recruiting and retaining staff.65, 141, 158, 160 The pressure of working in this environment is further complicated by neuropsychiatric symptoms of residents, with more severe behaviours linked to an increase in time needed to provide direct care.662 Furthermore, RACF staff, particularly PCAs who provide most of the direct care, often have limited training on the management of older age mental health (see Chapter 2.2.5).65, 139, 170, 491 These short-falls can lead to the inappropriate utilisation of psychotropic medication. A 2018 study using semi-structured interviews of 40 staff from eight Sydney RACFs found that inappropriate psychotropic medication use was often the result of staff feeling helpless to provide person-centred care for neuropsychiatric symptoms due to the working environment.43
These symptoms, especially agitation/aggression, disinhibition, and irritability/lability,144 may also be distressing to staff and lead to lower quality of health, increased sick days, reduced workability, injury, high burnout levels and staff turnover.179, 180, 684
Consequently, high staff turnover can have negative implications for the residents of RACFs,153,
154 such as an increased risk of infections and hospitalisations,154 and worsening of behavioural issues.155
Explorations of barriers to antipsychotic and benzodiazepine withdrawal within Belgian
RACFs have revealed concerns related to increased workloads and higher job demands among nursing staff.44, 45 Similar barriers were also reported within the first iteration of RedUSe.497
Despite this, two antipsychotic dose reduction studies included in our systematic review
(Chapter 4) found that the number of stressful events reported by RACF staff remained stable
264 | Page Daniel J. Hoyle or reduced during the study.498, 499 It should be noted, though, that both studies were conducted in Canada and may not reflect the work environment and possible effect that antipsychotic dose reduction may have in the Australian RACF setting. Secondly, whilst there is evidence that severe night-time disturbances are one of the most distressing neuropsychiatric symptoms for
RACF staff,144 there has been a lack of studies investigating the effect of benzodiazepine dose reduction on RACF staff distress.
The job satisfaction of RACF staff should also be considered. Improved job satisfaction may improve staff retention rates and result in improvements in the satisfaction of residents and relatives and reduction in the risk of adverse outcomes, such as falls, weight loss, and pressure ulcers.685-689 Whilst none of the interventions included in our systematic review
(Chapter 4) reported on changes in job satisfaction, it is possible that antipsychotic and/or benzodiazepine dose reduction interventions may also lead to improvements in job satisfaction.
Feelings of control and support and use of person-centred care in such interventions have the potential to improve job satisfaction.690, 691 As an example, an intervention to improve the management of neuropsychiatric symptoms in 17 Dutch special care dementia units reported small improvements in job satisfaction without an increase in job demands.43, 692
For antipsychotic and/or benzodiazepine dose reduction interventions to be implemented successfully, interventions should demonstrate that they do not over-burden
RACF staff or result in lower job satisfaction. Considering the lack of studies investigating staff- related outcomes upon antipsychotic and/or benzodiazepine dose reduction, occupational disruptiveness and job satisfaction were investigated as secondary outcomes in this study
(Figure 58). Overall, the focus was to determine whether the magnitude of antipsychotic and/or benzodiazepine dose change was associated with changes in occupational disruptiveness related to neuropsychiatric symptoms in residents recruited into this study or whether the
265 | Page Daniel J. Hoyle intervention had significant effects on job satisfaction measures for the staff employed in the
28 RACFs involved in this study.
Figure 58: Flow of the analysis for the clinical outcomes. This sub-chapter is dedicated to investigating the association between antipsychotic and benzodiazepine dose reduction and change in occupational disruptiveness, and differences in job satisfaction between baseline and four months. GP=General Practitioner. 6.8.2 Occupational disruptiveness related to neuropsychiatric symptoms
6.8.2.1 Methods
6.8.2.1.1 Occupational disruptiveness scoring
Occupational disruptiveness associated with neuropsychiatric symptoms was assessed at baseline and four months, through psychometric testing of RACF nursing and care staff, using the NPI-NH occupational disruptiveness scale (Chapter 0); a component of the overall NPI-NH psychometric tool. Occupational disruptiveness is a measure of increased work, effort, time or distress the neuropsychiatric symptom causes to the caregiver (if any).525 The overall occupational disruptiveness score was calculated through the summation of occupational disruptiveness scores for each NPI-NH domain. The overall score ranged from 0 to 60 and domain scores ranged from 0 to 5, where higher scores represented greater occupational disruptiveness.525
266 | Page Daniel J. Hoyle 6.8.2.1.2 Statistical analysis
Given RedUSe aimed to reduce inappropriate prescribing of antipsychotic and benzodiazepine medication, two separate sets of analyses were run; one for antipsychotic users (those with a chlorpromazine daily dose equivalent >0 mg at baseline) and one for benzodiazepine users
(those with a diazepam daily dose equivalent >0 mg at baseline). The distribution of the total and domain NPI-NH occupational disruptiveness scores were initially investigated using beeswarm plots. Paired t-tests were utilised to detect statistically significant changes in scores between baseline and four months.
To investigate the relationship between the magnitude of antipsychotic and benzodiazepine dose changes and changes in the occupational disruptiveness total and domain scores, linear regression models were fitted with the ‘lm’ function in R.594 Potentially confounding variables, identified in Chapter 6.2, were controlled for within the models, where appropriate. In the benzodiazepine users’ model, we controlled for the potentially confounding variable ‘documented diagnosis of dementia’. From the output of these models, it was possible to estimate the unit change in the occupational disruptiveness scores associated with the percentage change in the antipsychotic and benzodiazepine doses between baseline and four months.
Inspection of Q-Q plots of the residuals revealed approximately normal distributions of the residuals with small violations of this assumption in the tails. To account for this violation, confidence intervals were bootstrapped. Plots of the total and domain occupational disruptiveness scores and antipsychotic and benzodiazepine doses at baseline and four months were inspected for non-linear distribution of the data.
Lastly, the change in the total occupational disruptiveness score was assessed for residents who completely ceased their antipsychotic or benzodiazepine between baseline and four months, using paired t-tests.
267 | Page Daniel J. Hoyle 6.8.2.2 Results
6.8.2.2.1 Antipsychotic users
6.8.2.2.1.1 Total occupational disruptiveness
The disruptiveness of neuropsychiatric symptoms was assessed at baseline and four months by
RACF nursing and care staff for 83 residents initially taking antipsychotics using the NPI-NH psychometric tool.
The total occupational disruptiveness score did not change significantly between baseline (mean=10.40, 95%CI=8.05, 12.75) and four months (mean=10.98, 95%CI=8.94, 13.01) among residents initially taking antipsychotics (t(82)=-0.497, p=0.621; Figure 59).
Figure 59: Beeswarm plot of the NPI-NH total occupational disruptiveness scores at baseline and four months for residents taking an antipsychotic at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval. NPI-NH=Neuropsychiatric Inventory-Nursing Home version.
Using linear regression modelling, there was no evidence of an appreciable association between the magnitude of antipsychotic dose change and change in occupational disruptiveness (β=0.012; i.e., total NPI-NH occupational disruptiveness score was 0.12 points
268 | Page Daniel J. Hoyle lower, on average, for each 10% reduction in the chlorpromazine daily dose equivalent, p=0.547,
95%CI=-0.028, 0.052; Table 54, Figure 60).
Furthermore, the occupational disruptiveness score remained constant between baseline (mean=10.3, 95%CI=4.0, 12.8) and four months (mean=10.3, 95%CI=5.6, 15.1) in the
18 residents who had their antipsychotic ceased (t(17)=-0.720, p=0.482).
Table 54: Estimated changes in the NPI-NH total occupational disruptiveness score associated with changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value
Intercept 0.834 -1.639, 3.307 0.671 0.504 Antipsychotic dose change (%) 0.012 -0.028, 0.052 0.604 0.547
2 NH total
- Adjusted R -0.008 NPI
occupational 2
disruptiveness Multiple R 0.005 NPI-NH=Neuropsychiatric Inventory-Nursing Home version, 95%CI=95% confidence interval
Figure 60: Scatterplot of the change in the NPI-NH total occupational disruptiveness score versus change (%) in the chlorpromazine (CPZ) daily dose equivalent (DDE). The blue line represents the predicted change in the NPI-NH total occupational disruptiveness score based on the change (%) in the chlorpromazine daily dose equivalent. NPI- NH=Neuropsychiatric Inventory-Nursing Home version.
269 | Page Daniel J. Hoyle 6.8.2.2.1.2 Domain-specific occupational disruptiveness
Occupational disruptiveness scores related to most NPI-NH domains remained constant between baseline and four months for antipsychotic users (Table 55, Figure 61 and Figure 62).
However, there was a trend towards increased disruptiveness associated with appetite-related behaviours.
Table 55: Mean Neuropsychiatric Inventory-Nursing Home version domain-specific occupational disruptiveness scores at baseline and four months.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Delusions 0.9 (0.5, 1.2) 0.9 (0.6, 1.3) -0.17 82 0.868 Hallucinations 0.7 (0.4, 1.0) 0.6 (0.3, 0.8) 0.52 82 0.602 Agitation 1.6 (1.2, 1.9) 1.8 (1.4, 2.2) -1.14 82 0.257 Depression 0.9 (0.6, 1.2) 1.1 (0.8, 1.4) -1.47 82 0.146 Anxiety 1.2 (0.8, 1.5) 1.3 (0.9, 1.6) -0.68 82 0.500 Elation 0.2 (0.0, 0.3) 0.2 (0.0, 0.4) -0.34 82 0.735 Apathy 1.0 (0.7, 1.3) 1.0 (0.7, 1.3) 0.25 82 0.804 Disinhibition 0.6 (0.4, 0.9) 0.8 (0.5, 1.1) -0.76 82 0.452 Irritability 1.4 (1.1, 1.8) 1.4 (1.0, 1.7) 0.12 82 0.905 Aberrant motor behaviour 0.7 (0.4, 0.9) 0.6 (0.3, 0.8) 0.77 82 0.445 Night-time behaviour 1.1 (0.8, 1.4) 0.9 (0.5, 1.2) 1.09 82 0.281 Appetite 0.3 (0.1, 0.5) 0.6 (0.3, 0.9) -1.84 82 0.070 df=degrees of freedom, 95%CI=95% confidence interval
270 | Page Daniel J. Hoyle
Figure 61: Beeswarm plot of the Neuropsychiatric Inventory-Nursing Home version domain-specific occupational disruptiveness scores at baseline and four months for residents taking an antipsychotic at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval.
271 | Page Daniel J. Hoyle
Figure 62: Beeswarm plot of the Neuropsychiatric Inventory-Nursing Home version domain-specific occupational disruptiveness scores at baseline and four months for residents taking an antipsychotic at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval.
Using linear regression, we generally found no association between the magnitude of antipsychotic dose change and change in domain-specific occupational disruptiveness scores
(Table 56). However, there was a trend towards reduced occupational disruptiveness related to agitation with a greater magnitude of antipsychotic dose reduction (β=0.006; i.e., NPI-NH occupational disruption score related to agitation was 0.06 points lower, on average, for each
10% reduction in the chlorpromazine daily dose equivalent, p=0.074; 95%CI=-0.001, 0.012;
Table 56, Figure 63 and Figure 64).
272 | Page Daniel J. Hoyle Table 56: Estimated changes in the Neuropsychiatric Inventory-Nursing Home version domain-specific occupational disruptiveness scores associated with changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value
Intercept 0.050 -0.412, 0.512 0.216 0.829
Antipsychotic dose change (%) 0.001 -0.007, 0.008 0.178 0.859 related - OD Adjusted R2 -0.012
Delusion Multiple R2 0.000
Intercept -0.155 -0.495, 0.185 -0.909 0.366 -
Antipsychotic dose change (%) -0.003 -0.009, 0.002 -1.218 0.227
Adjusted R2 0.006 related OD Hallucination Multiple R2 0.018
Intercept 0.340 -0.057, 0.737 1.704 0.092
Antipsychotic dose change (%) 0.006 -0.001, 0.012 1.811 0.074 related - OD Adjusted R2 0.027
Agitation Multiple R2 0.039
Intercept 0.304 -0.043, 0.651 1.746 0.085
- Antipsychotic dose change (%) 0.003 -0.003, 0.009 1.067 0.289
Adjusted R2 0.002 related OD Depression Multiple R2 0.014
Intercept 0.109 -0.269, 0.487 0.573 0.568
Antipsychotic dose change (%) -0.000 -0.007, 0.006 -0.179 0.858 related - OD Adjusted R2 -0.012
Anxiety Multiple R2 0.000
Intercept 0.068 -0.157, 0.294 0.603 0.548
Antipsychotic dose change (%) 0.002 -0.002, 0.005 0.833 0.407
related OD 2 - Adjusted R -0.004
Multiple R2 0.009 Elation
273 | Page Daniel J. Hoyle Intercept 0.021 -0.390, 0.431 0.101 0.920
Antipsychotic dose change (%) 0.003 -0.003, 0.010 0.982 0.329
related OD 2 - Adjusted R -0.000
Multiple R2 0.012 Apathy
Intercept 0.098 -0.241, 0.437 0.574 0.568 -
Antipsychotic dose change (%) -0.001 -0.007, 0.004 -0.393 0.696
Adjusted R2 -0.010 related OD Disinhibition Multiple R2 0.002
Intercept 0.025 -0.404, 0.454 0.116 0.908
Antipsychotic dose change (%) 0.002 -0.005, 0.009 0.669 0.506 related
-
OD Adjusted R2 -0.007
Multiple R2 0.006 Irritability
Intercept -0.094 -0.361, 0.173 -0.701 0.486
Antipsychotic dose change (%) 0.000 -0.004, 0.004 0.050 0.960 related
-
OD Adjusted R2 -0.012
2 Aberrant motor Multiple R 0.000 behaviour
Intercept -0.243 -0.668, 0.183 -1.136 0.259
Antipsychotic dose change (%) -0.001 -0.008, 0.006 -0.356 0.722 related
- time -
OD Adjusted R2 -0.011 Night Multiple R2 0.002 behaviour
Intercept 0.311 -0.009, 0.631 1.933 0.057
Antipsychotic dose change (%) 0.002 -0.004, 0.007 0.618 0.538
related -
OD Adjusted R2 -0.008
2
Appetite Multiple R 0.005
OD=Occupational disruptiveness, 95%CI=95% confidence interval
274 | Page Daniel J. Hoyle
Figure 63: Scatterplots of the changes in the Neuropsychiatric Inventory-Nursing Home version (NPI-NH) domain- specific occupational disruptiveness scores versus change (%) in the chlorpromazine (CPZ) daily dose equivalent (DDE). The blue line represents the predicted change in the NPI-NH occupational disruptiveness score based on the change (%) in the CPZ DDE.
275 | Page Daniel J. Hoyle
Figure 64: Scatterplots of the changes in the Neuropsychiatric Inventory-Nursing Home version (NPI-NH) domain- specific occupational disruptiveness scores versus change (%) in the chlorpromazine (CPZ) daily dose equivalent (DDE). The blue line represents the predicted change in the NPI-NH occupational disruptiveness score based on the change (%) in the CPZ DDE.
6.8.2.2.2 Benzodiazepine users
6.8.2.2.2.1 Total occupational disruptiveness
The disruptiveness of neuropsychiatric symptoms was assessed at baseline and four months by
RACF nursing and care staff for 118 residents initially taking benzodiazepines using the NPI-NH psychometric tool.
276 | Page Daniel J. Hoyle Overall, the total occupational disruptiveness score declined, albeit non-significantly, between baseline (mean=7.80, 95%CI=6.20, 9.40) and four months (mean=6.70, 95%CI=5.30,
8.11) among residents initially taking benzodiazepines (t(117)=1.394, p=0.166; Figure 65).
Figure 65: Beeswarm plot of the NPI-NH total occupational disruptiveness scores at baseline and four months for residents taking a benzodiazepine at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval. NPI-NH=Neuropsychiatric Inventory-Nursing Home version.
Using linear regression modelling to control for a documented diagnosis of dementia, there was no evidence of an association between change in benzodiazepine dose and change in total occupational disruptiveness (β=0.015; i.e., total NPI-NH occupational disruptiveness score was 0.15 points lower, on average, for each 10% reduction in the diazepam daily dose equivalent, p=0.198, 95%CI=-0.008, 0.037; Table 57 and Figure 66).
On average, the total occupational disruptiveness score tended to decrease slightly between baseline (mean=8.8, 95%CI=4.5, 13.1) and four months (mean=6.0, 95%CI=3.2, 8.9) in the 25 residents who had their benzodiazepine ceased (t(24)=1.703, p=0.102).
277 | Page Daniel J. Hoyle Table 57: Estimated changes in the NPI-NH total occupational disruptiveness score associated with changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value
Intercept 0.193 -1.895, 2.281 0.183 0.855
Benzodiazepine dose change (%) 0.015 -0.008, 0.037 1.294 0.198
Documented diagnosis of -2.336 -5.441, 0.770 -1.490 0.139
NH total dementia - Adjusted R2 0.017 NPI occupational disruptiveness Multiple R2 0.034
NPI-NH=Neuropsychiatric Inventory-Nursing Home version, 95%CI=95% confidence interval
Figure 66: Scatterplot of the change in the NPI-NH total occupational disruptiveness score versus change (%) in the diazepam daily dose equivalent (DDE). The blue line represents the predicted change in NPI-NH total occupational disruptiveness score based on the change (%) in the diazepam daily dose equivalent. NPI-NH=Neuropsychiatric Inventory-Nursing Home version.
6.8.2.2.2.2 Domain-specific occupational disruptiveness
A trend toward decreased disruptiveness related to depression and irritability between baseline and four months was noted among residents initially prescribed benzodiazepines (Table 58,
Figure 67 and Figure 68). On the other hand, all remaining domain-specific occupational disruptiveness scores remained relatively stable.
278 | Page Daniel J. Hoyle Table 58: Mean Neuropsychiatric Inventory-Nursing Home version domain-specific occupational disruptiveness scores at baseline and four months.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Delusions 0.6 (0.3, 0.8) 0.4 (0.2, 0.6) 1.01 117 0.314 Hallucinations 0.3 (0.1, 0.5) 0.2 (0.1, 0.4) 0.81 117 0.419 Agitation 1.1 (0.9, 1.4) 1.0 (0.7, 1.2) 1.19 117 0.236 Depression 1.1 (0.8, 1.3) 0.9 (0.7, 1.1) 1.76 117 0.080 Anxiety 1.0 (0.8, 1.2) 0.9 (0.7, 1.2) 0.63 117 0.529 Elation 0.1 (0.0, 0.2) 0.1 (0.0, 0.2) 0.47 117 0.641 Apathy 0.5 (0.4, 0.7) 0.5 (0.3, 0.6) 0.58 117 0.564 Disinhibition 0.5 (0.3, 0.7) 0.5 (0.3, 0.7) -0.07 117 0.942 Irritability 1.2 (0.9, 1.4) 0.9 (0.7, 1.2) 1.77 117 0.079 Aberrant motor behaviour 0.4 (0.2, 0.6) 0.3 (0.1, 0.5) 1.11 117 0.269 Night-time behaviour 0.8 (0.5, 1.0) 0.7 (0.5, 0.9) 0.81 117 0.422 Appetite 0.3 (0.1, 0.4) 0.4 (0.2, 0.6) -1.29 117 0.202 df=degrees of freedom, 95%CI=95% confidence interval
279 | Page Daniel J. Hoyle
Figure 67: Beeswarm plot of the Neuropsychiatric Inventory-Nursing Home version domain-specific occupational disruptiveness scores at baseline and four months for residents taking a benzodiazepine at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval.
280 | Page Daniel J. Hoyle
Figure 68: Beeswarm plot of the Neuropsychiatric Inventory-Nursing Home version domain-specific occupational disruptiveness scores at baseline and four months for residents taking a benzodiazepine at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval.
Using linear regression to control for having a documented diagnosis of dementia, there was a trend towards decreased occupational disruptiveness related to night-time behaviour associated with a greater magnitude of benzodiazepine dose reduction (β=0.004; i.e., NPI-NH occupational disruption score related to night-time behaviour was 0.04 points lower, on average, for each 10% reduction in the diazepam daily dose equivalent, p=0.071; 95%CI=-0.000,
0.007; Table 59). There was no evidence that benzodiazepine dose reduction was associated
281 | Page Daniel J. Hoyle with worsening in any domain-specific occupational disruptiveness scores (Table 59, Figure 69 and Figure 70).
Table 59: Estimated changes in the Neuropsychiatric Inventory-Nursing Home version domain-specific occupational disruptiveness scores associated with changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between baseline and four months.
Outcome Estimate (β) 95%CI t- p-value statistic
Intercept -0.008 -0.370, 0.353 -0.046 0.964
Benzodiazepine dose change (%) 0.001 -0.003, 0.004 0.229 0.819
related OD Documented diagnosis of -0.271 -0.808, 0.266 -0.999 0.320 - dementia Adjusted R2 -0.008 Delusion Multiple R2 0.009
Intercept -0.007 -0.261, 0.247 -0.054 0.957
Benzodiazepine dose change (%) 0.001 -0.002, 0.003 0.393 0.695 related -
Documented diagnosis of -0.136 -0.514, 0.242 -0.712 0.478 OD dementia Adjusted R2 -0.011
Hallucination Multiple R2 0.006
Intercept -0.059 -0.481, 0.363 -0.278 0.782
Benzodiazepine dose change (%) 0.001 -0.003, 0.006 0.475 0.636
related OD Documented diagnosis of -0.246 -0.873, 0.382 -0.775 0.440 - dementia Adjusted R2 -0.010
Agitation Multiple R2 0.007
Intercept -0.110 -0.407, 0.188 -0.730 0.467
Benzodiazepine dose change (%) -0.001 -0.004, 0.002 -0.759 0.449 related OD - Documented diagnosis of -0.242 -0.684, 0.199 -1.087 0.279 dementia Adjusted R2 -0.003
Depression Multiple R2 0.015
282 | Page Daniel J. Hoyle Intercept 0.070 -0.292, 0.431 0.381 0.704
Benzodiazepine dose change (%) 0.001 -0.003, 0.004 0.298 0.766
Documented diagnosis of -0.327 -0.864, 0.211 -1.205 0.231 related OD - dementia Adjusted R2 -0.004 Anxiety Multiple R2 0.014
Intercept -0.003 -0.150, 0.144 -0.04 0.965
Benzodiazepine dose change (%) 0.001 -0.001, 0.002 0.980 0.329
Documented diagnosis of -0.019 -0.238, 0.199 -0.175 0.862 related OD - dementia Adjusted R2 -0.009
Elation Multiple R2 0.009
Intercept -0.005 -0.319, 0.309 -0.033 0.974
Benzodiazepine dose change (%) 0.003 -0.001, 0.006 1.555 0.123
Documented diagnosis of -0.037 -0.504, 0.430 -0.157 0.876 related OD - dementia Adjusted R2 0.004
Apathy Multiple R2 0.021
Intercept 0.046 -0.266, 0.358 0.292 0.770
Benzodiazepine dose change (%) 0.002 -0.001, 0.005 1.199 0.233 related OD
- Documented diagnosis of -0.005 -0.468, 0.459 -0.020 0.984 dementia Adjusted R2 -0.005
Multiple R2 0.012 Disinhibition
Intercept -0.025 -0.396, 0.346 -0.132 0.895
Benzodiazepine dose change (%) 0.001 -0.003, 0.005 0.675 0.501
Documented diagnosis of -0.448 -0.999, 0.104 -1.607 0.111 related OD - dementia Adjusted R2 0.010
2 Irritability Multiple R 0.027
283 | Page Daniel J. Hoyle Intercept -0.139 -0.407, 0.129 -1.027 0.307
Benzodiazepine dose change (%) 0.001 -0.002, 0.004 0.576 0.566
related OD Documented diagnosis of 0.098 -0.300, 0.496 0.489 0.626 - dementia Adjusted R2 -0.013 Aberrant motor 2 behaviour Multiple R 0.005
Intercept 0.250 -0.104, 0.605 1.399 0.165 - Benzodiazepine dose change (%) 0.004 -0.000, 0.007 1.824 0.071
Documented diagnosis of -0.679 -1.206, - -2.550 0.012 dementia 0.152
time behaviour 2 related OD - Adjusted R 0.066
2
Night Multiple R 0.081
Intercept 0.183 -0.101, 0.467 1.277 0.204
Benzodiazepine dose change (%) 0.002 -0.001, 0.005 1.338 0.183
Documented diagnosis of -0.025 -0.447, 0.396 -0.119 0.905 related OD - dementia Adjusted R2 -0.002
Appetite Multiple R2 0.016
OD=Occupational disruptiveness, 95%CI=95% confidence interval
284 | Page Daniel J. Hoyle
Figure 69: Scatterplots of the changes in the Neuropsychiatric Inventory-Nursing Home version (NPI-NH) domain- specific occupational disruptiveness scores versus change (%) in the diazepam daily dose equivalent (DDE). The blue line represents the predicted change in the NPI-NH occupational disruptiveness score based on the change (%) in the diazepam DDE.
285 | Page Daniel J. Hoyle
Figure 70: Scatterplots of the changes in the Neuropsychiatric Inventory-Nursing Home version (NPI-NH) domain- specific occupational disruptiveness scores versus change (%) in the diazepam daily dose equivalent (DDE). The blue line represents the predicted change in the NPI-NH occupational disruptiveness score based on the change (%) in the diazepam DDE. 6.8.3 Job satisfaction
6.8.3.1 Methods
6.8.3.1.1 Adjusted-Measure of Job Satisfaction
The adjusted-MJS is a validated and reliable measure of job satisfaction within the Australian
RACF setting.592 The adjusted-MJS survey was distributed to RACF nurses and PCAs by the champion nurse at each facility. As per Chapter 5.7.1.12.2, the adjusted-MJS contains five factors; personal satisfaction, workload, team spirit/co-workers, training, and professional
286 | Page Daniel J. Hoyle support. Factor scores presented in this chapter were calculated by averaging the item scores within each factor.592 This provided a factor score between 1 and 5, where higher scores indicated greater job satisfaction.
Additionally, a total adjusted-MJS score was calculated by summing the scores for all
22-items. This provided a score between 22 and 110, where higher scores indicated greater job satisfaction.
6.8.3.1.2 Sample size
Assessing the change in job satisfaction was considered a secondary objective. Therefore, the sample size was not calculated prior to data collection. However, calculation of the required sample size was performed retrospectively to ensure the analyses were sufficiently powered.
Overall, it was calculated that 63 staff at baseline and four months were required to complete the adjusted-MJS survey to detect a meaningful change of 6.1 in the adjusted-MJS total score (calculation based on the 0.5 of a standard deviation rule using baseline data), given a mean of 84.1 (standard deviation= 12.2), with a significance level (alpha) of 0.05 and a power
(beta) of 0.80.
6.8.3.1.3 Statistical analysis
Incomplete responses were removed from the analysis.
Differences in demographic data between baseline and four months were investigated.
Chi-squared tests were used to compare categorical information (i.e., occupation, age groups, gender, employment in a dementia care unit). Mann-Whitney U tests were used to compare continuous variables at baseline and four months (i.e., number of years registered/enrolled, number of years of aged care experience).
The distribution of the MJS factor scores were initially investigated using beeswarm plots. Considering that it was not possible to match responses between baseline and four
287 | Page Daniel J. Hoyle months due to anonymous completion of the survey, independent t-tests were used to detect differences in the adjusted-MJS factor scores between baseline and four months. The data were re-analysed based on the occupation of the respondent (RN/EN/PCA). This re-analysis is reported in Appendix L.
6.8.3.2 Results
Overall, 261 and 250 staff members completed the survey at baseline and four months, respectively. Considering the anonymous completion of the survey, it is unclear whether this survey was completed by the same nursing staff at both time points. There were no significant differences in demographic data between baseline and four months. However, there was a trend toward more PCA staff and fewer females completing the survey at four months (Table
60).
288 | Page Daniel J. Hoyle Table 60: Demographic data for nursing and care staff participating in the job satisfaction survey at baseline and four months.
Demographic variable Baseline (n=261) Four months (n=250) p-value % occupational groups 0.050 registered nurses 36.8% (n=96) 29.6% (n=74) enrolled nurses 23.4% (n=61) 17.6% (n=44) personal care assistants 31.8% (n=83) 42.4% (n=106) managers 1.9% (n=5) 0.8% (n=2) other 4.6% (n=12) 7.2% (n=18) unspecified 1.5% (n=4) 2.4% (n=6) % age groups 0.921 <20 years 2.3% (n=6) 1.2% (n=3) 20 to 29 years 19.9% (n=52) 18.4% (n=46) 30 to 39 years 19.9% (n=52) 20.8% (n=52) 40 to 49 years 21.8% (n=57) 23.6% (n=59) 50 to 59 years 25.3% (n=66) 23.6% (n=59) ≥60 years 8.8% (n=23) 10.8% (n=27) unspecified 1.9% (n=5) 1.6% (n=4) % gender 0.062 female 80.5% (n=210) 72.8% (n=182) male 9.6% (n=25) 10.4% (n=26) unspecified 10.0% (n=26) 16.8% (n=42) Years registered; median (range)* 6 (0, 48) 7 (0.5, 40) 0.839 Years enrolled; median (range)† 4 (0, 43) 5 (1, 36) 0.481 Years of aged care experience; median 5.5 (0, 45) 7 (0, 40) 0.087 (range)‡ % of sample that works in a dementia unit 52.2% (n=129) 47.8% (n=118) 0.615 *data available for 87 registered nurses at baseline and 67 registered nurses at four months †data available for 52 enrolled nurses at baseline and 40 enrolled nurses at four months ‡data available for 256 respondents at baseline and 246 respondents at four months
There were no substantial differences in any of the adjusted-MJS factor or total scores between baseline and four months (Table 61). Similarly, the distribution of adjusted-MJS factor and total scores in the beeswarm plots (Figure 71 and Figure 72) remained consistent.
Re-analysis of the data based on the occupation of the respondent (PCA, EN or RN; data provided in Appendix L) did not reveal any substantial differences in any of the job satisfaction factors or total score between baseline and four months, except for changes in the Team
Spirit/Co-workers factor for ENs. On average, there was a trend towards lower satisfaction with
Team Spirit/Co-workers for ENs between baseline (mean=4.2, 95%CI=4.0, 4.3) and four months
(mean=3.9, 95%CI=3.7, 4.2) (t(77)=1.78, p=0.078).
289 | Page Daniel J. Hoyle Table 61: Mean adjusted-Measure of Job Satisfaction total and factor scores at baseline and four months.
Outcome Mean score at Mean score at t-statistic df p-value baseline (95%CI) four months (95%CI) Personal Satisfaction 4.1 (4.0, 4.1) 4.1 (4.0, 4.2) -0.76 500 0.450 Workload 3.3 (3.2, 3.4) 3.4 (3.3, 3.5) -1.12 491 0.264 Team Spirit/Co-workers 4.1 (4.0, 4.2) 4.1 (4.0, 4.2) 0.19 499 0.851 Training 3.8 (3.7, 3.9) 3.7 (3.6, 3.8) 0.70 508 0.484 Professional support 4.0 (3.9, 4.1) 3.9 (3.8, 4.0) 0.95 503 0.344 Total 84.1 (82.6, 85.6) 84.1 (82.5, 85.8) -0.05 500 0.959 df=degrees of freedom, 95%CI=95% confidence interval
Figure 71: Beeswarm plots of the adjusted-Measure of Job Satisfaction factor scores at baseline and four months. The blue dot represents mean and error bars indicate the 95% confidence interval.
290 | Page Daniel J. Hoyle
Figure 72: Beeswarm plot of the adjusted-Measure of Job Satisfaction (MJS) total scores at baseline and four months. The blue dot represents mean and error bars indicate the 95% confidence interval. 6.8.4 Discussion
This study aimed to explore the relationship between antipsychotic and/or benzodiazepine dose reduction and occupational disruptiveness, and the difference in job satisfaction between baseline and four months after implementing RedUSe. Overall, there was no evidence that antipsychotic and/or benzodiazepine dose reduction was associated with an increase in occupational disruptiveness. Moreover, there were trends toward improvements in disruptiveness to staff, particularly related to agitation with antipsychotic dose reduction and night-time behaviour with benzodiazepine dose reduction. Additionally, there was no indication of differences in job satisfaction between baseline and four months after the implementation of RedUSe.
The absence of evidence to suggest increased occupational disruptiveness with antipsychotic dose reduction aligns with two earlier Canadian antipsychotic reduction educational interventions, which found no increase in stressful events associated with the
291 | Page Daniel J. Hoyle intervention.498, 499 As discussed in the introduction, RACF staff often have limited training in the management of neuropsychiatric symptoms and are required to work in resource-poor environments (e.g., inadequate staffing).65, 170 These issues promote the inappropriate use of antipsychotic medications in RACFs.43, 170, 187, 491 Considering both RedUSe50 and the aforementioned Canadian interventions498, 499 utilised educational components, it is likely that improved training on the non-pharmacological management of neuropsychiatric symptoms may have facilitated the reduction in antipsychotic and/or benzodiazepine dose without worsening RACF staff distress.
Additionally, these results are consistent with our earlier findings that neuropsychiatric symptoms generally did not worsen with antipsychotic and/or benzodiazepine dose reduction
(Chapter 6.3). Ironically, despite antipsychotics commonly being utilised for agitation and aggression and benzodiazepines used for sleep disturbances, we found trends toward improvements in occupational disruptiveness related to agitation in residents who had their antipsychotic dose reduced and occupational disruptiveness related to night-time behaviours in residents who had their benzodiazepine dose reduced.
Our finding, suggesting an improvement in occupational disruptiveness related to agitation, aligns with our findings, in Chapter 6.3, where we reported that the frequency of agitation was reduced in those who had their antipsychotic dose reduced. In part, these results may represent improvements in antipsychotic-related akathisia, which can be misdiagnosed as agitation.693 Conversely, we found no evidence of a relationship between benzodiazepine dose reduction and an improvement in the NPI-NH factor related to sleep. It should be noted, though, that the NPI-NH domain related to sleep is relatively insensitive to small changes in the severity and frequency of these symptoms.545 Therefore, there may have been some improvements in sleep that were not captured with the NPI-NH. Alternatively, the training in person-centred care strategies taught within RedUSe may have improved the confidence and ability of RACF staff
292 | Page Daniel J. Hoyle members to manage sleep disturbances with non-pharmacological methods translating to improvements in the occupational disruptiveness score.
The trends toward improvements in occupational disruptiveness related to these symptoms are interesting considering that agitation, aggression and severe sleep disturbances are associated with high levels of disruption and distress of RACF staff.144, 593 The cause of these improvements is unclear. However, improvements (and lack of worsening) in occupational disruptiveness could be attributed to improvements (or lack of deterioration) in symptoms, improved capabilities of staff to implement non-pharmacological strategies as a result of training with RedUSe, or a combination of both. In any case, the fact that there was no evidence of worsening in occupational disruptiveness should mitigate barriers to antipsychotic and/or benzodiazepine dose reduction.
To our knowledge, earlier interventions to reduce antipsychotic and/or benzodiazepine use in RACFs have not reported on changes in job satisfaction. However, unlike our study, where we found no appreciable difference in job satisfaction between baseline and four months after the implementation of RedUSe, Zwijsen et al. reported an improvement in job satisfaction of care staff related to the implementation of a care programme for challenging behaviour in people with dementia within 17 Dutch dementia special care units.692 The Grip on Challenging
Behaviour care programme (GRIP) provided training to care staff on the identification of challenging behaviour and ongoing resident monitoring informed management strategies put in place by a multidisciplinary team.694 Job satisfaction was measured before the implementation of GRIP, halfway through implementation and after all care units had implemented the intervention.692 Although not conclusive, Zwijsen et al. suggest that the improvement in job satisfaction may be associated with a reduction in job demand resulting from the support offered by the intervention.692 Alternatively, increased job satisfaction may be related to improvements in staff knowledge as a result of the training sessions.692
293 | Page Daniel J. Hoyle Zwijsen et al. were able to review the change in job satisfaction among the same RACF nursing and care staff over the implementation of GRIP.692 However, in our study, anonymous completion of the job satisfaction survey meant that we did not know whether the same staff were completing the survey at both baseline and four months. Consequently, the lack of difference in job satisfaction between baseline and four months may be attributed to the completion of the survey by different staff between baseline and four months (e.g., new staff at four months) who were not impacted by the intervention. Alternatively, it needs to be acknowledged that job satisfaction was not expected to change between baseline and four months given that the RedUSe intervention required little input by most RACF staff (i.e., staff generally did not have to audit or review medication use). The absence of evidence to suggest worsening in job satisfaction with the implementation of RedUSe is a positive result as benefits may be derived elsewhere (e.g., reduced medication costs, reduction in adverse effects) without a worsening in job satisfaction.
6.8.4.1 Strengths and limitations
This research is one of few to investigate the outcomes for RACF staff within an intervention to ensure appropriate antipsychotic and/or benzodiazepine use and is strengthened by using validated scales to assess changes in occupational disruptiveness and differences in job satisfaction between baseline and four months. Whilst this research is an important contribution to the literature, there are several limitations which need to be accounted for.
First, the job satisfaction survey was completed anonymously as per ethical stipulations.
Consequently, the respondents could not be matched between baseline and four months.
Whilst there was no evidence to suggest that job satisfaction changed after implementation of
RedUSe, some of the surveys were likely completed by different RACF staff members at each time point. In some instances, new staff may have taken the survey without being exposed to the RedUSe intervention.
294 | Page Daniel J. Hoyle Anonymous completion of the survey also precluded our ability to identify whether surveys were completed multiple times by the same participants at each time point. However, the demographic details provided by the participants matched only three times (i.e., six participants recorded the same demographic details) at a single data collection time point indicating that it is unlikely the same staff completed the survey multiple times at each time point. Furthermore, re-analysis of the data excluding the results for these six staff members did not lead to any substantial changes in the results. Future studies that investigate longitudinal changes in job satisfaction should consider methods to track the results for individual respondents while maintaining participant privacy.
Similarly, identifying information for the RACF staff member who was interviewed with the NPI-NH occupational disruptiveness scale was not recorded. Thus, it is possible that differences between baseline and four months may be due to inter-rater differences. However, on average, the level of disruptiveness for most NPI-NH domains did not change substantially, suggesting that inter-rater differences were not substantially affecting our results.
It was not possible to quantitatively investigate whether RACF staff (i.e., RNs, ENs, PCAs) completing the job satisfaction survey were representative of the national RACF workforce.
However, qualitative comparison between demographic data from our survey and findings from the 2016 National Aged Care Workforce Census65 suggests that RNs were over-represented and
PCAs were under-represented in our job satisfaction survey. This likely relates to the timing of survey distribution. Most champion nurses indicated that the baseline job satisfaction surveys were distributed during the initial RedUSe training session, which were mostly attended by RNs and ENs. Considering no substantial differences in job satisfaction were detected in sub- analyses of the individual occupations (Appendix L), it is likely that the under-representation of
PCAs has not impacted the overall results of the survey. In any case, researchers should employ
295 | Page Daniel J. Hoyle strategies to ensure a representative sample is recruited in future studies of the aged care workforce.
Feelings of job satisfaction and occupational disruptiveness are influenced by a variety of factors external to RedUSe. Whilst it is unclear what other external factors may have impacted job satisfaction and occupational disruptiveness, our results did not change significantly suggesting a minimal impact of these external factors.
Recruitment bias is also possible considering that the champion nurse for the RACF was responsible for the distribution of the job satisfaction surveys. However, the absence of evidence to suggest differences in job satisfaction at baseline and four months suggests that recruitment bias is unlikely to be influencing the results in this instance.
Lastly, some of the limitations discussed in Chapter 6.3 also affected our analysis in
Chapter 6.8, such as:
• absence of a control group; and,
• lack of rater blinding.
6.8.5 Conclusion
Antipsychotic and/or benzodiazepine dose reduction was not associated with worsening in occupational disruptiveness. Ironically, there were trends towards improved occupational disruptiveness related to agitation with antipsychotic dose reduction and night-time behaviour with benzodiazepine dose reduction. Larger, controlled studies are needed to investigate these trends further.
Additionally, there was no evidence that the implementation of RedUSe worsened the job satisfaction of RACF staff. Controlled longitudinal studies that track individual RACF staff are needed to explore whether changes in job satisfaction are related to the extent of change in antipsychotic and/or benzodiazepine use within RACFs.
296 | Page Daniel J. Hoyle 6.9 Basic costing analysis
6.9.1 Introduction
Antipsychotic and benzodiazepine dose reduction has the potential to be cost saving through reduced prescribing and adverse effects of these medications. However, these savings may be offset by prescription of other medications or increased utilisation of the health care system
(e.g., increased GP consultations). Considering scarce resources, policymakers need to ensure that interventions to improve the prescribing of antipsychotic and/or benzodiazepine medications represent good value for money.
Despite calls for increased reporting of the cost-effectiveness of psychotropic reduction interventions,48, 507, 695, 696 our systematic review did not identify any interventions similar to
RedUSe that have reported on the cost benefits related to antipsychotic and/or benzodiazepine dose reduction.47 However, several studies, which did not meet the inclusion criteria for our systematic review or were published after our review, have indicated that interventions to reduce these medications can be cost-effective.527, 532, 697 Whilst these studies compared savings in RACFs receiving the intervention versus controls, none directly investigated the effects of dose reduction on the individual residents. Furthermore, the previous studies were based in the
UK,532 Ireland,697 and the Netherlands527; thus, they are difficult to generalise to the Australian setting. In the initial implementation of RedUSe, Associate Professor Breen reported an interim costing analysis.497 Associate Professor Breen’s analysis indicated potential cost savings due to reduced antipsychotic and benzodiazepine prescriptions. However, it was outside the scope of this study to investigate other factors (e.g., costs of other medications, utilisation of the health care system), which may offset some or all savings derived from the intervention.
Given this, a basic costing analysis was conducted to determine the potential cost of antipsychotic and/or benzodiazepine dose reduction. Figure 73 illustrates where these findings fit in this overall thesis. Limitations of this study, most notably the absence of a control group
297 | Page Daniel J. Hoyle and longitudinal data prevented a more comprehensive economic (cost-effectiveness) analysis from being carried out. However, it was possible to provide an estimate of costs and benefits, by drawing together data from a range of sources and making some key assumptions.
Figure 73: Flowchart of the analysis for the clinical outcomes study. This sub-chapter is dedicated to a basic costing analysis of antipsychotic and benzodiazepine dose reduction. GP=General Practitioner. 6.9.2 Methods
6.9.2.1 Grouping
Residents who were not followed-up at four months were excluded from the analysis. The choice to exclude these residents was made because of the absence of four-month data, such as the cost of medications, and, in many cases, the absence of GP consultation data and hospitalisation records.
To perform this basic costing analysis, residents were split into groups based on changes in chlorpromazine and diazepam daily dose equivalents. Residents who were initially taking an antipsychotic (i.e., chlorpromazine daily dose equivalent >0 mg at baseline) and who had a reduction in their chlorpromazine daily dose equivalent at the four-month data collection point were classified as ‘antipsychotic reducers’. Residents who were initially taking a benzodiazepine
298 | Page Daniel J. Hoyle (i.e., diazepam daily dose equivalent >0 mg at baseline) and who had a reduction in their diazepam daily dose equivalent at the four-month data collection point were classified as
‘benzodiazepine reducers’.
Residents who were classified as an ‘antipsychotic reducer’ and/or ‘benzodiazepine reducer’ were grouped in an overall ‘reducer’ group. Residents who were initially taking antipsychotics or benzodiazepine and were neither ‘antipsychotic reducers’ nor
‘benzodiazepine reducers’ were grouped in an overall ‘non-reducer’ group. Three of the 179 residents had an increase in dose for one medication but a reduction in the other. For the purposes of this study, they were classified as an overall ‘non-reducer’ considering that one class of medication had replaced the other.
6.9.2.2 Calculating the average cost benefit per resident
The average cost benefit was calculated on a per resident basis for all residents, non-reducers, reducers, antipsychotic reducers, and benzodiazepine reducers. Daily costs related to the PBS were calculated at baseline and four months. These amounts were extrapolated to four monthly then 12-monthly costs. GP consultation and hospitalisation costs were calculated over the four- month study period and were extrapolated to costs over 12 months. Costs are reported as
Australian Dollars.
To account for the variability in the data, two sets of savings data were calculated; termed “optimistic” and “conservative” savings. Optimistic savings involved calculating the potential savings associated with being a reducer (i.e., antipsychotic or benzodiazepine reducer) versus being a non-reducer of either medicine. This was considered an optimistic saving because it is likely that a proportion of residents in the non-reducers group would not have had their medication reduced due to valid clinical reasons (e.g., because they were considered too
‘sick’), rather than being simply due to apathy on behalf of their prescriber. Subsequently, it may be the case that the non-reducer group had, on average, worse health than the reducer
299 | Page Daniel J. Hoyle group, and would consequently represent a greater drain on the health care system (e.g., through increased medications, GP consultations). Conservative savings involved calculating the potential savings associated with being a reducer compared to the overall sample (i.e., including both reducers and non-reducers). This was done to minimise the influence of potentially greater health care utilisation costs associated with the potentially ‘sicker’ residents in the non-reducer group.
6.9.2.3 Medication costs based on the pharmaceutical benefits scheme
Data on all resident medications, including strength and number of doses (e.g., tablets) per day, were collected at baseline and four months. To calculate the total cost of medications per day, we first calculated the cost per dose using the Dispensed Price for Maximum Quantity (DPMQ) for medications on the PBS based on prices as of March 2016. The DPMQ is the price for dispensing the maximum quantity of a product under a given prescribing rule and includes the
“price ex-manufacturer, including all fees, mark-ups and patient contributions”.698 The DPMQ was then divided by the number of doses within a pack and multiplied by the number of doses per day taken by the resident to give the cost of medications per day.
6.9.2.4 General practitioner consultations
Information on GP visits to the residents involved in this study were collected by the champion nurses and through audit of the GP consultation records by the thesis author between baseline and four months (see Chapter 5.7.1.10). Unfortunately, it was not possible to identify the
Medicare Benefits Schedule (MBS) item number5 used when claiming for the consultations with the residents.
5 Medicare is Australia’s national public health insurance scheme. The scheme provides free or subsidised healthcare to all Australians and most permanent residents. Doctors, specialists, optometrists and other healthcare professionals (e.g., dentists) claim payment through the Medicare Benefits Schedule. Each claimable medical service is assigned a fee and item number (https://www.betterhealth.vic.gov.au/health/ServicesAndSupport/understanding-medicare).
300 | Page Daniel J. Hoyle Instead, the overall cost of GP consultations was estimated by multiplying a calculated average cost per consultation by the number of consultations each resident had over the four months.
To estimate the average cost per GP consultation, we used Medicare claims data from states involved in this project (New South Wales, Queensland, Victoria, Tasmania and South
Australia) from May 2016. First, all MBS items that could be claimed following a consultation with a RACF resident were identified.699 This included MBS items 20, 35, 43, 51, 92, 93, 95, 96,
5010, 5028, 5049, 5067, 5260, 5263, 5265, and 5267 but not 701, 703, 705 and 707, which can be claimed for a health assessment of people residing in and outside of a RACF.699 Table 62 describes these MBS items. The amount of money claimed through these MBS item numbers was then divided by the number of times these MBS items were utilised, providing an average cost per consultation.
Table 62: Descriptions of Medicare Benefits Scheme item numbers. Information from the Department of Health, 2015.699
Medicare Benefits Scheme item number Description 20*, 35**, 43***, 51**** Fees and benefits for GP attendances at a residential aged care facility 92*, 93**, 95***, 96**** Fees and benefits for other non-referred attendances at a residential aged care facility 5010*, 5028**, 5049***, 5067**** Fees and benefits for GP after-hours attendances at a residential aged care facility 5260*, 5263**, 5265***, 5267**** Fees and benefits for other non-referred after- hours attendances at a residential aged care facility 701, 703, 705, 707 Health assessment *Brief consultation, **standard consultation, ***long consultation, ****prolonged consultation
6.9.2.5 Hospital
Information on the reason and length of each hospitalisation for residents involved in this study was collected by the champion nurses and through audit of the resident’s progress notes by the thesis author between baseline and four months (see Chapter 5.7.1.11). Hospital admission costs per resident were estimated by multiplying the days of each hospital admission by the
301 | Page Daniel J. Hoyle average cost/day according to the reason for admission using data obtained from Victorian hospitals in 2015.6
6.9.2.6 Quality of life
The AQoL-4D tool was used to assess QoL at baseline and four months through psychometric testing of RACF nursing and care staff. As per Chapter 5.7.1.5.1, the weighted AQoL-4D utility scores, commonly used in economic analyses, were calculated and presented in this chapter.
These scores range from -0.04 (QoL worse than death) to 0 (QoL equivalent to death) to 1 (best
QoL).
6.9.3 Results
An average cost per resident in each group was calculated (Table 63). When total costs were extrapolated to 12 months, we can see modest savings in the total costs for antipsychotic and benzodiazepine reducers in both the conservative and optimistic analysis. This was mainly driven by savings in hospitalisations, particularly amongst the benzodiazepine reducer group.
The antipsychotic reducers demonstrated savings in overall PBS costs by $349.18 per year in the optimistic analysis and $254.28 per year in the conservative analysis. Whereas benzodiazepine reducers demonstrated savings of $163.22 per year in the optimistic analysis and $68.31 per year in the conservative analysis. Overall, reducers saved $257.40 per year in the optimistic analysis and $162.49 per year in the conservative analysis.
Using an estimated cost of $90.44 per GP consultation, we found that consultation costs were greater for reducers of either medication in both the optimistic ($192.27 more per year) and conservative ($121.38 more per year) analyses.
On average, reducers of antipsychotic and/or benzodiazepine medications cost the
Government $1,421.18 less in hospital fees than non-reducers and $897.17 less than the overall
6 http://health.vic.gov.au/feesman/fees1.htm
302 | Page Daniel J. Hoyle sample, over 12 months. However, there was limited hospitalisation data available in the dataset. Consequently, there is uncertainty around the estimation of these costs.
Lastly, there were clinically significant improvements in QoL for the ‘antipsychotic reducers’ group while, on average, QoL in the overall sample remained the same between baseline and four months. On average, there was also an improvement in QoL for
‘benzodiazepine reducers' and ‘reducers’ groups, although the change in QoL did not reach clinical significance.7
7 A change in the AQoL-4D score by ≥0.06 points is considered clinically significant.
303 | Page Daniel J. Hoyle Table 63: Average potential cost savings on a per resident basis.
Average resident All (n=179) Non-reducers (n=113) Reducers (n=66) Antipsychotic Benzodiazepine reducers (n=30) reducers (n=42) Before RedUSe PBS medicines (daily) $5.30 $5.34 $5.23 $5.12 $5.58 PBS medicines (120 days) $636 $640 $628 $614 $669 QoL 0.19 0.20 0.17 0.13 0.18 After RedUSe PBS medicines (daily) $5.31 $5.57 $4.87 $4.62 $5.13 PBS medicines (120 days) $638 $669 $584 $554 $615 QoL 0.19 0.17 0.21 0.22 0.19 During RedUSe (four-month period) GP consultations (n*) $545 (6.0) $521 (5.8) $585 (6.5) $642 (7.1) $571 (6.3) Hospitalisations (n**) $687 (0.92) $862 (1.15) $388 (0.52) $478 (0.63) $431 (0.57) Total costs extrapolated to 12 months PBS medicines $1,939 $2,034 $1,777 $1,685 $1,871 GP consultations $1,634 $1,563 $1,755 $1,926 $1,712 Hospitalisations $2,062 $2,586 $1,165 $1,433 $1,293 TOTAL $5,636 $6,184 $4,698 $5,044 $4,876 Optimistic total savings (versus non-reducers) $1,486 $1,140 $1,308 Conservative total savings (versus all residents) $938 $592 $760 Optimistic savings PBS meds only (versus non-reducers) $257.40 $349.18 $163.22 Conservative savings PBS meds only (versus all residents) $162.49 $254.28 $68.31 Optimistic savings GP consult only (versus non-reducers) -$192.27 -$363.28 -$148.81 Conservative savings GP consult only (versus all residents) -$121.38 -$292.39 -$77.92 Optimistic savings hospitalisation only (versus non-reducers) $1,421.18 $1,153.85 $1,293.88 Conservative savings hospitalisation only (versus all residents) $897.17 $629.84 $769.87 QoL improvements when reduced 0.04 0.09 0.01 n*=average number of GP consultations over four months, n**=average number of days admitted to hospital, RedUSe=Reducing Use of Sedatives project, PBS=Pharmaceutical Benefit Scheme, GP=General Practitioner
304 | Page Daniel J. Hoyle 6.9.4 Discussion
Our findings provide evidence that antipsychotic and benzodiazepine dose reduction is associated with less health care-related costs compared to residents who continue their psycholeptic medication. Whilst there are several limitations, the health care savings and improvements in QoL derived from antipsychotic and/or benzodiazepine dose reduction are relatively strong. Savings were mostly driven by lower hospital costs among residents who had their antipsychotic and/or benzodiazepine dose reduced. Dose reduction was also associated with savings in overall PBS medication costs and improvements in the QoL utility scores. On the other hand, MBS costs related to GP consultations were higher for residents who had their antipsychotic and/or benzodiazepine dose reduced, although this might be attributed to increased consultations associated with the review of dose and may have settled over a longer time frame.
In a study of 277 residents with dementia from 16 UK RACFs, Romeo et al. reported that hospital admissions account for 50% of costs, after excluding RACF fees.680 Furthermore, the authors reported that the use of psychotropic medication was associated with increased hospital costs.680 Although only based on a limited number of hospitalisations, we estimated savings of over $1,400 per resident per year with antipsychotic and/or benzodiazepine dose reduction compared to non-reducers. Antipsychotic and benzodiazepine medication are associated with several adverse effects, such as falls and fractures and pneumonia, which may lead to hospital admissions.195, 379, 700 Whilst the hospital savings with antipsychotic and/or benzodiazepine dose reduction may have been attributed to avoidance of these adverse effects, it should be acknowledged that it is also possible that some of the residents who did not have their antipsychotic and/or benzodiazepine dose reduced may be those with more severe health issues who tend to require longer and/or more frequent hospital admissions.
305 | Page Daniel J. Hoyle Residents who had their antipsychotic and/or benzodiazepine dose reduced had overall
PBS savings of approximately $257 per year compared to non-reducers. However, this was offset by an estimated increase in GP consultation costs (~$192 per resident per year greater for antipsychotic and/or benzodiazepine reducers compared to non-reducers). We initially expected a decrease in GP consultation costs considering that adverse effects would potentially be reduced with dose reduction. However, the increased GP costs may be related to a period of increased monitoring following dose reduction. It should be acknowledged that GP costs were estimated based on an average consultation cost that we calculated ($90.44). However, this average calculated consultation cost does not consider the type of GP consultation (e.g., a brief review may cost as little as $9.75 whereas an after-hours consultation may cost $164.45). It is possible that the consultation cost for a review following antipsychotic and/or benzodiazepine dose reduction may have been lower than the average consultation cost. Consequently, the overall GP consultation costs for residents who had their medication reduced may have been over-estimated. Future Australian studies should endeavour to obtain MBS claiming data to overcome this limitation when performing similar cost analyses.
6.9.4.1 Limitations
The presented information should be interpreted cautiously as there are several limitations.
First, without a true control group, it is difficult to rule out costs being related to clinical differences between the reducer and non-reducer groups. This was addressed, in part, by the conservative analysis which compared reducers with the overall sample.
Second, the largest cost savings were related to reduced hospitalisations among antipsychotic and benzodiazepine reducers. However, only 23 residents were hospitalised during the data collection period. Thus, we are uncertain whether the observed difference in hospitalisation costs between the reducer and non-reducer groups would hold true in a larger sample. Considering that most of the cost savings predicted in this model were associated with
306 | Page Daniel J. Hoyle reduced hospitalisations in the reducer groups, the overall cost savings predicted by this model would be significantly affected should the differences in hospitalisations be a statistical anomaly.
Third, the GP consultation data were collected by asking the champion nurse to keep a record of all GP visits. This was verified and amended based on resident notes at the RACF.
Whilst validation of these records helped to improve the accuracy of the data collected, if the physician had not recorded their visit in the notes and the nurse had not recorded this visit then that visit would not have been included in the data. Additionally, it was impossible to tell which
MBS item code the GP had used during their visit. Therefore, to estimate this cost we calculated the number of services for each relevant MBS item and the amount of money claimed for each item in order to calculate the average cost per resident per service. This may be grossly underestimating or overestimating the GP-related costs. Consequently, this estimation should be interpreted cautiously.
Fourth, it was beyond the scope of this study to investigate the impact that antipsychotic and/or benzodiazepine dose reduction had on the time spent caring for each resident by the RACF staff. This information is significant considering that some argue against antipsychotic and/or benzodiazepine dose reduction due to insufficient time and staff available to handle an increase in neuropsychiatric symptoms should they worsen. However, this information is difficult to collect and would likely require direct observation of the RACF staff for prolonged periods of time. This being said, our separate sub-study in Chapter 6.8 did not indicate any increase in occupational disruptiveness of staff following reduction, so it was considered unlikely that this factor would be substantial.
Fifth, this analysis was performed using Australia-specific data (e.g., MBS costs, PBS costs, hospitalisation costs). Consequently, this basic costing analysis is not generalisable internationally.
307 | Page Daniel J. Hoyle Lastly, this analysis did not include the costs of RedUSe, which also included substantial costs related to researchers and IT program development. However, we anticipate relatively few future costs considering that this program can be delivered largely within the existing framework of RMMR and QUM activities within RACFs. A sensitivity analysis including the costs of RedUSe has been reported elsewhere.217
6.9.5 Conclusion
This basic cost analysis indicates that antipsychotic and/or benzodiazepine dose reduction has the potential to be cost saving. However, larger, controlled studies with a dedicated health economic analysis embedded as part of the intervention are required to draw definitive conclusions. Savings associated with antipsychotic and/or benzodiazepine dose reduction were mainly related to decreased hospitalisations and reduced expenditure on PBS medications.
308 | Page Daniel J. Hoyle Chapter 7 Summary and conclusion
Internationally, the high prevalence of psychotropic drug use, predominantly antipsychotics and benzodiazepines, in RACFs has led to concern of residents being inappropriately sedated.10, 16,
138 Antipsychotics are commonly used to manage neuropsychiatric symptoms, such as aggression, agitation or psychosis,25 and benzodiazepines are widely used to mitigate acute agitation, anxiety and sleep disturbances.27 Antipsychotics and benzodiazepines are only modestly effective for these indications 22, 35 and pose a risk of severe adverse effects, such as an increased risk of falls and fractures,37, 281 cognitive impairment and possibly dementia with benzodiazepine use,308 and stroke and death with antipsychotic use.22, 248
Withdrawal or dose reduction of these medications may minimise the risk of adverse effects. However, these medications are often utilised long-term without dosage modification.12, 25 Concern of deterioration in QoL and worsened neuropsychiatric symptoms, consequently leading to increased time required to manage these symptoms, are some of the barriers to antipsychotic and benzodiazepine reduction in RACFs (Chapter 3).44, 45, 187, 491 Despite this, few interventions to improve antipsychotic and/or benzodiazepine prescribing in RACFs have reported on the clinical outcomes for the residents, changes in disruptiveness or job satisfaction of the RACF staff, or changes in health care-related costs (Chapter 4). To address this gap and minimise barriers to antipsychotic and/or benzodiazepine dose reduction, this thesis explored clinical and economic outcomes associated with dose changes to antipsychotics and/or benzodiazepines in RACFs. In particular, resident and staff outcomes, and health care- related cost outcomes were investigated.
Our key finding was that there was no evidence of worsened resident or staff outcomes with antipsychotic and/or benzodiazepine dose reduction. Furthermore, a basic costing analysis suggested potential health care-related savings associated with a reduction in these psychotropic medications. These results are discussed in detail throughout Chapter 6.
309 | Page Daniel J. Hoyle Overall, there was no evidence of deterioration in neuropsychiatric symptoms, QoL, social withdrawal, ADLs, or staff disruptiveness associated with antipsychotic and/or benzodiazepine dose reduction. The lack of change in the outcome measures with antipsychotic and/or benzodiazepine dose reduction is an overall positive result. These results, alongside those from earlier and other interventions, provide evidence that considered reduction is unlikely to cause harm to the resident or RACF staff. Furthermore, by minimising the exposure to antipsychotics and/or benzodiazepines, we can reduce the resident’s risk of experiencing adverse effects related to these medications. Considering that concern of potential deterioration in resident QoL and neuropsychiatric symptoms resulting in increased workloads are major barriers to the reduction of these medications,43-45 it is hoped that our results will lessen these concerns.
Substantial improvements in clinical outcomes were not expected with antipsychotic and/or benzodiazepine dose reduction considering that these medications confer only modest benefits for indications commonly encountered in the RACF setting (e.g., neuropsychiatric symptoms in dementia, anxiety, sleep disturbances and acute agitation). Despite this, in our study, antipsychotic dose reduction was associated with potential improvements in:
• physically nonaggressive behaviour (e.g., aimless wandering, general
restlessness)
• QoL
• staff disruptiveness related to agitation.
Additionally, benzodiazepine dose reduction showed potential to improve:
• physically nonaggressive behaviour
• verbally agitated behaviour (e.g., constant unwarranted requests for attention)
• staff disruptiveness related to night-time behaviours
310 | Page Daniel J. Hoyle Beyond this, benzodiazepine dose reduction was also associated with an improved rate of falls among mobile residents.
We also identified possible health care-related savings with antipsychotic and benzodiazepine dose reduction. These savings arose primarily from reduced hospitalisation and medication costs. Overall, annual savings of more than $1,400 per resident were estimated for residents who had a medication reduced.
7.1 Future research directions
This thesis was primarily focussed on investigating the clinical outcomes associated with antipsychotic and benzodiazepine dose change in a sample of residents within RedUSe.
Limitations directly related to the results reported in this thesis have been discussed in detail throughout Chapter 6. However, there were several broader topics outside of the scope of this thesis that should be considered in future research.
First, this thesis did not qualitatively investigate organisational or personal factors as to why antipsychotic and/or benzodiazepine medications were utilised or barriers to their reduction. However, our findings (Chapter 6.2) provide quantitative evidence that unknown differences in the RACFs and residents (e.g., differences in organisational culture, prescriber attitudes) heavily influences variation in antipsychotic and benzodiazepine doses. These results support findings from earlier research on the influence of organisational culture on psychotropic prescribing.43, 187, 489 Considering the potential impact of organisational culture, future qualitative research is needed to investigate factors that relate to the amplitude of change in psychotropic use within programs to improve their prescribing in RACFs. This could help to identify RACFs that are likely to respond to similar programs to optimise targeting of
RACFs in future implementations.
Second, this thesis focussed solely on the clinical outcomes of changes in antipsychotic and benzodiazepine prescribing considering these medications were targeted in the
311 | Page Daniel J. Hoyle overarching RedUSe intervention.50 However, there is concern that interventions targeting particular psychotropic medications may drive prescribing to other, potentially more dangerous and less efficacious, psychotropics.701 For example, interventions to reduce antipsychotic use may result in the prescribing of antidepressants. In 2011, a double-blind, placebo-controlled trial found no evidence that sertraline (n=107) or mirtazapine (n=108) are any more effective than placebo (n=111) in the management of depression in people with dementia.702
Furthermore, adverse reactions among participants given sertraline (43%) and mirtazapine
(41%) were almost double that compared to participants given placebo (26%).702 In our study, we found no evidence that antipsychotic or benzodiazepine prescribing was substituted with other medications (Chapter 6.3). However, it is important that future targeted interventions consider monitoring for increased prescribing of potentially less appropriate alternatives.
Additionally, future studies are needed to compare the outcomes of interventions targeting a narrow range of medications to interventions aiming to improve the prescribing of a broad range of medications.
Third, this study did not explicitly set out to assess the feasibility of measuring the clinical outcomes in RedUSe. However, the response rate was used as an indicator of method feasibility. Most tools in our study utilised nursing staff as proxy-raters. Consequently, response rates were close to, if not, 100% (Chapter 6), suggesting that interviewing nursing staff is a feasible method of data collection. We did, however, set out to measure QoL in the resident themselves, if able to provide informed consent, and the person responsible for them in order to perform a comparative analysis. Unfortunately, less than one-third of residents had sufficient cognitive capacity to be interviewed about their QoL. Furthermore, the response rate from the person responsible for the resident was relatively low despite reminder letters being sent.
Anecdotally, other challenges encountered included difficulty recruiting residents due to the need for proxy consent in the majority of cases and the ethical requirement for arms-length recruitment, limited nursing staff time available for assessments, inconsistent reporting of
312 | Page Daniel J. Hoyle events in RACF and GP consultation notes between and within RACFs, and the need to collect a broad range of data within a condensed timeframe due to travel requirements. Considering the challenges faced in this study, the following recommendations are suggested for future research in RACFs:
• where appropriate, reliance on resident-rated outcomes should be minimised,
particularly among populations with significant cognitive deficits;
• where ethically appropriate, studies should be designed using an opt-out recruitment
strategy;599
• use of RACF and GP records should be minimised given inconsistent reporting;
• the selection of assessment tools should be streamlined to minimise the impact on staff
time. Core outcome sets may provide guidance on pertinent outcomes;506 and,
• local, trained research assistants should be used for data collection in geographically
disperse projects;
Lastly, the results of this thesis report on the clinical outcomes of antipsychotic and/or benzodiazepine dose reduction within the implementation of the RedUSe intervention in the
Australian RACF setting. Whilst the generalisability of the overarching RedUSe intervention50 may be dependent on similarities in the aged care systems of other countries, the clinical outcomes of antipsychotic and/or benzodiazepine dose reduction reported in this thesis are mostly internationally applicable, especially in jurisdictions with high and inappropriate psycholeptic use. It is important, however, to acknowledge that the combination of the different facets of RedUSe (e.g., staff/pharmacist/prescriber education, sedative auditing, sedative reviews)50 likely played a significant role in the clinical outcomes. In particular, the delivery of non-pharmacological measures taught in the RedUSe staff education sessions may have mitigated some of the potentially negative effects associated with reducing antipsychotic and/or benzodiazepine use. In 2016, Ballard et al. showed that the addition of a social
313 | Page Daniel J. Hoyle interaction intervention can mitigate the negative effects an antipsychotic review intervention has on behavioural symptoms and QoL.514, 515
There are perceptions that insufficient resources, such as staff, may be a barrier to the implementation of non-pharmacological strategies in the RACF setting.44, 45, 162, 170, 703 To this end, clinical outcomes of antipsychotic and/or benzodiazepine dose reduction may differ based on the resourcing available to deliver non-pharmacological strategies. Countries with low proportions of RACF staff relative to the population aged 65 years and over, such as Turkey and
Portugal,6 may find it difficult to implement non-pharmacological strategies.704 Consequently, the clinical outcomes may be less certain in these countries. However, antipsychotic and/or benzodiazepine dose reduction in RACFs of countries with similar proportions of RACF staff relative to the population (e.g., Australia, New Zealand) are potentially more likely to report similar clinical outcomes.6 Studies are needed to investigate whether there are differences in the clinical outcomes of antipsychotic and/or benzodiazepine dose reduction on the basis of
RACF resourcing.
7.2 Implications for policy and practice
The findings in this thesis are highly relevant to Australia’s current aged care climate. On the 8th of October 2018, a Royal Commission into aged care quality and safety was announced by the
Australian Government as a result of increased complaints and reports of abuse and neglect.705,
706 An interim report was released on the 31st of October 2019 describing the systematic failure of Australia’s aged care system to meet the needs of older people. In particular, the overuse of psychotropic medications was highlighted as an area that should be immediately addressed.707
In their interim report, the Royal Commission recommended a regular and targeted review of people taking psychotropic medication through Australia’s RMMR and QUM framework.707 Specifically, the recommendation requests that the co-signatories of the next
(seventh) Community Pharmacy Agreement (Australian Government, the Pharmacy Guild of
314 | Page Daniel J. Hoyle Australia, and the Pharmaceutical Society of Australia) review the effectiveness of the current
RMMR and QUM program and:
a) provide funding to enable RMMRs to be performed by accredited pharmacists once
yearly (instead of biennially) or otherwise if there is a significant change to the
resident’s condition or medication regimen;
b) amend the RMMR eligibility criteria to include people receiving residential
respite/transitional care;
c) consider methods to improve the quality and consistency of RMMRs and the QUM
program (e.g. auditing of medication use); and,
d) consider additional government funding to support pharmacists involved in the
implementation and monitoring of recommendations made during the RMMR.707
Additionally, the Royal Commission has recommended increased training and education for aged care staff (e.g., PCAs, nurses, allied health) and GPs on the safe and appropriate management of neuropsychiatric symptoms.707
Despite calls for increased RMMR funding, the impact RMMRs have on the prescribing of antipsychotic and benzodiazepine medications is unclear. In 2009, Nishtala et al. used changes in the Drug Burden Index score as a surrogate marker for reduced prescribing of anticholinergic and sedative medications within a retrospective analysis of a random sample of
500 RMMRs performed by accredited pharmacists.186 Overall, the median Drug Burden Index score reduced from 0.5 (equivalent to one minimum efficacious dose of an anticholinergic or sedative medication per resident) to 0.33 (equivalent to half a minimum efficacious dose of an anticholinergic or sedative medication per resident) after the RMMR (p<0.001).186 However, the impact that RMMRs had specifically on sedative use was not stated. In 2011, a retrospective analysis of a random sample of 500 RMMRs found that whilst CNS medications were commonly implicated in drug-related problems (36.2% of drug-related problems), recommendations to
315 | Page Daniel J. Hoyle resolve these issues were the least likely to be enacted by the GPs (recommendations enacted in 51.9% of cases).605
Furthermore, clinical outcomes associated with reduced psychotropic prescribing as a result of an RMMR intervention have not been reported.185, 186 RMMRs are medication reviews that are delivered independently from staff education/training or other non-pharmacological interventions. Further, in Australia, nurses and aged care staff are not involved in the medication review process, with funding for service being restricted to GPs and pharmacists.
Reduction of psychotropic medications through medication review alone may expose the resident to harm. For example, in 2016, Ballard et al. reported deterioration in neuropsychiatric symptoms and QoL when an antipsychotic review intervention was delivered separately from a social interaction intervention.514, 515 Additionally, in 2005, Holland et al. reported an increased rate of hospital admissions and no improvements in QoL or mortality associated with a home- based medication review intervention among community-dwelling older people recently discharged from hospital.708 Considering the limited evidence base to suggest that RMMRs effectively reduce the use of antipsychotic and/or benzodiazepine prescribing, and the potential for worsened clinical outcomes when antipsychotic review interventions are delivered alone, a multifaceted and targeted approach to antipsychotic and/or benzodiazepine dose reduction, through the QUM program, may be more appropriate.
Australia’s RACF QUM program is flexible in nature. Pharmacists, in consultation with
RACFs, are responsible for selecting the QUM activities delivered each quarter. For this reason, it is difficult to report on the effectiveness of the QUM program, as a whole, to reduce antipsychotic and/or benzodiazepine use in RACFs.709 Additionally, the QUM program, in its current form, is not explicitly designed to require evidence-based practice. Limited alignment with the evidence base poses an increased risk of inequitable outcomes across RACFs. Thus, the most recent evaluation of the QUM program has called for increased alignment with the
316 | Page Daniel J. Hoyle evidence base and implementation of a performance measurement system using Key
Performance Indicators (KPIs) to assess the effectiveness of QUM.709
RedUSe, which aligns closely with the QUM program, addresses both these recommendations.50 RedUSe provides evidence that the combined use of validated components of the QUM program, notably auditing of prescribing, staff education and a targeted multidisciplinary sedative review, can effectively reduce antipsychotic and/or benzodiazepine medication50 without deleterious effects for the resident or staff, alongside possible financial savings. Considering close alignment with the already established QUM program, there is potential for RedUSe to be expanded further as a dedicated QUM service.
With an earlier follow-up study suggesting that antipsychotic prevalence increases to pre- intervention levels 12 months post-intervention,156 RedUSe should be repeated on an annual to biennial basis in order to maintain a sustainable effect on antipsychotic prescribing. Should
RedUSe (or a similar strategy) be incorporated into the QUM program, evaluation of the program should include robust clinical and economic outcome measures to further explore our findings in a larger representative sample.
Internationally, a high prevalence of antipsychotic and/or benzodiazepine use has been reported in RACFs.10, 16, 138, 257, 471, 710 Whilst significant barriers to a reduction in the use of these medications include the belief among RACF staff and GPs that the initial symptoms (e.g., agitation, anxiety, sleep disturbances) and QoL may worsen,44, 45, 170, 187 most interventions that have been implemented in the UK, USA and in Europe to reduce the use of these medications have not reported on these clinical outcomes.47, 137 Consequently, the research reported in this thesis has important practice implications in an international context to mitigate barriers to antipsychotic and benzodiazepine dose reduction.
Our findings suggest that carefully considered antipsychotic and benzodiazepine dose reduction is achievable, via a multifaceted intervention (RedUSe), without deleterious clinical
317 | Page Daniel J. Hoyle outcomes for most RACF residents. Instead, we noted possible improvements in physically nonaggressive behaviour, QoL and occupational disruptiveness related to agitation with antipsychotic dose reduction. Additionally, verbal aggression, physically nonaggressive behaviour and occupational disruptiveness related to night-time behaviour tended to reduce with benzodiazepine dose reduction. Overall, these findings should provide a degree of confidence that these medications can be reduced, in most cases, without harm to the resident or increased disruptiveness of RACF staff.
7.3 Conclusion
The findings in this thesis support the growing body of evidence that antipsychotic and benzodiazepine dose reduction in most residents of RACFs is unlikely to result in clinical deterioration. Furthermore, there was no evidence of increased staff disruptiveness or worsened job satisfaction. Lastly, antipsychotic and/or benzodiazepine dose reduction is potentially cost saving due to lower costs related to medications and hospitalisations. Possible improvements in agitation, QoL and occupational disruptiveness related to agitation with antipsychotic dose reduction, and improved agitation and occupational disruptiveness related to night-time behaviour with benzodiazepine dose reduction require further evaluation in larger, controlled, prospective studies.
318 | Page Daniel J. Hoyle References
1. Australian Institute of Health and Welfare. Australia's changing age and gender profile [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2017 [cited 2017 Aug 02]. Available from: http://www.aihw.gov.au/ageing/older-australia-at-a- glance/demographics/age-and-gender/. 2. Australian Bureau of Statistics. Population projections, Australia, 2012 (base) to 2101. [Internet]. Canberra (AU): Commonwealth of Australia; 2013 [cited 2017 Aug 15]. Available from: http://www.abs.gov.au/ausstats/[email protected]/Lookup/3222.0main+features32012%20(b ase)%20to%202101. 3. Life expectancy at birth, total (years) [Internet]. [place unknown]: The World Bank Group; 2019 [cited 2017 Aug 06]. Available from: http://data.worldbank.org/indicator/SP.DYN.LE00.IN. 4. Australian Bureau of Statistics. Australian demographic statistics, Jun 2017 [Internet]. Canberra (AU): Commonwealth of Australia; 2018 [cited 2018 Jul 25]. Available from: http://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/3101.0Jun%202017?OpenD ocument. 5. Australian Institute of Health and Welfare. Explore people using aged care [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2018 [cited 2018 Apr 30]. Available from: https://www.gen-agedcaredata.gov.au/Topics/People-using-aged- care/Explore-people-using-aged-care. 6. Australian Institute of Health and Welfare. International health data comparisons, 2018 [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2018 [updated 09/11/2018; cited 2020 Aug 3]. Available from: https://www.aihw.gov.au/reports/international-comparisons/international-health- data-comparisons-2018/contents/long-term-care. 7. Australian Institute of Health and Welfare. People's care needs in aged care [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2018 [cited 2018 Jul 31]. Available from: https://gen-agedcaredata.gov.au/Topics/Care-needs-in-aged-care. 8. GEN Aged Care Data. People's care needs in aged care [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2019 [cited 2019 Dec 27]. Available from: https://www.gen-agedcaredata.gov.au/Topics/Care-needs-in-aged-care. 9. Hilmer SN, Gnjidic D. Rethinking psychotropics in nursing homes. Med J Aust. 2013;198(2):77. 10. Westbury J, Gee P, Ling T, Kitsos A, Peterson G. More action needed: psychotropic prescribing in Australian residential aged care. Aust N Z J Psychiatry. 2019;53(2):136-47. 11. Westbury J, Jackson S, Peterson G. Psycholeptic use in aged care homes in Tasmania, Australia. J Clin Pharm Ther. 2010;35(2):189-93. 12. Westbury J, Beld K, Jackson S, Peterson G. Review of psychotropic medication in Tasmanian residential aged care facilities. Australas J Ageing. 2010;29(2):72-6. 13. Snowdon J, Galanos D, Vaswani D. Patterns of psychotropic medication use in nursing homes: surveys in Sydney, allowing comparisons over time and between countries. Int Psychogeriatr. 2011;23(9):1520-5. 14. Snowdon J, Day S, Baker W. Why and how antipsychotic drugs are used in 40 Sydney nursing homes. Int J Geriatr Psychiatry. 2005;20(12):1146-52. 15. Snowdon J, Vaughan R, Miller R, Burgess EE, Tremlett P. Psychotropic drug use in Sydney nursing homes. Med J Aust. 1995;163(2):70-2. 16. McMaster M, Fielding E, Lim D, Moyle W, Beattie E. A cross-sectional examination of the prevalence of psychotropic medications for people living with dementia in
319 | Page Daniel J. Hoyle Australian long-term care facilities: issues of concern. Int Psychogeriatr. 2018;30(7):1019-26. 17. WHO Collaborating Centre for Drug Statistics Methodology. ATC/DDD Index [Internet]. Oslo (NO): Norwegian Institute of Public Health; 2019 [updated 2019 Dec 16; cited 2019 Dec 20]. Available from: https://www.whocc.no/atc_ddd_index/?code=N05&showdescription=yes. 18. Psychotropic Expert Group. Therapeutic guidelines: Schizophrenia and related psychoses [Internet]. Melbourne (AU): Therapeutic Guidelines Limited; 2013 [updated 2019 Dec; cited 2020 Oct 3]. Available from: https://tgldcdp.tg.org.au.acs.hcn.com.au/viewTopic?topicfile=schizophrenia-and- related-psychoses&guidelineName=Psychotropic&topicNavigation=navigateTopic. 19. Psychotropic Expert Group. Therapeutic guidelines: Bipolar disorder [Internet]. Melbourne (AU): Therapeutic Guidelines Limited; 2013 [cited 2020 Oct 3]. Available from: https://tgldcdp.tg.org.au.acs.hcn.com.au/viewTopic?topicfile=bipolar- disorder&guidelineName=Psychotropic&topicNavigation=navigateTopic#toc_d1e561. 20. Leucht S, Leucht C, Huhn M, Chaimani A, Mavridis D, Helfer B, et al. Sixty years of placebo-controlled antipsychotic drug trials in acute schizophrenia: systematic review, Bayesian meta-analysis, and meta-regression of efficacy predictors. Am J Psychiatry. 2017;174(10):927-42. 21. Lindström L, Lindström E, Nilsson M, Höistad M. Maintenance therapy with second generation antipsychotics for bipolar disorder - a systematic review and meta-analysis. J Affect Disord. 2017;213:138-50. 22. Ballard CG, Waite J, Birks J. Atypical antipsychotics for aggression and psychosis in Alzheimer's disease. Cochrane Database Syst Rev. 2006;1:CD00476. 23. Maher AR, Maglione M, Bagley S, Suttorp M, Hu JH, Ewing B, et al. Efficacy and comparative effectiveness of atypical antipsychotic medications for off-label uses in adults: a systematic review and meta-analysis. JAMA. 2011;306(12):1359-69. 24. Sultzer DL, Davis SM, Tariot PN, Dagerman KS, Lebowitz BD, Lyketsos CG, et al. Clinical symptom responses to atypical antipsychotic medications in Alzheimer's disease: phase 1 outcomes from the CATIE-AD effectiveness trial. Am J Psychiatry. 2008;165(7):844-54. 25. Helvik A-S, Šaltytė Benth J, Wu B, Engedal K, Selbæk G. Persistent use of psychotropic drugs in nursing home residents in Norway. BMC Geriatr. 2017;17(1):52. 26. Chen L, Bell JS, Visvanathan R, Hilmer SN, Emery T, Robson L, et al. The association between benzodiazepine use and sleep quality in residential aged care facilities: a cross- sectional study. BMC Geriatr. 2016;16(1):196. 27. Defrancesco M, Marksteiner J, Fleischhacker WW, Blasko I. Use of benzodiazepines in Alzheimer's disease: a systematic review of literature. Int J Neuropsychopharmacol. 2015;18(10):1-11. 28. Psychotropic Expert Group. Therapeutic guidelines: Dementia [Internet]. Melbourne (AU): Therapeutic Guidelines Limited; 2013 [updated 2015 Oct; cited 2017 Jun 16]. Available from: https://tgldcdp.tg.org.au.acs.hcn.com.au/viewTopic?topicfile=dementia&guidelineNa me=Psychotropic#toc_d1e345. 29. Psychotropic Expert Group. Therapeutic guidelines: Insomnia, parasomnias and jet lag [Internet]. Melbourne (AU): Therapeutic Guidelines Limited; 2013 [updated 2019 Jun; cited 2019 Jun 16]. Available from: https://tgldcdp.tg.org.au.acs.hcn.com.au/viewTopic?topicfile=insomnia-parasomnias- jet-lag&guidelineName=Psychotropic#toc_d1e47. 30. Psychotropic Expert Group. Therapeutic guidelines: Anxiety and associated disorders [Internet]. Melbourne (AU): Therapeutic Guidelines Limited; 2013 [cited 2017 Jun 16].
320 | Page Daniel J. Hoyle Available from: https://tgldcdp.tg.org.au.acs.hcn.com.au/viewTopic?topicfile=anxiety- disorders&guidelineName=Psychotropic#toc_d1e189. 31. International Psychogeriatric Association. The IPA complete guides to behavioural and psychological symptoms of dementia (BPSD): specialists guide. Northfield, IL: International Psychogeriatric Association; 2015. 32. The Royal Australian College of General Practitioners. Prescribing drugs of dependence in general practice, Part B [Internet]. East Melbourne (AU): The Royal Australian College of General Practitioners; 2015 [cited 2017 Jul 31]. Available from: https://www.racgp.org.au/your-practice/guidelines/drugs-of-dependence-b. 33. The Royal Australian College of General Practitioners. The use of antipsychotics in residential aged care. East Melbourne (AU): The Royal Australian College of General Practitioners; 2011 34. Seitz DP, Gill SS, Herrmann N, Brisbin S, Rapoport MJ, Rines J, et al. Pharmacological treatments for neuropsychiatric symptoms of dementia in long-term care: a systematic review. Int Psychogeriatr. 2013;25(2):185-203. 35. Glass J, Lanctôt KL, Herrmann N, Sproule BA, Busto UE. Sedative hypnotics in older people with insomnia: meta-analysis of risks and benefits. BMJ. 2005;331(7526):1169. 36. Maust DT, Kim HM, Seyfried LS, Chiang C, Kavanagh J, Schneider LS, et al. Antipsychotics, other psychotropics, and the risk of death in patients with dementia: number needed to harm. JAMA Psychiatry. 2015;72(5):438-45. 37. Cox CA, van Jaarsveld HJ, Houterman S, van der Stegen JC, Wasylewicz AT, Grouls RJ, et al. Psychotropic drug prescription and the risk of falls in nursing home residents. J Am Med Dir Assoc. 2016;17(12):1089-93. 38. Westbury JL, Peterson GM. Rethinking psychotropics in nursing homes. Med J Aust. 2013;199(2):98. 39. Snowdon J, Day S, Baker W. Current use of psychotropic medication in nursing homes. Int Psychogeriatr. 2006;18(2):241-50. 40. Hagen BF, Armstrong-Esther C, Quail P, Williams RJ, Norton P, Le Navenec CL, et al. Neuroleptic and benzodiazepine use in long-term care in urban and rural Alberta: characteristics and results of an education intervention to ensure appropriate use. Int Psychogeriatr. 2005;17(4):631-52. 41. Selbaek G, Kirkevold O, Engedal K. The course of psychiatric and behavioral symptoms and the use of psychotropic medication in patients with dementia in Norwegian nursing homes--a 12-month follow-up study. Am J Geriatr Psychiatry. 2008;16(7):528-36. 42. Snowdon J, Vaughan R. Changes in psychotropic medication use in nursing homes over a 9-month period. J Qual Clin Pract. 1997;17(2):83-90. 43. Sawan M, Jeon Y-H, Chen TF. Shaping the use of psychotropic medicines in nursing homes: a qualitative study on organisational culture. Soc Sci Med. 2018;202:70-8. 44. Azermai M, Stichele RR, Van Bortel LM, Elseviers MM. Barriers to antipsychotic discontinuation in nursing homes: an exploratory study. Aging Ment Health. 2014;18(3):346-53. 45. Bourgeois J, Elseviers MM, Azermai M, Van Bortel L, Petrovic M, Vander Stichele RR. Barriers to discontinuation of chronic benzodiazepine use in nursing home residents: perceptions of general practitioners and nurses. Eur Geriatr Med. 2013;5(3):181-7. 46. Thompson Coon J, Abbott R, Rogers M, Whear R, Pearson S, Lang I, et al. Interventions to reduce inappropriate prescribing of antipsychotic medications in people with dementia resident in care homes: a systematic review. J Am Med Dir Assoc. 2014;15(10):706-18. 47. Hoyle DJ, Bindoff IK, Clinnick LM, Peterson GM, Westbury JL. Clinical and economic outcomes of interventions to reduce antipsychotic and benzodiazepine use within nursing homes: a systematic review. Drugs Aging. 2018;35(2):123-34.
321 | Page Daniel J. Hoyle 48. Nishtala PS, McLachlan AJ, Bell JS, Chen TF. Psychotropic prescribing in long-term care facilities: impact of medication reviews and educational interventions. Am J Geriatr Psychiatry. 2008;16(8):621-32. 49. Westbury J, Jackson S, Gee P, Peterson G. An effective approach to decrease antipsychotic and benzodiazepine use in nursing homes: the RedUSe project. Int Psychogeriatr. 2010;22(1):26-36. 50. Westbury JL, Gee P, Ling T, Brown DT, Franks KH, Bindoff I, et al. RedUSe: reducing antipsychotic and benzodiazepine prescribing in residential aged care facilities. Med J Aust. 2018;208(9):398-403. 51. United Nations. World Population Ageing 2019: Highlights. New York: Department of Economic and Social Affairs, Population Division; 2019. 52. World Health Organization. Ageing and life course [Internet]. [place unknown]: World Health Organization; 2017 [cited 2017 Sep 20]. Available from: http://www.who.int/ageing/en/. 53. Parliamentary Budget Office. Australia's ageing population: understanding the fiscal impacts over the next decade. Canberra (AU): Commonwealth of Australia; 2019. Report No.: 02/2019. 54. Sanford AM, Orrell M, Tolson D, Abbatecola AM, Arai H, Bauer JM, et al. An international definition for “nursing home”. J Am Med Dir Assoc. 2015;16(3):181-4. 55. Sawan M, Jeon Y-H, Chen TF. Psychotropic medicines use in Residents And Culture: Influencing Clinical Excellence (PRACTICE) tool©. A development and content validation study. Res Social Adm Pharm. 2019;15(6):691-700. 56. myagedcare. About aged care homes (nursing homes) [Internet]. Australian Government; 2018 [cited 2018 Jul 25]. Available from: https://www.myagedcare.gov.au/aged-care-homes/considering-aged-care-home. 57. Productivity Commission. Report on Government services 2018: part f, chapter 14, aged care services report and attachment tables. Canberra (AU): Australian Government; 2018. 58. Australian Institute of Health and Welfare. Explore people using aged care [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2019 [cited 2019 Dec 27]. Available from: https://www.gen-agedcaredata.gov.au/Topics/People-using-aged- care/Explore-people-using-aged-care. 59. GEN. GEN data: people using aged care [Data file]. Canberra (AU): Australian Institute of Health and Welfare; 2019 [cited 2019 Sep 09]. Available from: https://www.gen- agedcaredata.gov.au/Resources/Access-data/2018/May/GEN-data-people-using- aged-care. 60. Australian Bureau of Statistics. Population by age and sex, Australia, states and territories [Internet]. Canberra (AU): Commonwealth of Australia; 2018 [cited 2019 Dec 27]. Available from: https://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/3101.0Jun%202018?Open Document. 61. Australian Institute of Health and Welfare. Services and places in aged care [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2017 [cited 2017 Oct 18]. Available from: https://www.gen-agedcaredata.gov.au/Topics/Services-and-places-in- aged-care. 62. Australian Institute of Health and Welfare. Residential aged care in Australia 2010-2011: a statistical overview. Canberra (AU): Australian Institute of Health and Welfare; 2012. 63. Australian Institute of Health and Welfare. Residential aged care in Australia 2008-2009: a statistical overview. Canberra (AU): Australian Institute of Health and Welfare; 2010.
322 | Page Daniel J. Hoyle 64. GEN Aged Care Data. Sevices and places in aged care [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2019 [cited 2019 Dec 27]. Available from: https://gen-agedcaredata.gov.au/Topics/Services-and-places-in-aged-care. 65. Mavromaras K, Knight G, Isherwood L, Crettenden A, Flavel J, Karmel T, et al. The aged care workforce, 2016. Canberra: Australian Government Department of Health; 2017. 66. Productivity Commission. Chapter 14 Aged care services. Canberra (AU): Australian Government; 2019. 67. Australian Institute of Health and Welfare. Admissions into aged care [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2018 [cited 2018 Jul 31]. Available from: https://gen-agedcaredata.gov.au/Topics/Admissions-into-aged-care. 68. GEN Aged Care Data. Explore people's care needs in aged care [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2019 [cited 2019 Dec 27]. Available from: https://www.gen-agedcaredata.gov.au/Topics/Care-needs-in-aged-care/Explore-care- needs-in-aged-care. 69. Department of Health. Basic subsidy amount (Aged Care Funding Instrument) [Internet]. Canberra (AU): Australian Government; 2018 [cited 2019 Dec 27]. Available from: https://agedcare.health.gov.au/aged-care-funding/residential-care-subsidy/basic- subsidy-amount-aged-care-funding-instrument. 70. The Columnist. Ageing in place or rorting the system? [Internet]. agedcarecrisis.com; 2015 [cited 2018 Sep 24]. Available from: https://www.agedcarecrisis.com/opinion/the-column/185-ageing-in-place-or-rorting- the-system. 71. O'Keeffe D. Government clamps down on ACFI claims [Internet]. Australian Ageing Agenda; 2015 [cited 2018 Sep 24]. Available from: https://www.australianageingagenda.com.au/2015/12/16/government-clamps-down- on-acfi-claims/. 72. Department of Health and Ageing. Report to the Minister for Ageing on residential care and people with psychogeriatric disorders. Canberra (AU): Australian Government; 2008. 73. Jeste DV, Alexopoulos GS, Bartels SJ, et al. Consensus statement on the upcoming crisis in geriatric mental health: research agenda for the next 2 decades. Arch Gen Psychiatry. 1999;56(9):848-53. 74. Hall KA, Hassett AM. Assessing and managing old age psychiatric disorders in community practice. Med J Aust. 1998;168(6):299-305. 75. American Psychiatic Association. Neurocognitive disorders. In Diagnostic and Statistical Manual of Mental Disorders [Internet]. Arlington, VA: American Psychiatric Publishing; 2013 [cited 2020 Jan 14]. Available from: https://dsm.psychiatryonline.org/doi/full/10.1176/appi.books.9780890425596.dsm17. 76. World Health Organization. International statistical classification of diseases and related health problems 10th revision (ICD-10) version for 2010. Oslo (NO): World Health Organization; 2010. 77. The Royal Australian College of General Practitioners. Medical care of older persons in residential aged care facilities (silver book). 2006. 78. Australian Institute of Health and Welfare. Dementia [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2019 [updated 2019 Aug 15; cited 2019 Dec 30]. Available from: https://www.aihw.gov.au/reports-data/health-conditions- disability-deaths/dementia/overview. 79. Lie D. Mental health issues in Australian nursing homes. Aust Fam Physician. 2003;32(7):519-22.
323 | Page Daniel J. Hoyle 80. Selbaek G, Kirkevold O, Engedal K. The prevalence of psychiatric symptoms and behavioural disturbances and the use of psychotropic drugs in Norwegian nursing homes. Int J Geriatr Psychiatry. 2007;22(9):843-9. 81. Brodaty H, Draper B, Saab D, Low LF, Richards V, Paton H, et al. Psychosis, depression and behavioural disturbances in Sydney nursing home residents: prevalence and predictors. Int J Geriatr Psychiatry. 2001;16(5):504-12. 82. Pitkala KH, Laurila JV, Strandberg TE, Tilvis RS. Behavioral symptoms and the administration of psychotropic drugs to aged patients with dementia in nursing homes and in acute geriatric wards. Int Psychogeriatr. 2004;16(1):61-74. 83. Dementia Australia. Types of dementia [Internet]. [place unknown]: Dementia Australia; 2020 [cited 2020 Jan 27]. Available from: https://www.dementia.org.au/information/about-dementia/types-of-dementia. 84. Royal Commission into Aged Care Quality and Safety. Dementia in Australia: nature, prevalence and care. Canberra (AU): Commonwealth of Australia; 2019. 85. Arvanitakis Z, Shah RC, Bennett DA. Diagnosis and management of dementia: review. JAMA. 2019;322(16):1589-99. 86. Schneider JA, Arvanitakis Z, Leurgans SE, Bennett DA. The neuropathology of probable Alzheimer disease and mild cognitive impairment. Ann Neurol. 2009;66(2):200-8. 87. Querfurth HW, LaFerla FM. Alzheimer's disease. N Engl J Med. 2010;362(4):329-44. 88. Reitz C, Brayne C, Mayeux R. Epidemiology of Alzheimer disease. Nat Rev Neurol. 2011;7(3):137-52. 89. Alzheimer's Association. 2019 Alzheimer's disease facts and figures. Alzheimers Dement. 2019;15(3):321-87. 90. Goedert M, Spillantini MG. A century of Alzheimer's disease. Science. 2006;314(5800):777-81. 91. Whitehouse PJ, Price DL, Clark AW, Coyle JT, DeLong MR. Alzheimer disease: evidence for selective loss of cholinergic neurons in the nucleus basalis. Ann Neurol. 1981;10(2):122-6. 92. Boeve BF. Update on the diagnosis and management of sleep disturbances in dementia. Sleep Med Clin. 2008;3(3):347-60. 93. Albert MS, DeKosky ST, Dickson D, Dubois B, Feldman HH, Fox NC, et al. The diagnosis of mild cognitive impairment due to Alzheimer's disease: recommendations from the National Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimers Dement. 2011;7(3):270-9. 94. McKhann GM, Knopman DS, Chertkow H, Hyman BT, Jack Jr CR, Kawas CH, et al. The diagnosis of dementia due to Alzheimer's disease: recommendations from the National Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimers Dement. 2011;7(3):263-9. 95. Henriksen A-L, St. Dennis C, Setter SM, Tran JT. Dementia with lewy bodies: therapeutic opportunities and pitfalls. Consult Pharm. 2006;21(7):563-75. 96. Cerejeira J, Lagarto L, Mukaetova-Ladinska EB. Behavioral and psychological symptoms of dementia. Front Neurol. 2012;3:1-21. 97. Seniors Health Strategic Clinical Network. Responsive behaviours [Internet]. St. Albert, AB: Alberta Health Services; 2016 [cited 2018 Sep 20]. Available from: https://www.albertahealthservices.ca/assets/about/scn/ahs-scn-srs-responsive- behaviours.pdf. 98. Burns K, Jayasinha R, Tsang R, Brodaty H. Behaviour management a guide to good practice: managing behavioural and psychological symptoms of dementia. Canberra (AU): Australian Government; 2012.
324 | Page Daniel J. Hoyle 99. Harrison SL, Cations M, Jessop T, Hilmer SN, Sawan M, Brodaty H. Approaches to deprescribing psychotropic medications for changed behaviours in long-term care residents living with dementia. Drugs Aging. 2018;36(2):125-36. 100. Savva GM, Zaccai J, Matthews FE, Davidson JE, McKeith I, Brayne C. Prevalence, correlates and course of behavioural and psychological symptoms of dementia in the population. Br J Psychiatry. 2009;194(3):212-9. 101. Declercq T, Petrovic M, Azermai M, Vander Stichele R, De Sutter AI, van Driel ML, et al. Withdrawal versus continuation of chronic antipsychotic drugs for behavioural and psychological symptoms in older people with dementia. Cochrane Database Syst Rev. 2013;3:CD007726. 102. Liperoti R, Pedone C, Corsonello A. Antipsychotics for the treatment of behavioral and psychological symptoms of dementia (BPSD). Curr Neuropharmacol. 2008;6(2):117-24. 103. Seitz D, Purandare N, Conn D. Prevalence of psychiatric disorders among older adults in long-term care homes: a systematic review. Int Psychogeriatr. 2010;22(7):1025-39. 104. Van Leeuwen E, Petrovic M, van Driel ML, De Sutter AI, Vander Stichele R, Declercq T, et al. Withdrawal versus continuation of long-term antipsychotic drug use for behavioural and psychological symptoms in older people with dementia. Cochrane Database Syst Rev. 2018;3:CD007726. 105. Zuidema SU, Derksen E, Verhey FR, Koopmans RT. Prevalence of neuropsychiatric symptoms in a large sample of Dutch nursing home patients with dementia. Int J Geriatr Psychiatry. 2007;22(7):632-8. 106. The Royal Australian & New Zealand College of Psychiatrists. Professional practice guideline 10: Antipsychotic medications as a treatment of behavioural and psychological symptoms of dementia [Internet]. Melbourne (AU): The Royal Australian & New Zealand College of Psychiatrists; 2016 [cited 2018 Apr 26]. Available from: https://www.ranzcp.org/Files/Resources/College_Statements/Practice_Guidelines/pg 10-pdf.aspx. 107. Quinn J, editor. Dementia. Hoboken, NJ: Wiley; 2013. 108. Jeste DV, Meeks TW, Kim DS, Zubenko GS. Research agenda for DSM-V: diagnostic categories and criteria for neuropsychiatric syndromes in dementia. J Geriatr Psychiatry Neurol. 2006;19(3):160-71. 109. Kales HC, Gitlin LN, Lyketsos CG. Assessment and management of behavioral and psychological symptoms of dementia. BMJ. 2015;350:h369. 110. Tible OP, Riese F, Savaskan E, von Gunten A. Best practice in the management of behavioural and psychological symptoms of dementia. Ther Adv Neurol Disord. 2017;10(8):297-309. 111. Ballard C, Neill D, O’brien J, McKeith IG, Ince P, Perry R. Anxiety, depression and psychosis in vascular dementia: prevalence and associations. J Affect Disord. 2000;59(2):97-106. 112. Srikanth S, Nagaraja AV, Ratnavalli E. Neuropsychiatric symptoms in dementia- frequency, relationship to dementia severity and comparison in Alzheimer's disease, vascular dementia and frontotemporal dementia. J Neurol Sci. 2005;236(1-2):43-8. 113. Rockwell E, Choure J, Galasko D, Olichney J, Jeste DV. Psychopathology at initial diagnosis in dementia with Lewy bodies versus Alzheimer disease: comparison of matched groups with autopsy-confirmed diagnoses. Int J Geriatr Psychiatry. 2000;15(9):819-23. 114. Chaudhury H, Cooke HA, Cowie H, Razaghi L. The influence of the physical environment on residents with dementia in long-term care settings: a review of the empirical literature. Gerontologist. 2018;58(5):e325-e37.
325 | Page Daniel J. Hoyle 115. Zuidema SU, de Jonghe JF, Verhey FR, Koopmans RT. Environmental correlates of neuropsychiatric symptoms in nursing home patients with dementia. Int J Geriatr Psychiatry. 2010;25(1):14-22. 116. Zuidema S, Koopmans R, Verhey F. Prevalence and predictors of neuropsychiatric symptoms in cognitively impaired nursing home patients. J Geriatr Psychiatry Neurol. 2007;20(1):41-9. 117. Sloane PD, Mitchell CM, Preisser JS, Phillips C, Commander C, Burker E. Environmental correlates of resident agitation in Alzheimer's disease special care units. J Am Geriatr Soc. 1998;46(7):862-9. 118. Jao YL, Liu W, Williams K, Chaudhury H, Parajuli J. Association between environmental stimulation and apathy in nursing home residents with dementia. Int Psychogeriatr. 2019;31(8):1109-20. 119. Algase DL, Beck C, Kolanowski A, Whall A, Berent S, Richards K, et al. Need-driven dementia-compromised behavior: an alternative view of disruptive behavior. Am J Alzheimers Dis Other Demen. 1996;11(6):10-9. 120. Ferreira AR, Simoes MR, Moreira E, Guedes J, Fernandes L. Modifiable factors associated with neuropsychiatric symptoms in nursing homes: the impact of unmet needs and psychotropic drugs. Arch Gerontol Geriatr. 2020;86:103919. 121. Hancock GA, Woods B, Challis D, Orrell M. The needs of older people with dementia in residential care. Int J Geriatr Psychiatry. 2006;21(1):43-9. 122. Mann AH, Schneider J, Mozley CG, Levin E, Blizard R, Netten A, et al. Depression and the response of residential homes to physical health needs. Int J Geriatr Psychiatry. 2000;15(12):1105-12. 123. Husebo BS, Ballard C, Sandvik R, Nilsen OB, Aarsland D. Efficacy of treating pain to reduce behavioural disturbances in residents of nursing homes with dementia: cluster randomised clinical trial. BMJ. 2011;343:d4065. 124. Husebo BS, Ballard C, Aarsland D. Pain treatment of agitation in patients with dementia: a systematic review. Int J Geriatr Psychiatry. 2011;26(10):1012-8. 125. Livingston G, Kelly L, Lewis-Holmes E, Baio G, Morris S, Patel N, et al. A systematic review of the clinical effectiveness and cost-effectiveness of sensory, psychological and behavioural interventions for managing agitation in older adults with dementia. Health Technol Assess. 2014;18(39):1-226, v-vi. 126. van der Linde RM, Dening T, Stephan BC, Prina AM, Evans E, Brayne C. Longitudinal course of behavioural and psychological symptoms of dementia: systematic review. Br J Psychiatry. 2016;209(5):366-77. 127. Selbaek G, Engedal K, Benth JS, Bergh S. The course of neuropsychiatric symptoms in nursing-home patients with dementia over a 53-month follow-up period. Int Psychogeriatr. 2014;26(1):81-91. 128. Voyer P, McCusker J, Cole MG, Monette J, Champoux N, Ciampi A, et al. Behavioral and psychological symptoms of dementia: how long does every behavior last, and are particular behaviors associated with PRN antipsychotic agent use? J Gerontol Nurs. 2014;41(1):22-37. 129. Lyketsos CG, Steinberg M, Tschanz JT, Norton MC, Steffens DC, Breitner JC. Mental and behavioral disturbances in dementia: findings from the Cache County Study on Memory in Aging. Am J Psychiatry. 2000;157(5):708-14. 130. Brodaty H, Heffernan M, Draper B, Reppermund S, Kochan NA, Slavin MJ, et al. Neuropsychiatric symptoms in older people with and without cognitive impairment. J Alzheimers Dis. 2012;31(2):411-20. 131. Cravello L, Palmer K, de Girolamo G, Caltagirone C, Spalletta G. Neuropsychiatric symptoms and syndromes in institutionalized elderly people without dementia. Int Psychogeriatr. 2011;23(3):425-34.
326 | Page Daniel J. Hoyle 132. Geda YE, Roberts RO, Knopman DS, Petersen RC, Christianson TJH, Pankratz VS, et al. Prevalence of neuropsychiatric symptoms in mild cognitive impairment and normal cognitive aging: population-based study. Arch Gen Psychiatry. 2008;65(10):1193-8. 133. Klerman GL, Weissman MM. Increasing rates of depression. JAMA. 1989;261(15):2229- 35. 134. Isaac ML, Larson EB. Medical conditions with neuropsychiatric manifestations. Med Clin North Am. 2014;98(5):1193-208. 135. Aarsland D, Kramberger MG. Neuropsychiatric symptoms in Parkinson's Disease. J Parkinsons Dis. 2015;5(3):659-67. 136. Evrard P, Henrard S, Foulon V, Spinewine A. Benzodiazepine use and deprescribing in Belgian nursing homes: results from the COME-ON study. J Am Geriatr Soc. 2020. 137. Hoyle DJ, Peterson GM, Bindoff IK, Clinnick LM, Bindoff AD, Breen JL. Clinical impact of antipsychotic and benzodiazepine reduction: findings from a multicomponent psychotropic reduction program within long-term aged care. Int Psychogeriatr. 2020:1- 13. 138. van der Spek K, Gerritsen DL, Smalbrugge M, Nelissen-Vrancken MH, Wetzels RB, Smeets CH, et al. Only 10% of the psychotropic drug use for neuropsychiatric symptoms in patients with dementia is fully appropriate. The PROPER I-study. Int Psychogeriatr. 2016;28(10):1589-95. 139. Jones TS, Matias M, Powell J, Jones EG, Fishburn J, Looi JC. Who cares for older people with mental illness? A survey of residential aged care facilities in the Australian Capital Territory: implications for mental health nursing. Int J Ment Health Nurs. 2007;16(5):327-37. 140. Martin B, King D. Who cares for older Australians? A picture of the residential and community based aged care workforce, 2007. Adelaide (AU): National Institute of Labour Studies; 2008. 141. Alzheimer's Australia. The future of Australia's aged care sector workforce. [place unknown]; 2016. 142. Reus VI, Fochtmann LJ, Eyler AE, Hilty DM, Horvitz-Lennon M, Jibson MD, et al. The American Psychiatric Association practice guideline on the use of antipsychotics to treat agitation or psychosis in patients with dementia. Am J Psychiatry. 2016;173(5):543-6. 143. Zuidema SU, de Jonghe JF, Verhey FR, Koopmans RT. Psychotropic drug prescription in nursing home patients with dementia: influence of environmental correlates and staff distress on physicians' prescription behavior. Int Psychogeriatr. 2011;23(10):1632-9. 144. Zwijsen SA, Kabboord A, Eefsting JA, Hertogh CM, Pot AM, Gerritsen DL, et al. Nurses in distress? An explorative study into the relation between distress and individual neuropsychiatric symptoms of people with dementia in nursing homes. Int J Geriatr Psychiatry. 2014;29(4):384-91. 145. Volicer L, Frijters DHM, van der Steen JT. Underdiagnosis and undertreatment of depression in nursing home residents. Eur Geriatr Med. 2011;2(6):332-7. 146. Australian Medical Association. 2017 AMA aged care survey report. Barton (AU); 2018. 147. Aberdeen SM, Leggat SG, Barraclough S. Validating a marking rubric for evaluating staff knowledge of dementia for competency in residential aged care. J Vocat Educ Train. 2009;61(4):535-52. 148. Ervin K, Finlayson S, Cross M. The management of behavioural problems associated with dementia in rural aged care. Collegian. 2012;19(2):85-95. 149. Hsu MC, Moyle W, Creedy D, Venturato L. An investigation of aged care mental health knowledge of Queensland aged care nurses. Int J Ment Health Nurs. 2005;14(1):16-23. 150. Davison TE, McCabe MP, Mellor D, Karantzas G, George K. Knowledge of late-life depression: an empirical investigation of aged care staff. Aging Ment Health. 2009;13(4):577-86.
327 | Page Daniel J. Hoyle 151. Castle NG, Engberg J. Organizational characteristics associated with staff turnover in nursing homes. Gerontologist. 2006;46(1):62-73. 152. Community Affairs References Committee. Care and management of younger and older Australians living with dementia and behavioural and psychological symptoms of dementia (BPSD). Canberra (AU): Commonwealth of Australia; 2014. 153. Leiter MP, Harvie P, Frizzell C. The correspondence of patient satisfaction and nurse burnout. Soc Sci Med. 1998;47(10):1611-7. 154. Zimmerman S, Gruber‐Baldini AL, Hebel JR, Sloane PD, Magaziner J. Nursing home facility risk factors for infection and hospitalization: importance of registered nurse turnover, administration, and social factors. J Am Geriatr Soc. 2002;50(12):1987-95. 155. Access Economics. Nurses in residential aged care. [place unknown]: Access Economics Pty Limited; 2009. 156. Westbury J, Tichelaar L, Peterson G, Gee P, Jackson S. A 12-month follow-up study of "RedUSe": a trial aimed at reducing antipsychotic and benzodiazepine use in nursing homes. Int Psychogeriatr. 2011;23(8):1260-9. 157. Moniz‐Cook E, Agar S, Silver M, Woods R, Wang M, Elston C, et al. Can staff training reduce behavioural problems in residential care for the elderly mentally ill? Int J Geriatr Psychiatry. 1998;13(3):149-58. 158. Chenoweth L, Jeon YH, Merlyn T, Brodaty H. A systematic review of what factors attract and retain nurses in aged and dementia care. J Clin Nurs. 2010;19(1-2):156-67. 159. Australian Nursing & Midwifery Federation. Fact Sheet: A snapshot of residential aged care. 2017. 160. O'Keefe D. Aged care wages: Tackling pay in 'the forgotten industry': Australian Ageing Agenda; 2017 [cited 2018 15/08]. Available from: https://www.australianageingagenda.com.au/2017/02/08/aged-care-wages-tackling- pay-forgotten-industry/. 161. King D, Mavromaras K, Wei Z, He B, Healy J, Macaitis KK, et al. The aged care workforce, 2012. Canberra (AU): Australian Government Department of Health and Ageing; 2012. 162. Cousins J, Bereznicki L, Cooling N, Peterson G. Prescribing of psychotropic medication for nursing home residents with dementia: a general practitioner survey. Clin Interv Aging. 2017;12:1573-8. 163. Looi JC, Macfarlane S. Psychotropic drug use in aged care facilities: a reflection of a systemic problem? Med J Aust. 2014;200(1):13-4. 164. Gadzhanova S, Reed R. Medical services provided by general practitioners in residential aged-care facilities in Australia. Med J Aust. 2007;187(2):92-4. 165. Pearson R, Mullan J, Ujvary E, Bonney A, Dijkmans-Hadley B. Australian general practitioner attitudes to residential aged care facility visiting. Health Soc Care Community. 2018;26(4):e497-e504. 166. Flannery J, Hawthorne M. Crisis in aged care looming [Internet]. Barton (AU): Australian Medical Association; 2018 [cited 2018 Aug 17]. Available from: https://ama.com.au/media/crisis-aged-care. 167. Cody M, Beck C, Svarstad BL. Challenges to the use of nonpharmacologic interventions in nursing homes. Psychiatr Serv. 2002;53(11):1402-6. 168. Burgess S, Davis J, Morgans A. General practice and residential aged care: a qualitative study of barriers to access to care and the role of remuneration. AMJ. 2015;8(5):161-9. 169. Australian Medical Association. 2012 AMA aged care survey report. Barton (AU); 2012. 170. Sawan M, Jeon Y-H, Fois RA, Chen TF. Exploring the link between organizational climate and the use of psychotropic medicines in nursing homes: a qualitative study. Res Social Adm Pharm. 2017;13(3):513-23.
328 | Page Daniel J. Hoyle 171. Wood-Mitchell A, James IA, Waterworth A, Swann A, Ballard C. Factors influencing the prescribing of medications by old age psychiatrists for behavioural and psychological symptoms of dementia: a qualitative study. Age Ageing. 2008;37(5):547-52. 172. Cornege-Blokland E, Kleijer BC, Hertogh CM, van Marum RJ. Reasons to prescribe antipsychotics for the behavioral symptoms of dementia: a survey in Dutch nursing homes among physicians, nurses, and family caregivers. J Am Med Dir Assoc. 2012;13(1):80 e1-6. 173. Janus SIM, van Manen JG, Ijzerman MJ, Bisseling M, Drossaert CHC, Zuidema SU. Determinants of the nurses’ and nursing assistants’ request for antipsychotics for people with dementia. Int Psychogeriatr. 2017;29(3):475-84. 174. Smeets CH, Gerritsen DL, Zuidema SU, Teerenstra S, van der Spek K, Smalbrugge M, et al. Psychotropic drug prescription for nursing home residents with dementia: prevalence and associations with non-resident-related factors. Aging Ment Health. 2017;22(9):1239-46. 175. Levy C, Palat S-IT, Kramer AM. Physician practice patterns in nursing homes. J Am Med Dir Assoc. 2007;8(9):558-67. 176. Hickie IB, Davenport TA, Scott EM, Hadzi-Pavlovic D, Naismith SL, Koschera A. Unmet need for recognition of common mental disorders in Australian general practice. Med J Aust. 2001;175(S1):S18-24. 177. Gleeson LE, Jennings S, Gavin R, McConaghy D, Collins DR. Primary care in nursing homes revisited: survey of the experiences of primary care physicians. Ir Med J. 2014;107(8):234-6. 178. McSweeney K, Jeffreys A, Griffith J, Plakiotis C, Kharsas R, O'Connor DW. Specialist mental health consultation for depression in Australian aged care residents with dementia: a cluster randomized trial. Int J Geriatr Psychiatry. 2012;27(11):1163-71. 179. Roufeil L, Radermacher H. At long last, mental health support for residents [Internet]. [place unknown]: Australian Ageing Agenda; 2018 [cited 2018 Aug 07]. Available from: https://www.australianageingagenda.com.au/2018/05/11/at-long-last-mental-health- support-for-residents/. 180. Margarey J. Nursing home residents denied GP mental health treatment plans and psychological therapy [Internet]. Pyrmont (AU): The Sydney Morning Herald; 2017 [cited 2018 Aug 22]. Available from: https://www.smh.com.au/healthcare/nursing- homes-story-headline-20161228-gtiqc6.html. 181. Department of Health. Psychological treatment services for people with mental illness in residential aged care facilities. Canberra (AU): Commonwealth of Australia; 2018. 182. Pharmacy Programs Administrator. Program rules: residential medication management review and quality use of medicines Canberra (AU): Australian Government Department of Health; 2019 [cited 2019 13/11]. Available from: https://www.ppaonline.com.au/wp-content/uploads/2019/01/RMMR-and-QUM- Program-Rules.pdf. 183. Guidelines for pharmacists providing residential medication management review (RMMR) and quality use of medicine (QUM) services [Internet]. Canberra (AU): Pharmaceutical Society of Australia; 2011 [cited 2020 Jan 28]. Available from: https://www.ppaonline.com.au/wp-content/uploads/2019/01/PSA-RMMR-and-QUM- Guidelines.pdf. 184. Koria LG, Zaidi TS, Peterson G, Nishtala P, Hannah PJ, Castelino R. Impact of medication reviews on inappropriate prescribing in aged care. Curr Med Res Opin. 2018;34(5):833- 8. 185. Chen EY, Wang KN, Sluggett JK, Ilomaki J, Hilmer SN, Corlis M, et al. Process, impact and outcomes of medication review in Australian residential aged care facilities: a systematic review. Australas J Ageing. 2019;38 Suppl 2:9-25.
329 | Page Daniel J. Hoyle 186. Nishtala PS, Hilmer SN, McLachlan AJ, Hannan PJ, Chen TF. Impact of residential medication management reviews on drug burden index in aged-care homes: a retrospective analysis. Drugs Aging. 2009;26(8):677-86. 187. Sawan M, Jeon Y-H, Chen TF. Relationship between organizational culture and the use of psychotropic medicines in nursing homes: a systematic integrative review. Drugs Aging. 2018;35(3):189-211. 188. Boyce RD, Hanlon JT, Karp JF, Kloke J, Saleh A, Handler SM. A review of the effectiveness of antidepressant medications for depressed nursing home residents. J Am Med Dir Assoc. 2012;13(4):326-31. 189. Farina N, Morrell L, Banerjee S. What is the therapeutic value of antidepressants in dementia? A narrative review. Int J Geriatr Psychiatry. 2017;32(1):32-49. 190. WHO Collaborating Centre for Drug Statistics Methodology. Guidelines for ATC classification and DDD assignment 2017. Oslo (NO): Norwegian Institute of Public Health; 2016. 191. Therapeutic Goods Administration. Rivotril (clonazepam) [[Internet]]. Canberra (AU): Australian Government; 1991 [updated 2019 Apr 29; cited 2020 Sept 29]. Available from: https://www.ebs.tga.gov.au/ebs/picmi/picmirepository.nsf/pdf?OpenAgent&id=CP- 2010-PI-01135-3. 192. Australian Medicines Handbook. Clonazepam [Internet]. Adelaide (AU): Australian Medicines Handbook Pty Ltd.; 2018 [updated Feb 04 2018; cited 2018 Feb 04]. Available from: https://amhonline-amh-net-au.ezproxy.utas.edu.au/chapters/neurological- drugs/antiepileptics/benzodiazepines-neurology/clonazepam. 193. Davies SJC, Jacob B, Rudoler D, Zaheer J, de Oliveira C, Kurdyak P. Benzodiazepine prescription in Ontario residents aged 65 and over: a population-based study from 1998 to 2013. Therapeutic advances in psychopharmacology. 2018;8(3):99-114. 194. Therapeutic Goods Administration. Paxam [[Internet]]. Canberra (AU): Australian Government; 2010 [updated 2019 Jun 6; cited 2020 Sept 29]. Available from: https://www.ebs.tga.gov.au/ebs/picmi/picmirepository.nsf/pdf?OpenAgent&id=CP- 2010-PI-04583-3&d=202009291016933. 195. Seppala LJ, Wermelink A, de Vries M, Ploegmakers KJ, van de Glind EM, Daams JG, et al. Fall-risk-increasing drugs: a systematic review and meta-analysis: II. Psychotropics. J Am Med Dir Assoc. 2018;19(4):371.e11-.e17. 196. Markota M, Rummans TA, Bostwick JM, Lapid MI. Benzodiazepine use in older adults: dangers, management, and alternative therapies. Mayo Clin Proc. 2016;91(11):1632-9. 197. de Groot MH, van Campen JP, Moek MA, Tulner LR, Beijnen JH, Lamoth CJ. The effects of fall-risk-increasing drugs on postural control: a literature review. Drugs Aging. 2013;30(11):901-20. 198. Tampi RR, Tampi DJ. Efficacy and tolerability of benzodiazepines for the treatment of behavioral and psychological symptoms of dementia: a systematic review of randomized controlled trials. Am J Alzheimers Dis Other Demen. 2014;29(7):565-74. 199. Jasovic-Gasic M, Vukovic O, Pantovic M, Cvetic T, Maric-Bojovic N. Antipsychotics- history of development and field of indication, new wine-old glassess. Psychiatria Danubina. 2012;24(Suppl 3):12-45. 200. Ramachandraiah CT, Subramaniam N, Tancer M. The story of antipsychotics: past and present. Indian J Psychiatry. 2009;51(4):324-6. 201. Bryan J. After 30 years, clozapine is still best for treatment-resistant patients [Internet]. London (UK): Royal Pharmaceutical Society; 2014 [cited 2020 Sep 8]. Available from: https://www.pharmaceutical-journal.com/cpd-and-learning/learning-article/after-30- years-clozapine-is-still-best-for-treatment-resistant- patients/11132880.article?firstPass=false#fn_link_4.
330 | Page Daniel J. Hoyle 202. Idänpään-Heikkilä J, Alhava E, Olkinuora M, Palva IP. Agranulocytosis during treatment with clozapine. Eur J Clin Pharmacol. 1977;11(3):193-8. 203. Newman WJ, Newman BM. Rediscovering clozapine: after a turbulent history, current guidance on initiating and monitoring. Current Psychiatry. 2016;15(7):42-6. 204. Meyer J, Simpson G. From chlorpromazine to olanzapine: a brief history of antipsychotics. Psychiatr Serv. 1997;48(9):1137-9. 205. Australian Medicines Handbook. Antipsychotics [Internet]. Adelaide (AU): Australian Medicines Handbook Pty Ltd.; 2018 [updated Feb 04 2018; cited 2018 Feb 04]. Available from: https://amhonline-amh-net-au.ezproxy.utas.edu.au/chapters/psychotropic- drugs/antipsychotics?menu=vertical. 206. Mandal N, Singh OP, Sen S. Extrapyramidal side effects with low doses of amisulpride. Indian J Psychiatry. 2014;56(2):197-9. 207. Neil W, Curran S, Wattis J. Antipsychotic prescribing in older people. Age Ageing. 2003;32(5):475-83. 208. Pharmaceutical Benefits Scheme. Clozapine [Internet]. Canberra (AU): Australian Government Department of Health; 2018 [cited 2018 Dec 31]. Available from: http://www.pbs.gov.au/medicine/item/10288L-10289M-10302F-10341G-10358E- 11415T-11422E-11433R-5626D-5627E-5628F-5629G-5630H-6101D-6102E-6417R- 6418T-9632Y. 209. Pharmaceutical Benefits Scheme. Risperidone [Internet]. Canberra (AU): Australian Government Department of Health; 2018 [cited 2018 May 15]. Available from: http://www.pbs.gov.au/medicine/item/8787L-8869T. 210. DiPiro J, Talbert R, Yee G, Matzke G, Wells B, Posey M, editors. Pharmacotherapy: a pathophysiological approach. 10th ed. New York, NY: McGraw-Hill Education; 2017. 211. Miyamoto S, Miyake N, Jarskog LF, Fleischhacker WW, Lieberman JA. Pharmacological treatment of schizophrenia: a critical review of the pharmacology and clinical effects of current and future therapeutic agents. Mol Psychiatry. 2012;17(12):1206-27. 212. Kapur S, Zipursky RB, Remington G. Clinical and theoretical implications of 5-HT2 and D2 receptor occupancy of clozapine, risperidone, and olanzapine in schizophrenia. Am J Psychiatry. 1999;156(2):286-93. 213. Hollingworth SA, Siskind DJ, Nissen LM, Robinson M, Hall WD. Patterns of antipsychotic medication use in Australia 2002-2007. Aust N Z J Psychiatry. 2010;44(4):372-7. 214. O'Halloran J, Britt H, Valenti L. General practitioner consultations at residential aged- care facilities. Med J Aust. 2007;187(2):88-91. 215. Leucht S, Tardy M, Komossa K, Heres S, Kissling W, Salanti G, et al. Antipsychotic drugs versus placebo for relapse prevention in schizophrenia: a systematic review and meta- analysis. The Lancet. 2012;379(9831):2063-71. 216. Cipriani A, Barbui C, Salanti G, Rendell J, Brown R, Stockton S, et al. Comparative efficacy and acceptability of antimanic drugs in acute mania: a multiple-treatments meta- analysis. The Lancet. 2011;378(9799):1306-15. 217. Westbury J, Gee P, Ling T, Bindoff I, Brown D, Franks K, et al. Reducing the use of sedative medication in aged care facilities: Implementation of the 'RedUSe' Project into everyday practice. Hobart (AU): Wicking Dementia Research & Education Centre; 2016. 218. Guideline Adaptation Committee. Clinical practice guidelines and principles of care for people with dementia. Sydney (AU): Guideline Adaptation Committee; 2016. 219. Zuidema SU, Johansson A, Selbaek G, Murray M, Burns A, Ballard C, et al. A consensus guideline for antipsychotic drug use for dementia in care homes. Bridging the gap between scientific evidence and clinical practice. Int Psychogeriatr. 2015;27(11):1849- 59. 220. New South Wales Ministry of Health. Assessment and management of people with behavioural and psychological symptoms of dementia (BPSD): A handbook for NSW
331 | Page Daniel J. Hoyle clinicians. North Sydney (AU): Royal Australian & New Zealand College of Psychiatrists; 2013. 221. The Royal Australian College of General Practitioners. Part A. Behavioural and psychological symptoms of dementia. RACGP aged care clinical guide (silver book). 5th ed. East Melbourne (AU): The Royal Australian College of General Practitioners; 2019. 222. Abraha I, Rimland JM, Trotta FM, Dell'Aquila G, Cruz-Jentoft A, Petrovic M, et al. Systematic review of systematic reviews of non-pharmacological interventions to treat behavioural disturbances in older patients with dementia. The SENATOR-OnTop series. BMJ Open. 2017;7(3):e012759. 223. Watt JA, Goodarzi Z, Veroniki AA, Nincic V, Khan PA, Ghassemi M, et al. Comparative efficacy of interventions for aggressive and agitated behaviors in dementia: a systematic review and network meta-analysis. Ann Intern Med. 2019;171(9):633-42. 224. Janssen-Cilag Ltd. Risperdal 0.5mg Film-Coated Tablets [Internet]. Bucks (UK): Datapharm; 2000 [updated 2018 Sept; cited 2020 Aug 18]. Available from: https://www.medicines.org.uk/emc/product/6856/smpc#companyDetails. 225. Bjerre LM, Farrell B, Hogel M, Graham L, Lemay G, McCarthy L, et al. Deprescribing antipsychotics for behavioural and psychological symptoms of dementia and insomnia. Can Fam Physician. 2018;64(1):17-27. 226. Sylliaas H, Selbaek G, Bergland A. Do behavioral disturbances predict falls among nursing home residents? Aging Clin Exp Res. 2012;24(3):251-6. 227. Sato S, Kakamu T, Hayakawa T, Kumagai T, Hidaka T, Masuishi Y, et al. Predicting falls from behavioral and psychological symptoms of dementia in older people residing in facilities. Geriatr Gerontol Int. 2018;18(11):1573-7. 228. Banerjee S, Smith SC, Lamping DL, Harwood RH, Foley B, Smith P, et al. Quality of life in dementia: more than just cognition. An analysis of associations with quality of life in dementia. J Neurol Neurosurg Psychiatry. 2006;77(2):146-8. 229. Shin IS, Carter M, Masterman D, Fairbanks L, Cummings JL. Neuropsychiatric symptoms and quality of life in Alzheimer disease. Am J Geriatr Psychiatry. 2005;13(6):469-74. 230. Samus QM, Rosenblatt A, Steele C, Baker A, Harper M, Brandt J, et al. The association of neuropsychiatric symptoms and environment with quality of life in assisted living residents with dementia. Gerontologist. 2005;45 Spec No 1(1):19-26. 231. van de Ven-Vakhteeva J, Bor H, Wetzels RB, Koopmans RT, Zuidema SU. The impact of antipsychotics and neuropsychiatric symptoms on the quality of life of people with dementia living in nursing homes. Int J Geriatr Psychiatry. 2013;28(5):530-8. 232. Toot S, Swinson T, Devine M, Challis D, Orrell M. Causes of nursing home placement for older people with dementia: a systematic review and meta-analysis. Int Psychogeriatr. 2017;29(2):195-208. 233. Isik AT, Soysal P, Solmi M, Veronese N. Bidirectional relationship between caregiver burden and neuropsychiatric symptoms in patients with Alzheimer's disease: A narrative review. Int J Geriatr Psychiatry. 2019;34(9):1326-34. 234. van Duinen-van den IJssel J, Mulders A, Smalbrugge M, Zwijsen SA, Appelhof B, Zuidema SU, et al. Nursing staff distress associated with neuropsychiatric symptoms in young- onset dementia and late-onset dementia. J Am Med Dir Assoc. 2018;19(7):627-32. 235. Kennedy BR. Stress and burnout of nursing staff working with geriatric clients in long- term care. J Nurs Scholarsh. 2005;37(4):381-2. 236. Cheng S-T. Dementia caregiver burden: a research update and critical analysis. Curr Psychiatry Rep. 2017;19(9):64. 237. Rapaport P, Livingston G, Hamilton O, Turner R, Stringer A, Robertson S, et al. How do care home staff understand, manage and respond to agitation in people with dementia? A qualitative study. BMJ Open. 2018;8(6):e022260.
332 | Page Daniel J. Hoyle 238. Morris S, Patel N, Baio G, Kelly L, Lewis-holmes E, Omar RZ, et al. Monetary costs of agitation in older adults with Alzheimer’s disease in the UK: prospective cohort study. BMJ Open. 2015;5(3):e007382. 239. Murman DL, Chen Q, Powell MC, Kuo SB, Bradley CJ, Colenda CC. The incremental direct costs associated with behavioral symptoms in AD. Neurology. 2002;59(11):1721-9. 240. Beeri MS, Werner P, Davidson M, Noy S. The cost of behavioral and psychological symptoms of dementia (BPSD) in community dwelling Alzheimer's disease patients. Int J Geriatr Psychiatry. 2002;17(5):403-8. 241. Lonergan E, Luxenberg J, Colford JM, Birks J. Haloperidol for agitation in dementia. Cochrane Database Syst Rev. 2002(2):CD002852. 242. Schneider LS, Dagerman K, Insel PS. Efficacy and adverse effects of atypical antipsychotics for dementia: meta-analysis of randomized, placebo-controlled trials. Am J Geriatr Psychiatry. 2006;14(3):191-210. 243. Dyer SM, Harrison SL, Laver K, Whitehead C, Crotty M. An overview of systematic reviews of pharmacological and non-pharmacological interventions for the treatment of behavioral and psychological symptoms of dementia. Int Psychogeriatr. 2017;30(3):295-309. 244. Schneider LS, Tariot PN, Dagerman KS, Davis SM, Hsiao JK, Ismail MS, et al. Effectiveness of atypical antipsychotic drugs in patients with Alzheimer's disease. N Engl J Med. 2006;355(15):1525-38. 245. Ballard C, Waite J, Birks J. The effectiveness of atypical antipsychotics for the treatment of aggression and psychosis in Alzheimer's disease. Cochrane Database Syst Rev. 2006(1):CD003476-CD. 246. Masopust J, Protopopova D, Valis M, Pavelek Z, Klimova B. Treatment of behavioral and psychological symptoms of dementias with psychopharmaceuticals: a review. Neuropsychiatr Dis Treat. 2018;14:1211-20. 247. US Food and Drug Administration. Public health advisory: deaths with antipsychotics in elderly patients with behavioral disturbances [Internet]. Silver Spring, MD: Department of Health and Human Services; 2005 [cited 2017 Jun 16]. Available from: http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsa ndProviders/ucm053171.htm. 248. Schneider LS, Dagerman KS, Insel P. Risk of death with atypical antipsychotic drug treatment for dementia: meta-analysis of randomized placebo-controlled trials. JAMA. 2005;294(15):1934-43. 249. US Food and Drug Administration. Information for healthcare professionals: conventional antipsychotics [Internet]. Silver Spring, MD: Department of Health and Human Services; 2008 [cited 2017 Aug 21]. Available from: https://www.fda.gov/Drugs/DrugSafety/ucm124830.htm. 250. Gill SS, Bronskill SE, Normand S-LT, Anderson GM, Sykora K, Lam K, et al. Antipsychotic drug use and mortality in older adults with dementia. Ann Intern Med. 2007;146(11):775-86. 251. Schneeweiss S, Setoguchi S, Brookhart A, Dormuth C, Wang PS. Risk of death associated with the use of conventional versus atypical antipsychotic drugs among elderly patients. CMAJ. 2007;176(5):627-32. 252. Yang C, Hao Z, Tian J, Zhang W, Li W, Zhang LL, et al. Does antipsychotic drug use increase the risk of long term mortality? A systematic review and meta-analysis of observational studies. Oncotarget. 2018;9(19):15101-10. 253. Ralph SJ, Espinet AJ. Increased all-cause mortality by antipsychotic drugs: updated review and meta-analysis in dementia and general mental health care. J Alzheimers Dis Rep. 2018;2(1):287-312.
333 | Page Daniel J. Hoyle 254. Jennum P, Baandrup L, Ibsen R, Kjellberg J. Increased all-cause mortality with use of psychotropic medication in dementia patients and controls: a population-based register study. Eur Neuropsychopharmacol. 2015;25(11):1906-13. 255. Liperoti R, Onder G, Landi F, Lapane KL, Mor V, Bernabei R, et al. All-cause mortality associated with atypical and conventional antipsychotics among nursing home residents with dementia: a retrospective cohort study. J Clin Psychiatry. 2009;70(10):1340-7. 256. Yunusa I, Alsumali A, Garba AE, Regestein QR, Eguale T. Assessment of reported comparative effectiveness and safety of atypical antipsychotics in the treatment of behavioral and psychological symptoms of dementia: a network meta-analysis. JAMA Netw Open. 2019;2(3):e190828. 257. Banerjee S. The use of antipsychotic medication for people with dementia: Time for action. London (UK): Department of Health; 2009. 258. Crichton SL, Bray BD, McKevitt C, Rudd AG, Wolfe CD. Patient outcomes up to 15 years after stroke: survival, disability, quality of life, cognition and mental health. J Neurol Neurosurg Psychiatry. 2016;87(10):1091-8. 259. Medicines & Healthcare Products Regulatory Agency. Guidance: Antipsychotic medicines [Internet]. GOV.UK; 2005 [cited 2020 Jan 03]. Available from: https://www.gov.uk/government/publications/antipsychotic-medicines-licensed- products-uses-and-side-effects/antipsychotic-medicines. 260. Douglas IJ, Smeeth L. Exposure to antipsychotics and risk of stroke: self controlled case series study. BMJ (Clinical research ed). 2008;337:a1227-a. 261. Shin J-Y, Choi N-K, Lee J, Seong J-M, Park M-J, Lee SH, et al. Risk of ischemic stroke associated with the use of antipsychotic drugs in elderly patients: a retrospective cohort study in Korea. PLoS One. 2015;10(3):e0119931. 262. Zivkovic S, Koh CH, Kaza N, Jackson CA. Antipsychotic drug use and risk of stroke and myocardial infarction: a systematic review and meta-analysis. BMC Psychiatry. 2019;19(1):189. 263. Wallaschofski H, Donne M, Eigenthaler M, Hentschel B, Faber R, Stepan H, et al. PRL as a novel potent cofactor for platelet aggregation. J Clin Endocrinol Metab. 2001;86(12):5912-9. 264. Hagg S, Jonsson AK, Spigset O. Risk of venous thromboembolism due to antipsychotic drug therapy. Expert Opin Drug Saf. 2009;8(5):537-47. 265. Burghardt KJ, Seyoum B, Mallisho A, Burghardt PR, Kowluru RA, Yi Z. Atypical antipsychotics, insulin resistance and weight; a meta-analysis of healthy volunteer studies. Prog Neuropsychopharmacol Biol Psychiatry. 2018;83:55-63. 266. World Health Organization. Falls [Internet]. [place unknown]: World Health Organization; 2016 [cited 2017 Mar 09]. Available from: http://www.who.int/mediacentre/factsheets/fs344/en/. 267. Sterke CS, van Beeck EF, van der Velde N, Ziere G, Petrovic M, Looman CW, et al. New insights: dose-response relationship between psychotropic drugs and falls: a study in nursing home residents with dementia. J Clin Pharmacol. 2012;52(6):947-55. 268. Rubenstein LZ, Josephson KR, Robbins AS. Falls in the nursing home. Ann Intern Med. 1994;121(6):442-51. 269. Weir E, Culmer L. Fall prevention in the elderly population. Can Med Assoc J. 2004;171(7):724. 270. Australian Institute of Health and Welfare. Trends in hospitalisations due to falls by older people, Australia 2002-03 to 2012-13 [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2017 [cited 2019 Jul 02]. Available from: https://www.aihw.gov.au/reports/injury/hospitalisations-falls-older-people-2002- 2013/contents/summary.
334 | Page Daniel J. Hoyle 271. ANZHFR bi-national annual report for hip fracture care 2018. Australian & New Zealand Hip Fracture Registry; 2018. 272. Sylliaas H, Idland G, Sandvik L, Forsen L, Bergland A. Does mortality of the aged increase with the number of falls? Results from a nine-year follow-up study. Eur J Epidemiol. 2009;24(7):351-5. 273. Leucht S, Corves C, Arbter D, Engel RR, Li C, Davis JM. Second-generation versus first- generation antipsychotic drugs for schizophrenia: a meta-analysis. Lancet. 2009;373(9657):31-41. 274. Naidoo U, Goff DC, Klibanski A. Hyperprolactinemia and bone mineral density: the potential impact of antipsychotic agents. Psychoneuroendocrinology. 2003;28(Suppl 2):97-108. 275. Burns MJ. The pharmacology and toxicology of atypical antipsychotic agents. J Toxicol Clin Toxicol. 2001;39(1):1-14. 276. Meltzer HY, Matsubara S, Lee JC. Classification of typical and atypical antipsychotic drugs on the basis of dopamine D-1, D-2 and serotonin2 pKi values. J Pharmacol Exp Ther. 1989;251(1):238-46. 277. Howland R. Prescribing psychotropic medications for elderly patients. J Psychosoc Nurs Ment Health Serv. 2009;47(11):17-20. 278. Leipzig RM, Cumming RG, Tinetti ME. Drugs and falls in older people: a systematic review and meta-analysis: I. Psychotropic drugs. J Am Geriatr Soc. 1999;47(1):30-9. 279. Woolcott JC, Richardson KJ, Wiens MO, Patel B, Marin J, Khan KM, et al. Meta-analysis of the impact of 9 medication classes on falls in elderly persons. Arch Intern Med. 2009;169(21):1952-60. 280. Hartikainen S, Lönnroos E, Louhivuori K. Medication as a risk factor for falls: critical systematic review. J Gerontol A Biol Sci Med Sci. 2007;62(10):1172-81. 281. Bloch F, Thibaud M, Dugué B, Brèque C, Rigaud A-S, Kemoun G. Psychotropic drugs and falls in the elderly people: updated literature review and meta-analysis. J Aging Health. 2011;23(2):329-46. 282. Dzahini O, Singh N, Taylor D, Haddad PM. Antipsychotic drug use and pneumonia: systematic review and meta-analysis. J Psychopharmacol. 2018;32(11):1167-81. 283. Nose M, Recla E, Trifiro G, Barbui C. Antipsychotic drug exposure and risk of pneumonia: a systematic review and meta-analysis of observational studies. Pharmacoepidemiol Drug Saf. 2015;24(8):812-20. 284. Trifiro G. Antipsychotic drug use and community-acquired pneumonia. Curr Infect Dis Rep. 2011;13(3):262-8. 285. Trifiro G, Gambassi G, Sen EF, Caputi AP, Bagnardi V, Brea J, et al. Association of community-acquired pneumonia with antipsychotic drug use in elderly patients: a nested case-control study. Ann Intern Med. 2010;152(7):418-25, w139-40. 286. Yang SY, Liao YT, Liu HC, Chen WJ, Chen CC, Kuo CJ. Antipsychotic drugs, mood stabilizers, and risk of pneumonia in bipolar disorder: a nationwide case-control study. J Clin Psychiatry. 2013;74(1):e79-86. 287. Kuo CJ, Yang SY, Liao YT, Chen WJ, Lee WC, Shau WY, et al. Second-generation antipsychotic medications and risk of pneumonia in schizophrenia. Schizophr Bull. 2013;39(3):648-57. 288. Sultana J, Calabro M, Garcia-Serna R, Ferrajolo C, Crisafulli C, Mestres J, et al. Biological substantiation of antipsychotic-associated pneumonia: systematic literature review and computational analyses. PLoS One. 2017;12(10):e0187034. 289. Nagamine T. Serum substance P levels in patients with chronic schizophrenia treated with typical or atypical antipsychotics. Neuropsychiatr Dis Treat. 2008;4(1):289-94.
335 | Page Daniel J. Hoyle 290. Ishida T, Uchida H, Suzuki T, Nakajima S, Watanabe K, Kashima H. Plasma substance P level in patients with schizophrenia: a cross-sectional study. Psychiatry Clin Neurosci. 2011;65(5):526-8. 291. Kruk J, Sachdev P, Singh S. Neuroleptic-induced respiratory dyskinesia. J Neuropsychiatry Clin Neurosci. 1995;7(2):223-9. 292. Maher S, Cunningham A, O'Callaghan N, Byrne F, Mc Donald C, McInerney S, et al. Clozapine-induced hypersalivation: an estimate of prevalence, severity and impact on quality of life. Ther Adv Psychopharmacol. 2016;6(3):178-84. 293. Davydov L, Botts SR. Clozapine-induced hypersalivation. Ann Pharmacother. 2000;34(5):662-5. 294. Rang H, Dale M, Ritter J, Flower R, editors. Rang and Dale's Pharmacology. 6th ed. Philadelphia, PA: Elsevier; 2007. 295. Ng BJ, Le Couteur DG, Hilmer SN. Deprescribing benzodiazepines in older patients: impact of interventions targeting physicians, pharmacists, and patients. Drugs Aging. 2018;35(6):493-521. 296. Australian Medicines Handbook. Benzodiazepine comparative information [Internet]. Adelaide (AU): Australian Medicines Handbook Pty Ltd.; 2018 [updated Feb 04 2018; cited 2018 Feb 04]. Available from: https://amhonline-amh-net- au.ezproxy.utas.edu.au/chapters/psychotropic-drugs/tables/benzodiazepines- table?menu=vertical. 297. Bourgeois J, Elseviers MM, Azermai M, Van Bortel L, Petrovic M, Vander Stichele RR. Benzodiazepine use in Belgian nursing homes: a closer look into indications and dosages. Eur J Clin Pharmacol. 2012;68(5):833-44. 298. Tampi RR, Tampi DJ, Balachandran S. Antipsychotics, antidepressants, anticonvulsants, melatonin, and benzodiazepines for behavioral and psychological symptoms of dementia: a systematic review of meta-analyses. Curr Treat Options Psychiatry. 2017;4(1):55-79. 299. Stevenson DG, Decker SL, Dwyer LL, Huskamp HA, Grabowski DC, Metzger ED, et al. Antipsychotic and benzodiazepine use among nursing home residents: findings from the 2004 National Nursing Home Survey. Am J Geriatr Psychiatry. 2010;18(12):1078-92. 300. Qaseem A, Kansagara D, Forciea MA, Cooke M, Denberg TD. Management of chronic insomnia disorder in adults: a clinical practice guideline from the American College of Physicians. Ann Intern Med. 2016;165(2):125-33. 301. Riemann D, Baglioni C, Bassetti C, Bjorvatn B, Dolenc Groselj L, Ellis JG, et al. European guideline for the diagnosis and treatment of insomnia. J Sleep Res. 2017;26(6):675-700. 302. Cunnington D, Junge M. Chronic insomnia: diagnosis and non-pharmacological management. BMJ. 2016;355:i5819. 303. Andrews G, Bell C, Boyce P, Gale C, Lampe L, Marwat O, et al. Royal Australian and New Zealand College of Psychiatrists clinical practice guidelines for the treatment of panic disorder, social anxiety disorder and generalised anxiety disorder. Aust N Z J Psychiatry. 2018;52(12):1109-72. 304. Katzman MA, Bleau P, Blier P, Chokka P, Kjernisted K, Van Ameringen M, et al. Canadian clinical practice guidelines for the management of anxiety, posttraumatic stress and obsessive-compulsive disorders. BMC Psychiatry. 2014;14(Suppl 1):S1. 305. University of South Australia. Dementia and changes in behaviour. Therapeutic brief [Internet]. Adelaide (AU): Department of Veterans' Affairs; 2016 [cited 2018 Sep 20]. Available from: https://www.veteransmates.net.au/topic-44-therapeutic-brief. 306. Kalisch Ellett L, Moffat A, Gadzhanova S, Pratt N, Apajee J, Woodward M, et al. Reduction in use of risperidone for dementia in Australia following changed guidelines. Pharmacy. 2019;7(3):E100.
336 | Page Daniel J. Hoyle 307. Maust DT, Kim H, Chiang C, Kales HC. Association of the Centers for Medicare & Medicaid Services’ National Partnership to Improve Dementia Care with the use of antipsychotics and other psychotropics in long-term care in the United States from 2009 to 2014. JAMA Intern Med. 2018;178(5):640-7. 308. Picton JD, Marino AB, Nealy KL. Benzodiazepine use and cognitive decline in the elderly. Am J Health Syst Pharm. 2018;75(1):e6-e12. 309. Australian Medicines Handbook. Benzodiazepines [Internet]. Adelaide (AU): Australian Medicines Handbook Pty Ltd.; 2018 [updated Feb 04 2018; cited 2018 Feb 04]. Available from: https://amhonline-amh-net-au.ezproxy.utas.edu.au/chapters/psychotropic- drugs/drugs-anxiety-sleep-disorders/benzodiazepines?menu=vertical. 310. American Psychiatic Association. Sleep-wake disorders. In Diagnostic and Statistical Manual of Mental Disorders [Internet]. Arlington, VA: American Psychiatric Publishing; 2013 [cited 2017 Jan 24]. Available from: http://dsm.psychiatryonline.org.ezproxy.utas.edu.au/doi/full/10.1176/appi.books.978 0890425596.dsm12. 311. Roth T, Jaeger S, Jin R, Kalsekar A, Stang PE, Kessler RC. Sleep problems, comorbid mental disorders, and role functioning in the National Comorbidity Survey Replication (NCS-R). Biol Psychiatry. 2006;60(12):1364-71. 312. Ohayon MM. Epidemiology of insomnia: what we know and what we still need to learn. Sleep Med Rev. 2002;6(2):97-111. 313. Morin CM, LeBlanc M, Daley M, Gregoire JP, Merette C. Epidemiology of insomnia: prevalence, self-help treatments, consultations, and determinants of help-seeking behaviors. Sleep Med. 2006;7(2):123-30. 314. Skottheim A, Lovheim H, Isaksson U, Sandman PO, Gustafsson M. Insomnia symptoms among old people in nursing homes. Int Psychogeriatr. 2018;30(1):77-85. 315. Monane M, Glynn RJ, Avorn J. The impact of sedative-hypnotic use on sleep symptoms in elderly nursing home residents. Clin Pharmacol Ther. 1996;59(1):83-92. 316. Voyer P, Verreault R, Mengue PN, Morin CM. Prevalence of insomnia and its associated factors in elderly long-term care residents. Arch Gerontol Geriatr. 2006;42(1):1-20. 317. Makhlouf MM, Ayoub AI, Abdel-Fattah MM. Insomnia symptoms and their correlates among the elderly in geriatric homes in Alexandria, Egypt. Sleep Breath. 2007;11(3):187-94. 318. Buysse DJ, Angst J, Gamma A, Ajdacic V, Eich D, Rössler W. Prevalence, course, and comorbidity of insomnia and depression in young adults. Sleep. 2008;31(4):473-80. 319. Morin CM, Bélanger L, LeBlanc M, et al. The natural history of insomnia: a population- based 3-year longitudinal study. Arch Intern Med. 2009;169(5):447-53. 320. Zhang B, Wing YK. Sex differences in insomnia: a meta-analysis. Sleep. 2006;29(1):85- 93. 321. Van Cauter E, Leproult R, Plat L. Age-related changes in slow wave sleep and REM sleep and relationship with growth hormone and cortisol levels in healthy men. JAMA. 2000;284(7):861-8. 322. Ohayon MM, Carskadon MA, Guilleminault C, Vitiello MV. Meta-analysis of quantitative sleep parameters from childhood to old age in healthy individuals: developing normative sleep values across the human lifespan. Sleep. 2004;27(7):1255-73. 323. Luik AI, Zuurbier LA, Hofman A, Van Someren EJ, Tiemeier H. Stability and fragmentation of the activity rhythm across the sleep-wake cycle: the importance of age, lifestyle, and mental health. Chronobiol Int. 2013;30(10):1223-30. 324. Beaulieu-Bonneau S, Hudon C. Sleep disturbances in older adults with mild cognitive impairment. Int Psychogeriatr. 2009;21(4):654-66. 325. Yaffe K, Falvey CM, Hoang T. Connections between sleep and cognition in older adults. Lancet Neurol. 2014;13(10):1017-28.
337 | Page Daniel J. Hoyle 326. Porter VR, Buxton WG, Avidan AY. Sleep, cognition and dementia. Curr Psychiatry Rep. 2015;17(12):97. 327. Brown DT, Westbury JL, Schüz B. Sleep and agitation in nursing home residents with and without dementia. Int Psychogeriatr. 2015;27(12):1945-55. 328. Kaufmann CN, Canham SL, Mojtabai R, Gum AM, Dautovich ND, Kohn R, et al. Insomnia and health services utilization in middle-aged and older adults: results from the Health and Retirement Study. J Gerontol A Biol Sci Med Sci. 2013;68(12):1512-7. 329. Goldman SE, Ancoli-Israel S, Boudreau R, Cauley JA, Hall M, Stone KL, et al. Sleep problems and associated daytime fatigue in community-dwelling older individuals. J Gerontol A Biol Sci Med Sci. 2008;63(10):1069-75. 330. Martin JL, Fiorentino L, Jouldjian S, Josephson KR, Alessi CA. Sleep quality in residents of assisted living facilities: effect on quality of life, functional status, and depression. J Am Geriatr Soc. 2010;58(5):829-36. 331. Haimov I, Hanuka E, Horowitz Y. Chronic insomnia and cognitive functioning among older adults. Behav Sleep Med. 2008;6(1):32-54. 332. Lo CM, Lee PH. Prevalence and impacts of poor sleep on quality of life and associated factors of good sleepers in a sample of older Chinese adults. Health Qual Life Outcomes. 2012;10:72. 333. Buscemi N, Vandermeer B, Friesen C, Bialy L, Tubman M, Ospina M, et al. The efficacy and safety of drug treatments for chronic insomnia in adults: a meta-analysis of RCTs. J Gen Intern Med. 2007;22(9):1335-50. 334. Holbrook AM, Crowther R, Lotter A, Cheng C, King D. Meta-analysis of benzodiazepine use in the treatment of insomnia. Can Med Assoc J. 2000;162(2):225-33. 335. Vinkers CH, Olivier B. Mechanisms underlying tolerance after long-term benzodiazepine use: a future for subtype-selective GABA(A) receptor modulators? Adv Pharmacol Sci. 2012;2012:416864. 336. Bourgeois J, Elseviers MM, Van Bortel L, Petrovic M, Vander Stichele RH. Sleep quality of benzodiazepine users in nursing homes: a comparative study with nonusers. Sleep Med. 2013;14(7):614-21. 337. Manconi M, Ferri R, Miano S, Maestri M, Bottasini V, Zucconi M, et al. Sleep architecture in insomniacs with severe benzodiazepine abuse. Clin Neurophysiol. 2017;128(6):875- 81. 338. Harvey AG, Tang NK. (Mis)perception of sleep in insomnia: a puzzle and a resolution. Psychol Bull. 2012;138(1):77-101. 339. American Psychiatric Association. Anxiety disorders. In Diagnostic and Statistical Manual of Mental Disorders [Internet]. Arlington, VA: American Psychiatric Publishing; 2013 [cited 2017 Dec 6]. 5th ed:[Available from: https://dsm.psychiatryonline.org/doi/full/10.1176/appi.books.9780890425596.dsm05. 340. Creighton AS, Davison TE, Kissane DW. The prevalence of anxiety among older adults in nursing homes and other residential aged care facilities: a systematic review. Int J Geriatr Psychiatry. 2016;31(6):555-66. 341. Kirmizioglu Y, Dogan O, Kugu N, Akyuz G. Prevalence of anxiety disorders among elderly people. Int J Geriatr Psychiatry. 2009;24(9):1026-33. 342. Creighton AS, Davison TE, Kissane DW. The prevalence, reporting, and treatment of anxiety among older adults in nursing homes and other residential aged care facilities. J Affect Disord. 2018;227:416-23. 343. Copeland JRM, Dewey ME, Wood N, Searle R, Davidson IA, McWilliam C. Range of mental illness among the elderly in the community: prevalence in Liverpool using the GMS-AGECAT package. Br J Psychiatry. 1987;150(6):815-23. 344. Gliatto MF. Generalized anxiety disorder. Am Fam Physician. 2000;62(7):1591-600, 602.
338 | Page Daniel J. Hoyle 345. Schweizer E, Rickels K, Weiss S, Zavodnick S. Maintenance drug treatment of panic disorder. I. Results of a prospective, placebo-controlled comparison of alprazolam and imipramine. Arch Gen Psychiatry. 1993;50(1):51-60. 346. Rickels K, Schweizer E, Csanalosi I, Case WG, Chung H. Long-term treatment of anxiety and risk of withdrawal. Prospective comparison of clorazepate and buspirone. Arch Gen Psychiatry. 1988;45(5):444-50. 347. Bandelow B, Michaelis S, Wedekind D. Treatment of anxiety disorders. Dialogues Clin Neurosci. 2017;19(2):93-107. 348. 2019 American Geriatrics Society Beers Criteria Update Expert Panel. American Geriatrics Society 2019 Updated AGS Beers Criteria(R) for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2019;67(4):674-94. 349. Tiller JW. Depression and anxiety. Med J Aust. 2013;199(6 Suppl):S28-31. 350. Gorman JM. Comorbid depression and anxiety spectrum disorders. Depress Anxiety. 1996;4(4):160-8. 351. Slee A, Nazareth I, Bondaronek P, Liu Y, Cheng Z, Freemantle N. Pharmacological treatments for generalised anxiety disorder: a systematic review and network meta- analysis. Lancet. 2019;393(10173):768-77. 352. Chen T-R, Huang H-C, Hsu J-H, Ouyang W-C, Lin K-C. Pharmacological and psychological interventions for generalized anxiety disorder in adults: a network meta-analysis. J Psychiatr Res. 2019;118:73-83. 353. Rickels K, Downing R, Schweizer E, Hassman H. Antidepressants for the treatment of generalized anxiety disorder. A placebo-controlled comparison of imipramine, trazodone, and diazepam. Arch Gen Psychiatry. 1993;50(11):884-95. 354. Baldwin DS, Anderson IM, Nutt DJ, Bandelow B, Bond A, Davidson JR, et al. Evidence- based guidelines for the pharmacological treatment of anxiety disorders: recommendations from the British Association for Psychopharmacology. J Psychopharmacol. 2005;19(6):567-96. 355. National Institute for Health and Clinical Excellence. Generalised anxiety disorder and panic disorder in adults: management [place unknown]: NICE; 2011 [updated 2019 Jul; cited 2019 Sep 01]. Available from: https://www.nice.org.uk/Guidance/CG113. 356. Sable JA, Jeste DV. Anxiety disorders in older adults. Curr Psychiatry Rep. 2001;3(4):302- 7. 357. Bower ES, Wetherell JL, Mon T, Lenze EJ. Treating anxiety disorders in older adults: current treatments and future directions. Harv Rev Psychiatry. 2015;23(5):329-42. 358. Liao XY, Yamakawa M, Suto S, Shigenobu K, Makimoto K. Changes in activity patterns after the oral administration of brotizolam in institutionalized elderly patients with dementia. Psychogeriatrics. 2012;12(2):99-105. 359. McCarten JR, Kovera C, Maddox MK, Cleary JP. Triazolam in Alzheimer's disease: pilot study on sleep and memory effects. Pharmacol Biochem Behav. 1995;52(2):447-52. 360. Ancill RJ, Carlyle WW, Liang RA, Holliday SG. Agitation in the demented elderly: a role for benzodiazepines? Int Clin Psychopharmacol. 1991;6(3):141-6. 361. Coccaro E, Kramer E, Zemishlany Z, Thorne A, Rice C 3rd, Giordani B, et al. Pharmacologic treatment of noncognitive behavioral disturbances in elderly demented patients. Am J Psychiatry. 1990;147(12):1640-5. 362. Meehan KM, Wang H, David SR, Nisivoccia JR, Jones B, Beasley CM Jr, et al. Comparison of rapidly acting intramuscular olanzapine, lorazepam, and placebo: a double-blind, randomized study in acutely agitated patients with dementia. Neuropsychopharmacology. 2002;26(4):494-504. 363. Covington JS. Alleviating agitation, apprehension, and related symptoms in geriatric patients: a double-blind comparison of a phenothiazine and a benzodiazepine. South Med J. 1975;68(6):719-24.
339 | Page Daniel J. Hoyle 364. Christensen DB, Benfield WR. Alprazolam as an alternative to low-dose haloperidol in older, cognitively impaired nursing facility patients. J Am Geriatr Soc. 1998;46(5):620-5. 365. Butler JM, Begg EJ. Free drug metabolic clearance in elderly people. Clin Pharmacokinet. 2008;47(5):297-321. 366. McLachlan AJ, Pont LG. Drug metabolism in older people--a key consideration in achieving optimal outcomes with medicines. J Gerontol A Biol Sci Med Sci. 2012;67(2):175-80. 367. Peklar J, O'Halloran AM, Maidment ID, Henman MC, Kenny RA, Kos M. Sedative load and frailty among community-dwelling population aged ≥65 years. J Am Med Dir Assoc. 2015;16(4):282-9. 368. Muhlack DC, Hoppe LK, Saum KU, Haefeli WE, Brenner H, Schöttker B. Investigation of a possible association of potentially inappropriate medication for older adults and frailty in a prospective cohort study from Germany. Age Ageing. 2019;49(1):20-5. 369. Maclagan LC, Maxwell CJ, Gandhi S, Guan J, Bell CM, Hogan DB, et al. Frailty and potentially inappropriate medication use at nursing home transition. J Am Geriatr Soc. 2017;65(10):2205-12. 370. Clegg A, Young J, Iliffe S, Rikkert MO, Rockwood K. Frailty in elderly people. Lancet. 2013;381(9868):752-62. 371. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56(3):M146- 56. 372. Robertson DA, Savva GM, Kenny RA. Frailty and cognitive impairment-a review of the evidence and causal mechanisms. Ageing Res Rev. 2013;12(4):840-51. 373. American Geriatrics Society 2015 Beers Criteria Update Expert Panel. American Geriatrics Society 2015 updated Beers Criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2015;63(11):2227-46. 374. Holt S, Schmiedl S, Thürmann PA. Potentially inappropriate medications in the elderly: the PRISCUS list. Dtsch Arztebl Int. 2010;107(31-32):543-51. 375. Renom-Guiteras A, Meyer G, Thürmann PA. The EU(7)-PIM list: a list of potentially inappropriate medications for older people consented by experts from seven European countries. Eur J Clin Pharmacol. 2015;71(7):861-75. 376. Sun GQ, Zhang L, Zhang LN, Wu Z, Hu DF. Benzodiazepines or related drugs and risk of pneumonia: a systematic review and meta-analysis. Int J Geriatr Psychiatry. 2019;34(4):513-21. 377. Massoco C, Palermo-Neto J. Effects of midazolam on equine innate immune response: a flow cytometric study. Vet Immunol Immunopathol. 2003;95(1-2):11-9. 378. Hall AW, Moossa AR, Clark J, Cooley GR, Skinner DB. The effects of premedication drugs on the lower oesophageal high pressure zone and reflux status of rhesus monkeys and man. Gut. 1975;16(5):347-52. 379. Chen TY, Winkelman JW, Mao WC, Liu CL, Hsu CY, Wu CS. The use of benzodiazepine receptor agonists and the risk of hospitalization for pneumonia: a nationwide population-based nested case-control study. Chest. 2018;153(1):161-71. 380. Cheng SY, Chen WY, Liu HC, Yang TW, Pan CH, Yang SY, et al. Benzodiazepines and risk of pneumonia in schizophrenia: a nationwide case-control study. Psychopharmacology (Berl). 2018;235(11):3329-38. 381. Obiora E, Hubbard R, Sanders RD, Myles PR. The impact of benzodiazepines on occurrence of pneumonia and mortality from pneumonia: a nested case-control and survival analysis in a population-based cohort. Thorax. 2013;68(2):163-70. 382. Salat D, Serra-Prat M, Palomera E, Vendrell E, Lavado A, Bartolomé M, et al. Benzodiazepine use and risk of community-acquired pneumonia in a population-based
340 | Page Daniel J. Hoyle cohort study from north-eastern Spain. Clin Med Insights Ther. 2017;9:1179559X16689676. 383. Taipale H, Tolppanen AM, Koponen M, Tanskanen A, Lavikainen P, Sund R, et al. Risk of pneumonia associated with incident benzodiazepine use among community-dwelling adults with Alzheimer disease. CMAJ. 2017;189(14):E519-E29. 384. Almirall J, Bolibar I, Balanzo X, Gonzalez CA. Risk factors for community-acquired pneumonia in adults: a population-based case-control study. Eur Respir J. 1999;13(2):349-55. 385. Dublin S, Walker RL, Jackson ML, Nelson JC, Weiss NS, Von Korff M, et al. Use of opioids or benzodiazepines and risk of pneumonia in older adults: a population-based case- control study. J Am Geriatr Soc. 2011;59(10):1899-907. 386. Burns A, Jacoby R, Luthert P, Levy R. Cause of death in Alzheimer's disease. Age Ageing. 1990;19(5):341-4. 387. Thomas BM, Starr JM, Whalley LJ. Death certification in treated cases of presenile Alzheimer's disease and vascular dementia in Scotland. Age Ageing. 1997;26(5):401-6. 388. Chen JH, Lamberg JL, Chen YC, Kiely DK, Page JH, Person CJ, et al. Occurrence and treatment of suspected pneumonia in long-term care residents dying with advanced dementia. J Am Geriatr Soc. 2006;54(2):290-5. 389. Johnson CF, Frei C, Downes N, McTaggart SA, Akram G. Benzodiazepine and z-hypnotic prescribing for older people in primary care: a cross-sectional population-based study. Br J Gen Pract. 2016;66(647):e410-5. 390. Maust DT, Blow FC, Wiechers IR, Kales HC, Marcus SC. National trends in antidepressant, benzodiazepine, and other sedative-hypnotic treatment of older adults in psychiatric and primary care. J Clin Psychiatry. 2017;78(4):e363-e71. 391. Jackson G, Gerard C, Minko N, Parsotam N. Variation in benzodiazepine and antipsychotic use in people aged 65 years and over in New Zealand. N Z Med J. 2014;127(1396):67-78. 392. Mell T, Jacob L, Fuhr I, Dick S, Rapp MA, Kostev K. Patterns of benzodiazepine prescribing by neuropsychiatrists and general practitioners for elderly patients in Germany in 2014. Int J Clin Pharmacol Ther. 2017;55(6):466-71. 393. Brett J, Maust DT, Bouck Z, Ignacio RV, Mecredy G, Kerr EA, et al. Benzodiazepine use in older adults in the United States, Ontario, and Australia from 2010 to 2016. J Am Geriatr Soc. 2018;66(6):1180-5. 394. Valerie Curran H. Benzodiazepines, memory and mood: a review. Psychopharmacology (Berl). 1991;105(1):1-8. 395. Barker MJ, Greenwood KM, Jackson M, Crowe SF. Cognitive effects of long-term benzodiazepine use: a meta-analysis. CNS drugs. 2004;18(1):37-48. 396. Hanlon JT, Horner RD, Schmader KE, Fillenbaum GG, Lewis IK, Wall WE Jr, et al. Benzodiazepine use and cognitive function among community-dwelling elderly. Clin Pharmacol Ther. 1998;64(6):684-92. 397. Tannenbaum C, Paquette A, Hilmer S, Holroyd-Leduc J, Carnahan R. A systematic review of amnestic and non-amnestic mild cognitive impairment induced by anticholinergic, antihistamine, GABAergic and opioid drugs. Drugs Aging. 2012;29(8):639-58. 398. Gray SL, Dublin S, Yu O, Walker R, Anderson M, Hubbard RA, et al. Benzodiazepine use and risk of incident dementia or cognitive decline: prospective population based study. BMJ. 2016;352:i90. 399. Bourgeois J, Elseviers MM, Van Bortel L, Petrovic M, Vander Stichele RH. The impact of chronic benzodiazepine use on cognitive evolution in nursing home residents. Hum Psychopharmacol. 2015;30(2):85-93.
341 | Page Daniel J. Hoyle 400. Crowe SF, Stranks EK. The residual medium and long-term cognitive effects of benzodiazepine use: an updated meta-analysis. Arch Clin Neuropsychol. 2018;33(7):901-11. 401. Billioti de Gage S, Pariente A, Begaud B. Is there really a link between benzodiazepine use and the risk of dementia? Expert Opin Drug Saf. 2015;14(5):733-47. 402. Billioti de Gage S, Moride Y, Ducruet T, Kurth T, Verdoux H, Tournier M, et al. Benzodiazepine use and risk of Alzheimer’s disease: case-control study. BMJ. 2014;349:g5205. 403. Billioti de Gage S, Begaud B, Bazin F, Verdoux H, Dartigues JF, Peres K, et al. Benzodiazepine use and risk of dementia: prospective population based study. BMJ. 2012;345:e6231. 404. Lucchetta RC, da Mata BP, Mastroianni PC. Association between development of dementia and use of benzodiazepines: a systematic review and meta-analysis. Pharmacotherapy. 2018;38(10):1010-20. 405. Lagnaoui R, Begaud B, Moore N, Chaslerie A, Fourrier A, Letenneur L, et al. Benzodiazepine use and risk of dementia: a nested case-control study. J Clin Epidemiol. 2002;55(3):314-8. 406. Zhong G, Wang Y, Zhang Y, Zhao Y. Association between benzodiazepine use and dementia: a meta-analysis. PLoS One. 2015;10(5):e0127836. 407. Islam MM, Iqbal U, Walther B, Atique S, Dubey NK, Nguyen PA, et al. Benzodiazepine use and risk of dementia in the elderly population: a systematic review and meta- analysis. Neuroepidemiology. 2016;47(3-4):181-91. 408. Richardson K, Wharton SB, Grossi CM, Matthews FE, Fox C, Maidment I, et al. Neuropathological correlates of cumulative benzodiazepine and anticholinergic drug use. J Alzheimers Dis. 2020;74(3):999-1009. 409. Richardson K, Mattishent K, Loke YK, Steel N, Fox C, Grossi CM, et al. History of benzodiazepine prescriptions and risk of dementia: possible bias due to prevalent users and covariate measurement timing in a nested case-control study. Am J Epidemiol. 2019;188(7):1228-36. 410. Grossi CM, Richardson K, Fox C, Maidment I, Steel N, Loke YK, et al. Anticholinergic and benzodiazepine medication use and risk of incident dementia: a UK cohort study. BMC Geriatr. 2019;19(1):276. 411. Busse A, Hensel A, Guhne U, Angermeyer MC, Riedel-Heller SG. Mild cognitive impairment: long-term course of four clinical subtypes. Neurology. 2006;67(12):2176- 85. 412. Petersen RC, Doody R, Kurz A, Mohs RC, Morris JC, Rabins PV, et al. Current concepts in mild cognitive impairment. Arch Neurol. 2001;58(12):1985-92. 413. Hutchinson MA, Smith PF, Darlington CL. The behavioural and neuronal effects of the chronic administration of benzodiazepine anxiolytic and hypnotic drugs. Prog Neurobiol. 1996;49(1):73-97. 414. Shimohama S, Taniguchi T, Fujiwara M, Kameyama M. Changes in benzodiazepine receptors in Alzheimer-type dementia. Ann Neurol. 1988;23(4):404-6. 415. Stern Y. What is cognitive reserve? Theory and research application of the reserve concept. J Int Neuropsychol Soc. 2002;8(3):448-60. 416. Berry SD, Placide SG, Mostofsky E, Zhang Y, Lipsitz LA, Mittleman MA, et al. Antipsychotic and benzodiazepine drug changes affect acute falls risk differently in the nursing home. J Gerontol A Biol Sci Med Sci. 2016;71(2):273-8. 417. Xing D, Ma XL, Ma JX, Wang J, Yang Y, Chen Y. Association between use of benzodiazepines and risk of fractures: a meta-analysis. Osteoporos Int. 2014;25(1):105- 20.
342 | Page Daniel J. Hoyle 418. Panneman MJ, Goettsch WG, Kramarz P, Herings RM. The costs of benzodiazepine- associated hospital-treated fall Injuries in the EU: a Pharmo study. Drugs Aging. 2003;20(11):833-9. 419. Diaz-Gutierrez MJ, Martinez-Cengotitabengoa M, Saez de Adana E, Cano AI, Martinez- Cengotitabengoa MT, Besga A, et al. Relationship between the use of benzodiazepines and falls in older adults: a systematic review. Maturitas. 2017;101:17-22. 420. Machado-Duque ME, Castaño-Montoya JP, Medina-Morales DA, Castro-Rodríguez A, González-Montoya A, Machado-Alba JE. Association between the use of benzodiazepines and opioids with the risk of falls and hip fractures in older adults. Int Psychogeriatr. 2018;30(7):941-6. 421. Aspinall SL, Springer SP, Zhao X, Cunningham FE, Thorpe CT, Semla TP, et al. Central nervous system medication burden and risk of recurrent serious falls and hip fractures in Veterans Affairs nursing home residents. J Am Geriatr Soc. 2019;67(1):74-80. 422. Hanlon JT, Zhao X, Naples JG, Aspinall SL, Perera S, Nace DA, et al. Central nervous system medication burden and serious falls in older nursing home residents. J Am Geriatr Soc. 2017;65(6):1183-9. 423. Echt MA, Samelson EJ, Hannan MT, Dufour AB, Berry SD. Psychotropic drug initiation or increased dosage and the acute risk of falls: a prospective cohort study of nursing home residents. BMC Geriatr. 2013;13(19). 424. Gurvich T, Cunningham JA. Appropriate use of psychotropic drugs in nursing homes. Am Fam Physician. 2000;61(5):1437-46. 425. Cumming RG, Le Couteur DG. Benzodiazepines and risk of hip fractures in older people: a review of the evidence. CNS drugs. 2003;17(11):825-37. 426. Weich S, Pearce HL, Croft P, Singh S, Crome I, Bashford J, et al. Effect of anxiolytic and hypnotic drug prescriptions on mortality hazards: retrospective cohort study. BMJ. 2014;348:g1996. 427. Parsaik AK, Mascarenhas SS, Khosh-Chashm D, Hashmi A, John V, Okusaga O, et al. Mortality associated with anxiolytic and hypnotic drugs--a systematic review and meta- analysis. Aust N Z J Psychiatry. 2016;50(6):520-33. 428. Palmaro A, Dupouy J, Lapeyre-Mestre M. Benzodiazepines and risk of death: results from two large cohort studies in France and UK. Eur Neuropsychopharmacol. 2015;25(10):1566-77. 429. Saarelainen L, Tolppanen AM, Koponen M, Tanskanen A, Tiihonen J, Hartikainen S, et al. Risk of death associated with new benzodiazepine use among persons with Alzheimer disease: a matched cohort study. Int J Geriatr Psychiatry. 2017;33(4):583-90. 430. Kripke DF, Langer RD, Kline LE. Hypnotics' association with mortality or cancer: a matched cohort study. BMJ Open. 2012;2(1):e000850. 431. Shah AJ, Ghasemzadeh N, Zaragoza-Macias E, Patel R, Eapen DJ, Neeland IJ, et al. Sex and age differences in the association of depression with obstructive coronary artery disease and adverse cardiovascular events. J Am Heart Assoc. 2014;3(3):e000741. 432. Patorno E, Glynn RJ, Levin R, Lee MP, Huybrechts KF. Benzodiazepines and risk of all cause mortality in adults: cohort study. BMJ. 2017;358:j2941. 433. Jaussent I, Ancelin M-L, Berr C, Pérès K, Scali J, Besset A, et al. Hypnotics and mortality in an elderly general population: a 12-year prospective study. BMC Med. 2013;11(1):212. 434. Vinkers DJ, Gussekloo J, van der Mast RC, Zitman FG, Westendorp RG. Benzodiazepine use and risk of mortality in individuals aged 85 years or older. JAMA. 2003;290(22):2942-3. 435. Tan KR, Brown M, Labouèbe G, Yvon C, Creton C, Fritschy J-M, et al. Neural bases for addictive properties of benzodiazepines. Nature. 2010;463(7282):769-74.
343 | Page Daniel J. Hoyle 436. Luscher C, Ungless MA. The mechanistic classification of addictive drugs. PLoS Med. 2006;3(11):e437. 437. Saal D, Dong Y, Bonci A, Malenka RC. Drugs of abuse and stress trigger a common synaptic adaptation in dopamine neurons. Neuron. 2003;37(4):577-82. 438. Dell'osso B, Lader M. Do benzodiazepines still deserve a major role in the treatment of psychiatric disorders? A critical reappraisal. Eur Psychiatry. 2013;28(1):7-20. 439. American Psychiatric Association. Substance-related and addictive disorders. In Diagnostic and Statistical Manual of Mental Disorders [Internet]. Arlington, VA: American Psychiatric Publishing; 2013 [cited 2019 Oct 21]. 5th edition:[Available from: https://dsm.psychiatryonline.org/doi/full/10.1176/appi.books.9780890425596.dsm16. 440. Cheng S, Siddiqui TG, Gossop M, Kristoffersen ES, Lundqvist C. Sociodemographic, clinical and pharmacological profiles of medication misuse and dependence in hospitalised older patients in Norway: a prospective cross-sectional study. BMJ open. 2019;9(9):e031483. 441. Graves T, Hanlon JT, Schmader KE, Landsman PB, Samsa GP, Pieper CF, et al. Adverse events after discontinuing medications in elderly outpatients. Arch Intern Med. 1997;157(19):2205-10. 442. Reeve E, Moriarty F, Nahas R, Turner JP, Kouladjian O'Donnell L, Hilmer SN. A narrative review of the safety concerns of deprescribing in older adults and strategies to mitigate potential harms. Expert Opin Drug Saf. 2018;17(1):39-49. 443. Brett J, Murnion B. Management of benzodiazepine misuse and dependence. Aust Prescr. 2015;38(5):152-5. 444. Singh S, Sarkar S. Benzodiazepine abuse among the elderly. J Geriatr Ment Health. 2016;3(2):123-30. 445. Moore N, Pariente A, Bégaud B. Why are benzodiazepines not yet controlled substances? JAMA Psychiatry. 2015;72(2):110-1. 446. Soyka M. Treatment of benzodiazepine dependence. N Engl J Med. 2017;376(12):1147- 57. 447. Madhusoodanan S, Bogunovic OJ. Safety of benzodiazepines in the geriatric population. Expert Opin Drug Saf. 2004;3(5):485-93. 448. Pottie K, Thompson W, Davies S, Grenier J, Sadowski CA, Welch V, et al. Deprescribing benzodiazepine receptor agonists: evidence-based clinical practice guideline. Can Fam Physician. 2018;64(5):339-51. 449. Tenni P, Dunbabin D. A guide to deprescribing benzodiazepines. Primary Health Tasmania; 2016. 450. Schweizer E, Case WG, Rickels K. Benzodiazepine dependence and withdrawal in elderly patients. Am J Psychiatry. 1989;146(4):529-31. 451. Briesacher BA, Tjia J, Field T, Peterson D, Gurwitz JH. Antipsychotic use among nursing home residents. JAMA. 2013;309(5):440-2. 452. Kamble P, Sherer J, Chen H, Aparasu R. Off-label use of second-generation antipsychotic agents among elderly nursing home residents. Psychiatr Serv. 2010;61(2):130-6. 453. Kamble P, Chen H, Sherer J, Aparasu RR. Antipsychotic drug use among elderly nursing home residents in the United States. Am J Geriatr Pharmacother. 2008;6(4):187-97. 454. Allers K, Dorks M, Schmiemann G, Hoffmann F. Antipsychotic drug use in nursing home residents with and without dementia: keep an eye on the pro re nata medication. Int Clin Psychopharmacol. 2017;32(4):213-8. 455. Pitkala KH, Juola AL, Hosia H, Teramura-Gronblad M, Soini H, Savikko N, et al. Eight-year trends in the use of opioids, other analgesics, and psychotropic medications among institutionalized older people in Finland. J Am Med Dir Assoc. 2015;16(11):973-8. 456. Rochon PA, Stukel TA, Bronskill SE, Gomes T, Sykora K, Wodchis WP, et al. Variation in nursing home antipsychotic prescribing rates. Arch Intern Med. 2007;167(7):676-83.
344 | Page Daniel J. Hoyle 457. van der Putten MJ, Wetzels RB, Bor H, Zuidema SU, Koopmans RT. Antipsychotic drug prescription rates among Dutch nursing homes: the influence of patient characteristics and the dementia special care unit. Aging Ment Health. 2014;18(7):828-32. 458. Nijk RM, Zuidema SU, Koopmans RT. Prevalence and correlates of psychotropic drug use in Dutch nursing-home patients with dementia. Int Psychogeriatr. 2009;21(3):485- 93. 459. Ruths S, Sorensen PH, Kirkevold O, Husebo BS, Kruger K, Halvorsen KH, et al. Trends in psychotropic drug prescribing in Norwegian nursing homes from 1997 to 2009: a comparison of six cohorts. Int J Geriatr Psychiatry. 2013;28(8):868-76. 460. Azermai M, Elseviers M, Petrovic M, Van Bortel L, Stichele RV. Geriatric drug utilisation of psychotropics in Belgian nursing homes. Hum Psychopharmacol Clin Exp. 2011;26(1):12-20. 461. Lovheim H, Sandman PO, Karlsson S, Gustafson Y. Changes between 1982 and 2000 in the prevalence of behavioral symptoms and psychotropic drug treatment among old people with cognitive impairment in geriatric care. Int Psychogeriatr. 2009;21(5):941-8. 462. Mann E, Haastert B, Bohmdorfer B, Fruhwald T, Iglseder B, Roller-Wirnsberger R, et al. Prevalence and associations of potentially inappropriate prescriptions in Austrian nursing home residents: secondary analysis of a cross-sectional study. Wien Klin Wochenschr. 2013;125(7-8):180-8. 463. Mamun K, Goh-Tan CY, Ng LL. Prescribing psychoactive medications in nursing homes: current practice in Singapore. Singapore Med J. 2003;44(12):625-9. 464. Westaway K, Sluggett J, Alderman C, Moffat A, Procter N, Roughead E. The extent of antipsychotic use in Australian residential aged care facilities and interventions shown to be effective in reducing antipsychotic use: a literature review. Dementia (London). 2018:1471301218795792. 465. Lakey SL, Gray SL, Sales AE, Sullivan J, Hedrick SC. Psychotropic use in community residential care facilities: a prospective cohort study. Am J Geriatr Pharmacother. 2006;4(3):227-35. 466. Gustafsson M, Karlsson S, Lovheim H. Inappropriate long-term use of antipsychotic drugs is common among people with dementia living in specialized care units. BMC Pharmacol Toxicol. 2013;14(1):10. 467. Wetzels RB, Zuidema SU, de Jonghe JF, Verhey FR, Koopmans RT. Prescribing pattern of psychotropic drugs in nursing home residents with dementia. Int Psychogeriatr. 2011;23(8):1249-59. 468. Lustenberger I, Schupbach B, von Gunten A, Mosimann U. Psychotropic medication use in Swiss nursing homes. Swiss Med Wkly. 2011;141:w13254. 469. Kalisch Ellett LM, Kassie GM, Pratt NL, Kerr M, Roughead EE. Prevalence and duration of use of medicines recommended for short-term use in aged care facility residents. Pharmacy (Basel). 2019;7(2):E55. 470. Shah SM, Carey IM, Harris T, Dewilde S, Cook DG. Antipsychotic prescribing to older people living in care homes and the community in England and Wales. Int J Geriatr Psychiatry. 2011;26(4):423-34. 471. Hasan SS, Zaidi STR, Nirwan JS, Ghori MU, Javid F, Ahmadi K, et al. Use of central nervous system (CNS) medicines in aged care homes: a systematic review and meta-analysis. J Clin Med. 2019;8(9):E1292. 472. Snowdon J. A follow-up survey of psychotropic drug use in Sydney nursing homes. Med J Aust. 1999;170(7):299-301. 473. Ray WA, Blazer DG, Schaffner W, Federspiel CF. Reducing antipsychotic drug prescribing for nursing home patients: a controlled trial of the effect of an educational visit. Am J Public Health. 1987;77(11):1448-50.
345 | Page Daniel J. Hoyle 474. Ray WA, Federspiel CF, Schaffner W. A study of antipsychotic drug use in nursing homes: epidemiologic evidence suggesting misuse. Am J Public Health. 1980;70(5):485-91. 475. Janus SI, van Manen JG, MJ IJ, Zuidema SU. Psychotropic drug prescriptions in Western European nursing homes. Int Psychogeriatr. 2016;28(11):1775-90. 476. Kleijer BC, van Marum RJ, Frijters DH, Jansen PA, Ribbe MW, Egberts AC, et al. Variability between nursing homes in prevalence of antipsychotic use in patients with dementia. Int Psychogeriatr. 2014;26(3):363-71. 477. Richter T, Mann E, Meyer G, Haastert B, Kopke S. Prevalence of psychotropic medication use among German and Austrian nursing home residents: a comparison of 3 cohorts. J Am Med Dir Assoc. 2012;13(2):187 e7- e13. 478. Ågotnes G, McGregor MJ, Lexchin J, Doupe MB, Müller B, Harrington C. An international mapping of medical care in nursing homes. Health Services Insights. 2019;12:1178632918825083. 479. New South Wales Ministerial Taskforce on Psychotropic Medication Use in Nursing Homes, Raphael B. Psychotropic medication use in nursing homes: report. Sydney (AU): Department of Health, Public Health Division; 1998. 480. Therapeutic Goods Administration. Medicines safety update. Canberra (AU): Australian Government; 2015. Report No.: 4. 481. Lonergan E, Luxenberg J. Valproate preparations for agitation in dementia. Cochrane Database Syst Rev. 2009(3):CD003945. 482. van der Spek K, Gerritsen DL, Smalbrugge M, Nelissen-Vrancken MH, Wetzels RB, Smeets CH, et al. A reliable and valid index was developed to measure appropriate psychotropic drug use in dementia. J Clin Epidemiol. 2015;68(8):903-12. 483. Australian Institute of Health and Welfare. Older Australia at a glance [Internet]. Canberra (AU): Australian Institute of Health and Welfare; 2018 [cited 2019 Dec 27]. Available from: https://www.aihw.gov.au/reports/older-people/older-australia-at-a- glance/contents/summary. 484. University of South Australia. Veterans' MATES Therapeutic Brief 12: Antipsychotics in dementia [Internet]. Adelaide (AU): Department of Veterans' Affairs; 2007 [cited 2018 May 22]. Available from: https://www.veteransmates.net.au/topic-12-therapeutic- brief. 485. University of South Australia. Veterans' MATES [Internet]. Adelaide (AU): Department of Veterans' Affairs; 2018 [cited 2020 Jan 13]. Available from: https://www.veteransmates.net.au/. 486. Walsh KA, Dennehy R, Sinnott C, Browne J, Byrne S, McSharry J, et al. Influences on decision-making regarding antipsychotic prescribing in nursing home residents with dementia: a systematic review and synthesis of qualitative evidence. J Am Med Dir Assoc. 2017;18(10):897. 487. Janus SI, van Manen JG, Zuidema SU, Snijder C, Drossaert CH, Ijzerman MJ. Reasons for (not) discontinuing antipsychotics in dementia. Psychogeriatrics. 2017;18(1):13-20. 488. Anderson K, Stowasser D, Freeman C, Scott I. Prescriber barriers and enablers to minimising potentially inappropriate medications in adults: a systematic review and thematic synthesis. BMJ Open. 2014;4(12):e006544. 489. Huybrechts KF, Rothman KJ, Brookhart MA, Silliman RA, Crystal S, Gerhard T, et al. Variation in antipsychotic treatment choice across US nursing homes. J Clin Psychopharmacol. 2012;32(1):11-7. 490. Hughes CM, Lapane K, Watson MC, Davies HT. Does organisational culture influence prescribing in care homes for older people? A new direction for research. Drugs Aging. 2007;24(2):81-93.
346 | Page Daniel J. Hoyle 491. Smeets CH, Smalbrugge M, Zuidema SU, Derksen E, de Vries E, van der Spek K, et al. Factors related to psychotropic drug prescription for neuropsychiatric symptoms in nursing home residents with dementia. J Am Med Dir Assoc. 2014;15(11):835-40. 492. Ballard C, Thomas A, Fossey J, Lee L, Jacoby R, Lana MM, et al. A 3-month, randomized, placebo-controlled, neuroleptic discontinuation study in 100 people with dementia: the Neuropsychiatric Inventory median cutoff is a predictor of clinical outcome. J Clin Psychiatry. 2004;65(1):114-9. 493. Devanand DP, Pelton GH, Cunqueiro K, Sackeim HA, Marder K. A 6-month, randomized, double-blind, placebo-controlled pilot discontinuation trial following response to haloperidol treatment of psychosis and agitation in Alzheimer's disease. Int J Geriatr Psychiatry. 2011;26(9):937-43. 494. Bourgeois J, Elseviers MM, Van Bortel L, Petrovic M, Vander Stichele RH. Feasibility of discontinuing chronic benzodiazepine use in nursing home residents: a pilot study. Eur J Clin Pharmacol. 2014;70(10):1251-60. 495. Campbell N, Maidment ID, Randle E, Shaw RL. Preparing care home staff to manage challenging behaviours among residents living with dementia: A mixed-methods evaluation. Health Psychology Open. 2020;7(2):2055102920933065. 496. Maidment ID, Shaw RL, Killick K, Damery S, Hilton A, Wilcock J, et al. Improving the management of behaviour that challenges associated with dementia in care homes: protocol for pharmacy–health psychology intervention feasibility study. BMJ Open. 2016;6(3):e010279. 497. Westbury J. Roles for pharmacists in improving the quality use of psychotropic medicines in residential aged care facilities [dissertation]. Hobart (AU): University of Tasmania; 2011. 418 p. 498. Monette J, Monette M, Sourial N, Vandal AC, Wolfson C, Champoux N, et al. Effect of an interdisciplinary educational program on antipsychotic prescribing among residents with dementia in two long-term care centers. J Appl Gerontol. 2013;32(7):833-54. 499. Monette J, Champoux N, Monette M, Fournier L, Wolfson C, du Fort GG, et al. Effect of an interdisciplinary educational program on antipsychotic prescribing among nursing home residents with dementia. Int J Geriatr Psychiatry. 2008;23(6):574-9. 500. Harrison SL, Bradley C, Milte R, Liu E, Kouladjian O'Donnell L, Hilmer SN, et al. Psychotropic medications in older people in residential care facilities and associations with quality of life: a cross-sectional study. BMC Geriatr. 2018;18(1):60. 501. Forsetlund L, Eike M, Gjerberg E, Vist G. Effect of interventions to reduce potentially inappropriate use of drugs in nursing homes: a systematic review of randomised controlled trials. BMC Geriatr. 2011;11(1):16. 502. Loganathan M, Singh S, Franklin BD, Bottle A, Majeed A. Interventions to optimise prescribing in care homes: systematic review. Age Ageing. 2011;40(2):150-62. 503. Marcum ZA, Handler SM, Wright R, Hanlon JT. Interventions to improve suboptimal prescribing in nursing homes: a narrative review. Am J Geriatr Pharmacother. 2010;8(3):183-200. 504. Richter T, Meyer G, Mohler R, Kopke S. Psychosocial interventions for reducing antipsychotic medication in care home residents. Cochrane Database Syst Rev. 2012;12:CD008634. 505. Gurwitz JH, Bonner A, Berwick DM. Reducing excessive use of antipsychotic agents in nursing homes. JAMA. 2017;318(2):118-9. 506. Millar AN, Daffu-O'Reilly A, Hughes CM, Alldred DP, Barton G, Bond CM, et al. Development of a core outcome set for effectiveness trials aimed at optimising prescribing in older adults in care homes. Trials. 2017;18(1):175.
347 | Page Daniel J. Hoyle 507. Sheehan R, Strydom A, Brown E, Marston L, Hassiotis A. Association of focused medication review with optimization of psychotropic drug prescribing: a systematic review and meta-analysis. JAMA Netw Open. 2018;1(6):e183750. 508. Alldred DP, Kennedy MC, Hughes C, Chen TF, Miller P. Interventions to optimise prescribing for older people in care homes. Cochrane Database Syst Rev. 2016;2:CD009095. 509. Reeve E, Ong M, Wu A, Jansen J, Petrovic M, Gnjidic D. A systematic review of interventions to deprescribe benzodiazepines and other hypnotics among older people. Eur J Clin Pharmacol. 2017;73(8):927-35. 510. Cochrane Effective Practice and Organisation of Care Group. Data collection checklist [Internet]. Ottawa, ON 2002 [cited 2016 Dec 16]. Available from: http://methods.cochrane.org/sites/methods.cochrane.org.bias/files/public/uploads/E POC%20Data%20Collection%20Checklist.pdf. 511. Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group. Preferred Reporting Items for Systematic reviews and Meta-Analyses: the PRISMA statement. PLoS Med. 2009;6(7):e1000097. 512. Effective Public Health Practice Project. Quality assessment tool for quantitative studies. [Internet]. Hamilton, ON 2009 [cited 2017 Jul 4]. Available from: http://www.ephpp.ca/tools.html. 513. Tjia J, Hunnicutt JN, Herndon L, Blanks CR, Lapane KL, Wehry S. Association of a communication training program with use of antipsychotics in nursing homes. JAMA Intern Med. 2017;177(6):846-53. 514. Ballard C, Orrell M, Yong Zhong S, Moniz-Cook E, Stafford J, Whittaker R, et al. Impact of antipsychotic review and nonpharmacological intervention on antipsychotic use, neuropsychiatric symptoms, and mortality in people with dementia living in nursing homes: a factorial cluster-randomized controlled trial by the well-being and health for people with dementia (WHELD) program. Am J Psychiatry. 2016;173(3):252-62. 515. Ballard C, Orrell M, Sun Y, Moniz-Cook E, Stafford J, Whitaker R, et al. Impact of antipsychotic review and non-pharmacological intervention on health-related quality of life in people with dementia living in care homes: WHELD-a factorial cluster randomised controlled trial. Int J Geriatr Psychiatry. 2016;173(3):252-62. 516. Rajkumar AP, Ballard C, Fossey J, Corbett A, Woods B, Orrell M, et al. Apathy and its response to antipsychotic review and nonpharmacological interventions in people with dementia living in nursing homes: WHELD, a factorial cluster randomized controlled trial. J Am Med Dir Assoc. 2016;17(8):741-7. 517. Fossey J, Ballard C, Juszczak E, James I, Alder N, Jacoby R, et al. Effect of enhanced psychosocial care on antipsychotic use in nursing home residents with severe dementia: cluster randomised trial. BMJ. 2006;332(7544):756-61. 518. Gilbert A, Owen N, Innes JM, Sansom L. Trial of an intervention to reduce chronic benzodiazepine use among residents of aged-care accommodation. Aust N Z J Med. 1993;23(4):343-7. 519. Avorn J, Soumerai S, Everitt D, Ross-Degnan D, Beers M, Sherman D, et al. A randomized trial of a program to reduce the use of psychoactive drugs in nursing homes. N Engl J Med. 1992;327(3):168-73. 520. Ballard C, Powell I, James I, Reichelt K, Myint P, Potkins D, et al. Can psychiatric liaison reduce neuroleptic use and reduce health service utilization for dementia patients residing in care facilities. Int J Geriatr Psychiatry. 2002;17(2):140-5. 521. Ray WA, Taylor JA, Meador KG, Lichtenstein MJ, Griffin MR, Fought R, et al. Reducing antipsychotic drug use in nursing homes. A controlled trial of provider education. Arch Intern Med. 1993;153(6):713-21.
348 | Page Daniel J. Hoyle 522. Vida S, Monette J, Wilchesky M, Monette M, Friedman R, Nguyen A, et al. A long-term care center interdisciplinary education program for antipsychotic use in dementia: program update five years later. Int Psychogeriatr. 2012;24(4):599-605. 523. Earthy A, Collins J, Wong S, Myckatyn M. Ensuring appropriate use of neuroleptics. Canadian Nursing Home. 2000;11(5):5-10. 524. Thapa PB, Meador KG, Gideon P, Fought RL, Ray WA. Effects of antipsychotic withdrawal in elderly nursing home residents. J Am Geriatr Soc. 1994;42(3):280-6. 525. Cummings J. Neuropsychiatric Inventory-Nursing Home Version (NPI-NH) [Internet]. Los Angeles, CA 2009 [cited 2017 Jul 31]. Available from: http://www.dementiamanagementstrategy.com/File.axd?id=c9f7405f-3596-4f5f- 9f8d-f322e5188678. 526. Simmons SF, Bonnett KR, Hollingsworth E, Kim J, Powers J, Habermann R, et al. Reducing antipsychotic medication use in nursing homes: a qualitative study of nursing staff perceptions. Gerontologist. 2017;58(4):e239-e50. 527. Zwijsen SA, Bosmans JE, Gerritsen DL, Pot AM, Hertogh CM, Smalbrugge M. The cost- effectiveness of grip on challenging behaviour: an economic evaluation of a care programme for managing challenging behaviour. Int J Geriatr Psychiatry. 2016;31(6):567-74. 528. Olivieri-Mui BL, Devlin JW, Ochoa A, Schenck D, Briesacher B. Perceptions vs. evidence: therapeutic substitutes for antipsychotics in patients with dementia in long-term care. Aging Ment Health. 2017;22(4):544-9. 529. NHS Centre for Reviews and Dissemination. Systematic review: CRD's guidance for undertaking reviews in health care. York (GB): University of York; 2009. 530. Brodaty H, Aerts L, Harrison F, Jessop T, Cations M, Chenoweth L, et al. Antipsychotic deprescription for older adults in long-term care: the HALT study. J Am Med Dir Assoc. 2018;19(7):592-600.e7. 531. Ballard C, Corbett A, Orrell M, Williams G, Moniz-Cook E, Romeo R, et al. Impact of person-centred care training and person-centred activities on quality of life, agitation, and antipsychotic use in people with dementia living in nursing homes: a cluster- randomised controlled trial. PLoS Med. 2018;15(2):e1002500. 532. Romeo R, Zala D, Knapp M, Orrell M, Fossey J, Ballard C. Improving the quality of life of care home residents with dementia: cost-effectiveness of an optimized intervention for residents with clinically significant agitation in dementia. Alzheimers Dement. 2018;15(2):282-91. 533. Richter C, Berg A, Langner H, Meyer G, Kopke S, Balzer K, et al. Effect of person-centred care on antipsychotic drug use in nursing homes (EPCentCare): a cluster-randomised controlled trial. Age Ageing. 2019;48(3):419-25. 534. Weeks WB, Mishra MK, Curto D, Petersen CL, Cano P, Hswen Y, et al. Comparing three methods for reducing psychotropic use in older demented Spanish care home residents. J Am Geriatr Soc. 2019;67(7):1444-53. 535. Cossette B, Bruneau M-A, Couturier Y, Gilbert S, Boyer D, Ricard J, et al. Optimizing practices, Use, Care and Services–Antipsychotics (OPUS-AP) in long-term care centers in Québec, Canada: a strategy for best practices. J Am Med Dir Assoc. 2019;21(2):212- 9. 536. Ling T, Gee P, Westbury J, Bindoff I, Peterson G. An internet-based method for extracting nursing home resident sedative medication data from pharmacy packing systems: descriptive evaluation. J Med Internet Res. 2017;19(8):e283. 537. Lindsey PL. Psychotropic medication use among older adults: what all nurses need to know. J Gerontol Nurs. 2009;35(9):28-38.
349 | Page Daniel J. Hoyle 538. Brown DT, Westbury JL. Assessing health practitioner knowledge of appropriate psychotropic medication use in nursing homes: validation of the Older Age Psychotropic quiz. J Gerontol Nurs. 2016;42(9):21-7. 539. Australian Institute of Health and Welfare. Use of aged care services before death. Canberra (AU): Australian Institute of Health and Welfare; 2015. 540. Pruchno RA, Brill JE, Shands Y, Gordon JR, Genderson MW, Rose M, et al. Convenience samples and caregiving research: how generalizable are the findings? Gerontologist. 2008;48(6):820-7. 541. Hawthorne G, Richardson J, Osborne R. The Assessment of Quality of Life (AQoL) instrument: a psychometric measure of health-related quality of life. Qual Life Res. 1999;8(3):209-24. 542. Woods SW. Chlorpromazine equivalent doses for the newer atypical antipsychotics. J Clin Psychiatry. 2003;64(6):663-7. 543. WHO Collaboration Centre for Drug Statistics Methodology. Guidelines for ATC classification and DDD assignment 2014. 17th ed. Oslo (NO): Norwegian Institute of Public Health; 2013. 544. Taylor D, Paton C, Kapur S, editors. The Maudsley Prescribing Guidelines in Psychiatry. 11th ed. Hoboken, NJ: Wiley-Blackwell; 2012. 545. Zuidema SU, Buursema AL, Gerritsen MG, Oosterwal KC, Smits MM, Koopmans RT, et al. Assessing neuropsychiatric symptoms in nursing home patients with dementia: reliability and Reliable Change Index of the Neuropsychiatric Inventory and the Cohen- Mansfield Agitation Inventory. Int J Geriatr Psychiatry. 2011;26(2):127-34. 546. Ballard C, Hanney ML, Theodoulou M, Douglas S, McShane R, Kossakowski K, et al. The dementia antipsychotic withdrawal trial (DART-AD): long-term follow-up of a randomised placebo-controlled trial. Lancet Neurol. 2009;8(2):151-7. 547. Hawthorne G, Osborne R. Population norms and meaningful differences for the Assessment of Quality of Life (AQoL) measure. Aust N Z J Public Health. 2005;29(2):136- 42. 548. Giles LC, Hawthorne G, Crotty M. Health-related quality of life among hospitalized older people awaiting residential aged care. Health Qual Life Outcomes. 2009;7:71-. 549. The National Health and Medical Research Council, The Australian Research Council, The Australian Vice-Chancellors' Committee. National Statement on ethical conduct in human research 2007 (Updated May 2015). Canberra (AU): Commonwealth of Australia; 2015. 550. The ATC/DDD Methodology [Internet]. [place unknown]: World Health Organization; 2020 [cited 2020 Jan 23]. Available from: https://www.who.int/medicines/regulation/medicines- safety/toolkit_methodology/en/. 551. International Classification of Diseases (ICD) information sheet [Internet]. [place unknown]: World Health Organization; 2020 [cited 2020 Jan 23]. Available from: https://www.who.int/classifications/icd/factsheet/en/. 552. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373-83. 553. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi J-C, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43(11):1130-9. 554. Quan H, Li B, Couris CM, Fushimi K, Graham P, Hider P, et al. Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol. 2011;173(6):676-82.
350 | Page Daniel J. Hoyle 555. Wood S, Cummings JL, Hsu MA, Barclay T, Wheatley MV, Yarema KT, et al. The use of the neuropsychiatric inventory in nursing home residents: characterization and measurement. Am J Geriatr Psychiatry. 2000;8(1):75-83. 556. Selbaek G, Kirkevold O, Sommer OH, Engedal K. The reliability and validity of the Norwegian version of the Neuropsychiatric Inventory, Nursing Home version (NPI-NH). Int Psychogeriatr. 2008;20(2):375-82. 557. Cummings JL. The Neuropsychiatric Inventory assessing psychopathology in dementia patients. Neurology. 1997;48(5 Suppl 6):10S-6S. 558. Lange RT, Hopp GA, Kang N. Psychometric properties and factor structure of the Neuropsychiatric Inventory Nursing Home version in an elderly neuropsychiatric population. Int J Geriatr Psychiatry. 2004;19(5):440-8. 559. Kaufer DI, Cummings JL, Christine D, Bray T, Castellon S, Masterman D, et al. Assessing the impact of neuropsychiatric symptoms in Alzheimer's disease: the Neuropsychiatric Inventory Caregiver Distress Scale. J Am Geriatr Soc. 1998;46(2):210-5. 560. Iverson GL, Hopp GA, DeWolfe K, Solomons K. Measuring change in psychiatric symptoms using the Neuropsychiatric Inventory: Nursing Home version. Int J Geriatr Psychiatry. 2002;17(5):438-43. 561. Cohen-Mansfield J, Marx MS, Rosenthal AS. A description of agitation in a nursing home. J Gerontol. 1989;44(3):M77-M84. 562. Rabinowitz J, Davidson M, De Deyn PP, Katz I, et al. Factor analysis of the Cohen- Mansfield Agitation Inventory in three large samples of nursing home patients with dementia and behavioral disturbance. Am J Geriatr Psychiatry. 2005;13(11):991-8. 563. Cohen-Mansfield J. Instruction manual for the Cohen-Mansfield Agitation Inventory (CMAI) [Internet]. [place unknown]1991 [cited 2017 Jul 31]. Available from: https://www.pdx.edu/ioa/sites/www.pdx.edu.ioa/files/CMAI_Scale%20(1).pdf. 564. Zuidema SU, de Jonghe JF, Verhey FR, Koopmans RT. Agitation in Dutch institutionalized patients with dementia: factor analysis of the Dutch Version of the Cohen-Mansfield Agitation Inventory. Dement Geriatr Cogn Disord. 2007;23(1):35-41. 565. Hawthorne G, Richardson J, Day NA. A comparison of the Assessment of Quality of Life (AQoL) with four other generic utility instruments. Ann Med. 2001;33(5):358-70. 566. Hawthorne G, Richardson J. Measuring the value of program outcomes: a review of multiattribute utility measures. Expert Rev Pharmacoecon Outcomes Res. 2001;1(2):215-28. 567. Richardson J. Psychometric validity and multi attribute utility (MAU) instruments. Melbourne (AU): Monash University, Centre for Health Economics; 2010. 568. Hawthorne G, Richardson J, Osborne R, McNeil H. The Assessment of Quality of Life (AQoL) instrument. Melbourne (AU): Centre for Health Program Evaluation; 1997. 569. Richardson J, Chen G, Iezzi A, Khan MA. Transformations between the Assessment of Quality of Life AQoL instruments and test-retest reliability. Melbourne (AU): Centre for Health Economics, Monash University; 2011. 570. Hawthorne G, Richardson J, Day N, Osborne R, McNeil H. Construction and utility scaling of the Assessment of Quality of Life (AQoL) instrument. Melbourne (AU): Centre for Health Program Evaluation; 2000. 571. Hawthorne G, Richardson J, Day N. Using the Assessment of Quality of Life (AQoL) Version 1. Melbourne (AU): Centre for Health Program Evaluation; 2009. 572. AQoL. Scoring- Psychometric (unweighted) or Utility (weighted)? [Internet]. Clayton (AU): Centre for Health Economics; 2014 [cited 2018 Jun 21]. Available from: http://www.aqol.com.au/index.php/scoring-algorithms. 573. Osborne RH, Hawthorne G, Lew EA, Gray LC. Quality of life assessment in the community-dwelling elderly: validation of the Assessment of Quality of Life (AQoL) instrument and comparison with the SF-36. J Clin Epidemiol. 2003;56(2):138-47.
351 | Page Daniel J. Hoyle 574. Holland R, Smith RD, Harvey I, Swift L, Lenaghan E. Assessing quality of life in the elderly: a direct comparison of the EQ-5D and AQoL. Health Econ. 2004;13(8):793-805. 575. Harris A, Gospodarevskaya E, Callaghan J, Story I. The cost effectiveness of a pharmacist reviewing medication among the elderly in the community. Australas J Ageing. 2001;20(4):179-86. 576. Wang L, Crawford JD, Reppermund S, Trollor J, Campbell L, Baune BT, et al. Body mass index and waist circumference predict health-related quality of life, but not satisfaction with life, in the elderly. Qual Life Res. 2018;27(10):2653-65. 577. Bonner A, Chambers S, Healy H, Hoy W, Mitchell G, Kark A, et al. Tracking patients with advanced kidney disease in the last 12 months of life. J Ren Care. 2018;44(2):115-22. 578. Nikmat AW, Hawthorne G, Al-Mashoor SH. The comparison of quality of life among people with mild dementia in nursing home and home care--a preliminary report. Dementia (London). 2015;14(1):114-25. 579. IBM Corporation. IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corporation; 2016. 580. Slee J, Agius S, Keller J, Gilchrist J, Gray L, Hawthorne G. North Eastern Coordinated Care Trial. Final evaluation report. 2000. 581. Kennedy M, Osborne R. Care coordination in emergency medicine. Evaluation report. 2002. 582. Segal L, Day N, Day S, Dunt D, Piterman H, Robertson I. Evaluation of the Southern Health Care Network Coordinated Care Trial. Final report. Volume I: executive summary. 2000. 583. Helmes E, Csapo KG, Short JA. Standardization and validation of the Multidimensional Observation Scale for Elderly Subjects (MOSES). J Gerontol. 1987;42(4):395-405. 584. Zimmerman S, Sloane PD, Eckert JK, Gruber-Baldini AL, Morgan LA, Hebel JR, et al. How good is assisted living? Findings and implications from an outcomes study. J Gerontol B Psychol Sci Soc Sci. 2005;60(4):S195-204. 585. Pruchno RA, Kleban MH, Resch NL. Psychometric assessment of the Multidimensional Observation Scale for Elderly Subjects (MOSES). J Gerontol. 1988;43(6):P164-9. 586. Woods B, O'Philbin L, Farrell EM, Spector AE, Orrell M. Reminiscence therapy for dementia. Cochrane Database Syst Rev. 2018(3):CD001120. 587. Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist. 1969;9(3_Part_1):179-86. 588. Hokoishi K, Ikeda M, Maki N, Nomura M, Torikawa S, Fujimoto N, et al. Interrater reliability of the Physical Self-Maintenance Scale and the Instrumental Activities of Daily Living Scale in a variety of health professional representatives. Aging Ment Health. 2001;5(1):38-40. 589. Helvik A-S, Høgseth LD, Bergh S, Šaltytė-Benth J, Kirkevold Ø, Selbæk G. A 36-month follow-up of decline in activities of daily living in individuals receiving domiciliary care. BMC Geriatr. 2015;15(1):47. 590. ePROVIDE. Physical Self-Maintenance Scale (PSMS) [Internet]. [place unknown]: MAPI Research Trust; 2016 [cited 2018 May 01]. Available from: https://eprovide.mapi- trust.org/instruments/physical-self-maintenance-scale. 591. Stafford A. A clinical and economic evaluation of medication reviews conducted by pharmacists for communty-dwelling Australians [dissertation]. Hobart (AU): University of Tasmania; 2012. 725 p. 592. Chou SC, Boldy DP, Lee AH. Measuring job satisfaction in residential aged care. Int J Qual Health Care. 2002;14(1):49-54. 593. Aasmul I, Husebo BS, Flo E. Staff distress improves by treating pain in nursing home patients with dementia: results from a cluster-randomized controlled trial. J Pain Symptom Manage. 2016;52(6):795-805.
352 | Page Daniel J. Hoyle 594. R Core Team. R: A language and environment for statistical computing. Vienna (AT): R Foundation for Statistical Computing; 2017. 595. Australian Institute of Health and Welfare. Residential aged care and home care 2014- 15: supporting data [Data file]. Canberra (AU): GEN Aged Care Data; 2015 [cited 2018 Nov 15]. Available from: https://www.gen-agedcaredata.gov.au/Resources/Access- data/2015/December/Residential-aged-care-and-Home-Care-2014%E2%80%9315- suppo. 596. Lavrakas P. Encyclopedia of survey research methods. Thousand Oaks, CA: Sage Publications; 2008. 597. Swinscow T, Campbell M. Statistics at Square One 11. Correlation and regression 9th ed. London (GB): BMJ Publishing Group; 1997. 598. Elfil M, Negida A. Sampling methods in clinical research; an educational review. Emerg (Tehran). 2017;5(1):e52-e. 599. Lacey RJ, Wilkie R, Wynne-Jones G, Jordan JL, Wersocki E, McBeth J. Evidence for strategies that improve recruitment and retention of adults aged 65 years and over in randomised trials and observational studies: a systematic review. Age Ageing. 2017;46(6):895-903. 600. Lam HR, Chow S, Taylor K, Chow R, Lam H, Bonin K, et al. Challenges of conducting research in long-term care facilities: a systematic review. BMC Geriatr. 2018;18(1):242. 601. Maidment ID, Damery S, Campbell N, Seare N, Fox C, Iliffe S, et al. Medication review plus person-centred care: a feasibility study of a pharmacy-health psychology dual intervention to improve care for people living with dementia. BMC Psychiatry. 2018;18(1):340. 602. Whelan PJ, Walwyn R, Gaughran F, Macdonald A. Impact of the demand for 'proxy assent' on recruitment to a randomised controlled trial of vaccination testing in care homes. J Med Ethics. 2013;39(1):36-40. 603. Lasseter G, Donald I, Hapeshi J, McNulty CA. Is the Mental Capacity Act reducing generalizable research in care homes? Public Health. 2011;125(9):604-8. 604. Wang MQ, Youssef T, Smerdely P. Incidence and outcomes of humeral fractures in the older person. Osteoporos Int. 2018;29(7):1601-8. 605. Nishtala PS, McLachlan AJ, Bell JS, Chen TF. A retrospective study of drug-related problems in Australian aged care homes: medication reviews involving pharmacists and general practitioners. J Eval Clin Pract. 2011;17(1):97-103. 606. Davies HTO, Nutley SM, Mannion R. Organisational culture and quality of health care. Qual Health Care. 2000;9(2):111. 607. Meador KG, Taylor JA, Thapa PB, Fought RL, Ray WA. Predictors of antipsychotic withdrawal or dose reduction in a randomized controlled trial of provider education. J Am Geriatr Soc. 1997;45(2):207-10. 608. Draper B, Brodaty H, Low L-F, Saab D, Lie D, Richards V, et al. Use of psychotropics in Sydney nursing homes: associations with depression, psychosis, and behavioral disturbances. Int Psychogeriatr. 2001;13(01):107-20. 609. Hajduk G. Introduction to linear mixed models [Internet]. [place unknown]: Coding Club; 2017 [cited 2018 Oct 18]. Available from: https://ourcodingclub.github.io/2017/03/15/mixed-models.html#one. 610. Pallant J, editor. SPSS survival manual. 5th ed. Crows Nest (AU): Allen & Unwin; 2013. 611. Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Softw. 2015;67(1):1-48. 612. Stan Development Team. The Stan Core Library, Version 2.18. [Internet]. [place unknown] 2018 [cited 2018 Aug 30]. Available from: http://mc-stan.org. 613. Bürkner P-C. brms: an R package for Bayesian multilevel models using Stan. J Stat Softw. 2017;80(1):28.
353 | Page Daniel J. Hoyle 614. Vehtari A, Mononen T, Tolvanen V, Sivula T, Winther O. Bayesian leave-one-out cross- validation approximations for Gaussian latent variable models. J Mach Learn Res. 2016;17(1):3581-618. 615. Gelman A, Carlin JB, Stern HS, Dunson DB, Vehtari A, Rubin DB. Bayesian data analysis. 3rd ed. Boca Raton, FL: Chapman and Hall/CRC; 2013. 616. Lüdecke D. sjstats: statistical functions for regression models [Internet]. Geneva (CH): Zenodo; 2018 [cited 2018 30/08]. Available from: http://doi.org/10.5281/zenodo.1400701. 617. Bonner AF, Field TS, Lemay CA, Mazor KM, Andersen DA, Compher CJ, et al. Rationales that providers and family members cited for the use of antipsychotic medications in nursing home residents with dementia. J Am Geriatr Soc. 2015;63(2):302-8. 618. Cohen-Mansfield J. Agitated behavior in persons with dementia: The relationship between type of behavior, its frequency, and its disruptiveness. J Psychiatr Res. 2008;43(1):64-9. 619. Bozat-Emre S, Doupe M, Kozyrskyj AL, Grymonpre R, Mahmud SM. Atypical antipsychotic drug use and falls among nursing home residents in Winnipeg, Canada. Int J Geriatr Psychiatry. 2014;30(8):842-50. 620. Mathews M, Gratz S, Adetunji B, George V, Mathews M, Basil B. Antipsychotic-induced movement disorders: evaluation and treatment. Psychiatry (Edgmont). 2005;2(3):36-41. 621. Brimelow RE, Wollin JA, Byrne GJ, Dissanayaka NN. Prescribing of psychotropic drugs and indicators for use in residential aged care and residents with dementia. Int Psychogeriatr. 2019;31(6):837-47. 622. van Doorn C, Gruber-Baldini AL, Zimmerman S, Hebel JR, Port CL, Baumgarten M, et al. Dementia as a risk factor for falls and fall injuries among nursing home residents. J Am Geriatr Soc. 2003;51(9):1213-8. 623. Olazaran J, Valle D, Serra JA, Cano P, Muniz R. Psychotropic medications and falls in nursing homes: a cross-sectional study. J Am Med Dir Assoc. 2013;14(3):213-7. 624. Baandrup L, Ebdrup BH, Rasmussen JO, Lindschou J, Gluud C, Glenthoj BY. Pharmacological interventions for benzodiazepine discontinuation in chronic benzodiazepine users. Cochrane Database Syst Rev. 2018;3:CD011481. 625. Thompson W, Quay TA, Rojas-Fernandez C, Farrell B, Bjerre LM. Atypical antipsychotics for insomnia: a systematic review. Sleep Med. 2016;22:13-7. 626. Strawn JR, Geracioti L, Rajdev N, Clemenza K, Levine A. Pharmacotherapy for generalized anxiety disorder in adult and pediatric patients: an evidence-based treatment review. Expert Opin Pharmacother. 2018;19(10):1057-70. 627. Dyer SM, Gnanamanickam ES, Liu E, Whitehead C, Crotty M. Diagnosis of dementia in residential aged care settings in Australia: an opportunity for improvements in quality of care? Australas J Ageing. 2018;37(4):E155-E8. 628. Brodaty H, Howarth GC, Mant A, Kurrle SE. General practice and dementia. A national survey of Australian GPs. Med J Aust. 1994;160(1):10-4. 629. Bliwise DL. Sleep disorders in Alzheimer's disease and other dementias. Clin Cornerstone. 2004;6 Suppl 1A:S16-28. 630. Peri K, Kerse N, Moyes S, Scahill S, Chen C, Hong JB, et al. Is psychotropic medication use related to organisational and treatment culture in residential care. J Health Organ Manag. 2015;29(7):1065-79. 631. Field A. Discovering Statistics Using SPSS. Third ed. London (UK): SAGE Publications; 2009. 632. Ballard C, Lana MM, Theodoulou M, Douglas S, McShane R, Jacoby R, et al. A randomised, blinded, placebo-controlled trial in dementia patients continuing or stopping neuroleptics (the DART-AD trial). PLoS Med. 2008;5(4):e76-e.
354 | Page Daniel J. Hoyle 633. Lai CK, Arthur DG. Wandering behaviour in people with dementia. J Adv Nurs. 2003;44(2):173-82. 634. Kane RA, Kling KC, Bershadsky B, Kane RL, Giles K, Degenholtz HB, et al. Quality of life measures for nursing home residents. J Gerontol. 2003;58(3):M240-M8. 635. Saleh N, Penning M, Cloutier D, Mallidou A, Nuernberger K, Taylor D. Social engagement and antipsychotic use in addressing the behavioral and psychological symptoms of dementia in long-term care facilities. Can J Nurs Res. 2017;49(4):144-52. 636. WHOQOL: measuring quality of life [place unknown]: World Health Organization; 2020 [cited 2020 Feb 02]. Available from: https://www.who.int/healthinfo/survey/whoqol- qualityoflife/en/. 637. Federal Register of Legislation. Quality of care principles 2014 [Internet]. Canberra (AU): Australian Government; 2014 [updated 2019 Jan 16; cited 2019 Feb 04]. Available from: https://www.legislation.gov.au/Series/F2014L00830. 638. myagedcare. Quality indicators in aged care [Internet]. 2016 [cited 2018 Jul 23]. Available from: https://www.myagedcare.gov.au/quality-and-complaints/quality- indicators-in-aged-care. 639. Department of Health. Key findings and outcomes: national aged care quality indicator program- consumer experience and quality of life pilot outcomes for residential care. [place unknown]: KPMG International; 2017. 640. Wetzels RB, Zuidema SU, de Jonghe JF, Verhey FR, Koopmans RT. Determinants of quality of life in nursing home residents with dementia. Dement Geriatr Cogn Disord. 2010;29(3):189-97. 641. Ballard C, O'Brien J, James I, Mynt P, Lana M, Potkins D, et al. Quality of life for people with dementia living in residential and nursing home care: the impact of performance on activities of daily living, behavioral and psychological symptoms, language skills, and psychotropic drugs. Int Psychogeriatr. 2001;13(1):93-106. 642. Thorgrimsen L, Selwood A, Spector A, Royan L, de Madariaga Lopez M, Woods RT, et al. Whose quality of life is it anyway? The validity and reliability of the Quality of Life- Alzheimer's Disease (QoL-AD) scale. Alzheimer Dis Assoc Disord. 2003;17(4):201-8. 643. Romhild J, Fleischer S, Meyer G, Stephan A, Zwakhalen S, Leino-Kilpi H, et al. Inter-rater agreement of the Quality of Life-Alzheimer's Disease (QoL-AD) self-rating and proxy rating scale: secondary analysis of RightTimePlaceCare data. Health Qual Life Outcomes. 2018;16(1):131. 644. Sands LP, Ferreira P, Stewart AL, Brod M, Yaffe K. What explains differences between dementia patients' and their caregivers' ratings of patients' quality of life? Am J Geriatr Psychiatry. 2004;12(3):272-80. 645. Parker B, Petrou S, Underwood M, Madan J. Can care staff accurately assess health- related quality of life of care home residents? A secondary analysis of data from the OPERA trial. BMJ Open. 2017;7(4):e012779. 646. Koo TK, Li MY. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. J Chiropr Med. 2016;15(2):155-63. 647. Richardson J, Iezzi A, Khan MA, Maxwell A. Validity and reliability of the Assessment of Quality of Life (AQoL)-8D multi-attribute utility instrument. The patient. 2014;7(1):85- 96. 648. Robertson S, Cooper C, Hoe J, Hamilton O, Stringer A, Livingston G. Proxy rated quality of life of care home residents with dementia: a systematic review. Int Psychogeriatr. 2017;29(4):569-81. 649. Lyketsos CG, Gonzales-Salvador T, Chin JJ, Baker A, Black B, Rabins P. A follow-up study of change in quality of life among persons with dementia residing in a long-term care facility. Int J Geriatr Psychiatry. 2003;18(4):275-81.
355 | Page Daniel J. Hoyle 650. Tatsumi H, Nakaaki S, Torii K, Shinagawa Y, Watanabe N, Murata Y, et al. Neuropsychiatric symptoms predict change in quality of life of Alzheimer disease patients: a two-year follow-up study. Psychiatry Clin Neurosci. 2009;63(3):374-84. 651. Ydstebø AE, Bergh S, Selbæk G, Benth JŠ, Brønnick K, Vossius C. Longitudinal changes in quality of life among elderly people with and without dementia. Int Psychogeriatr. 2018;30(11):1607-18. 652. Mlinac ME, Feng MC. Assessment of activities of daily living, self-care, and independence. Arch Clin Neuropsychol. 2016;31(6):506-16. 653. Kiely DK, Simon SE, Jones RN, Morris JN. The protective effect of social engagement on mortality in long-term care. J Am Geriatr Soc. 2000;48(11):1367-72. 654. Harper Ice G. Daily life in a nursing home: Has it changed in 25 years? J Aging Stud. 2002;16(4):345-59. 655. Walker H, Paliadelis P. Older peoples’ experiences of living in a residential aged care facility in Australia. Australas J Ageing. 2016;35(3):E6-E10. 656. Lou VW, Chi I, Kwan CW, Leung AY. Trajectories of social engagement and depressive symptoms among long-term care facility residents in Hong Kong. Age Ageing. 2013;42(2):215-22. 657. Chen LY, Liu LK, Liu CL, Peng LN, Lin MH, Chen LK, et al. Predicting functional decline of older men living in veteran homes by minimum data set: implications for disability prevention programs in long term care settings. J Am Med Dir Assoc. 2013;14(4):309.e9- 13. 658. Freeman S, Spirgiene L, Martin-Khan M, Hirdes JP. Relationship between restraint use, engagement in social activity, and decline in cognitive status among residents newly admitted to long-term care facilities. Geriatr Gerontol Int. 2017;17(2):246-55. 659. Porcelli S, Van Der Wee N, van der Werff S, Aghajani M, Glennon JC, van Heukelum S, et al. Social brain, social dysfunction and social withdrawal. Neurosci Biobehav Rev. 2019;97:10-33. 660. Graf C. The Lawton instrumental activities of daily living (IADL) scale. Gerontologist. 2009;9(3):179-86. 661. Gaugler JE, Duval S, Anderson KA, Kane RL. Predicting nursing home admission in the U.S: a meta-analysis. BMC Geriatr. 2007;7:13. 662. Vossius C, Selbaek G, Saltyte Benth J, Wimo A, Bergh S. The use of direct care in nursing home residents: a longitudinal cohort study over 3 years. Int J Geriatr Psychiatry. 2019;34(2):337-51. 663. Luttenberger K, Graessel E. Recording care time in nursing homes: development and validation of the “RUD-FOCA”(Resource Utilization in Dementia–Formal Care). Int Psychogeriatr. 2010;22(8):1291-300. 664. Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index: a simple index of independence useful in scoring improvement in the rehabilitation of the chronically ill. Md State Med J. 1965;14:61-5. 665. Quinn Terence J, Langhorne P, Stott David J. Barthel Index for stroke trials. Stroke. 2011;42(4):1146-51. 666. Young J. Rehabilitation: general principles. In: Fillit HM, Rockwood K, Woodhouse K, editors. Brocklehurst's textbook of geriatric medicine and gerontology. 7th ed. Philadelphia, PA: W.B. Saunders; 2010. p. 1-21. 667. Tsunoda K, Uchida H, Suzuki T, Watanabe K, Yamashima T, Kashima H. Effects of discontinuing benzodiazepine-derivative hypnotics on postural sway and cognitive functions in the elderly. Int J Geriatr Psychiatry. 2010;25(12):1259-65. 668. Qian S, Yu P, Hailey DM, Zhang Z, Davy PJ, Nelson MI. Time spent on daytime direct care activities by personal carers in two Australian residential aged care facilities: a time- motion study. Aust Health Rev. 2014;38(2):230-7.
356 | Page Daniel J. Hoyle 669. Hartholt KA, van Beeck EF, Polinder S, van der Velde N, van Lieshout EM, Panneman MJ, et al. Societal consequences of falls in the older population: injuries, healthcare costs, and long-term reduced quality of life. J Trauma Acute Care Surg. 2011;71(3):748-53. 670. Javelot H, Marquis A, Antoine-Bernard E, Grandidier J, Weiner L, Javelot T, et al. Benzodiazepines withdrawal: initial outcomes and long-term impact on falls in a French nursing home. Pharmacy (Basel). 2018;6(2):30. 671. Campbell AJ, Robertson MC, Gardner MM, Norton RN, Buchner DM. Psychotropic medication withdrawal and a home-based exercise program to prevent falls: a randomized, controlled trial. J Am Geriatr Soc. 1999;47(7):850-3. 672. Briesacher BA, Soumerai SB, Field TS, Fouayzi H, Gurwitz JH. Medicare part D's exclusion of benzodiazepines and fracture risk in nursing homes. Arch Intern Med. 2010;170(8):693-8. 673. Page C. Benzodiazepine use in nursing homes: too much or too little? Arch Intern Med. 2010;170(17):1609. 674. Somers M, Rose E, Simmonds D, Whitelaw C, Calver J, Beer C. Quality use of medicines in residential aged care. Aust Fam Physician. 2010;39(6):413-6. 675. Poudel A, Peel NM, Mitchell CA, Gray LC, Nissen LM, Hubbard RE. Geriatrician interventions on medication prescribing for frail older people in residential aged care facilities. Clin Interv Aging. 2015;10:1043-51. 676. Hoffmann F, Allers K. Age and sex differences in hospitalisation of nursing home residents: a systematic review. BMJ Open. 2016;6(10):e011912. 677. Arendts G, Howard K. The interface between residential aged care and the emergency department: a systematic review. Age Ageing. 2010;39(3):306-12. 678. Dwyer R, Gabbe B, Stoelwinder JU, Lowthian J. A systematic review of outcomes following emergency transfer to hospital for residents of aged care facilities. Age Ageing. 2014;43(6):759-66. 679. Pedone C, Ercolani S, Catani M, Maggio D, Ruggiero C, Quartesan R, et al. Elderly patients with cognitive impairment have a high risk for functional decline during hospitalization: the GIFA study. J Gerontol A Biol Sci Med Sci. 2005;60(12):1576-80. 680. Romeo R, Knapp M, Salverda S, Orrell M, Fossey J, Ballard C. The cost of care homes for people with dementia in England: a modelling approach. Int J Geriatr Psychiatry. 2016;32(12):1466-75. 681. Liperoti R, Gambassi G, Lapane KL, Chiang C, Pedone C, Mor V, et al. Cerebrovascular events among elderly nursing home patients treated with conventional or atypical antipsychotics. J Clin Psychiatry. 2005;66(9):1090-6. 682. Rothberg MB, Herzig SJ, Pekow PS, Avrunin J, Lagu T, Lindenauer PK. Association between sedating medications and delirium in older inpatients. J Am Geriatr Soc. 2013;61(6):923-30. 683. Hill AM, Hoffmann T, Hill K, Oliver D, Beer C, McPhail S, et al. Measuring falls events in acute hospitals-a comparison of three reporting methods to identify missing data in the hospital reporting system. J Am Geriatr Soc. 2010;58(7):1347-52. 684. Schmidt SG, Dichter MN, Palm R, Hasselhorn HM. Distress experienced by nurses in response to the challenging behaviour of residents – evidence from German nursing homes. J Clin Nurs. 2012;21(21-22):3134-42. 685. Plaku-Alakbarova B, Punnett L, Gore RJ. Nursing home employee and resident satisfaction and resident care outcomes. Saf Health Work. 2018;9(4):408-15. 686. Castle NG, Engberg J, Anderson R, Men A. Job satisfaction of nurse aides in nursing homes: intent to leave and turnover. Gerontologist. 2007;47(2):193-204. 687. Radford K, Meissner E. Job satisfaction and intention to stay within community and residential aged care employees. Australas J Ageing. 2017;36(3):E1-E6.
357 | Page Daniel J. Hoyle 688. King D, Wei Z, Howe A. Work satisfaction and intention to leave among direct care workers in community and residential aged care in Australia. J Aging Soc Policy. 2013;25(4):301-19. 689. Decker FH, Harris-Kojetin LD, Bercovitz A. Intrinsic job satisfaction, overall satisfaction, and intention to leave the job among nursing assistants in nursing homes. Gerontologist. 2009;49(5):596-610. 690. Choi J, Flynn L, Aiken LH. Nursing practice environment and registered nurses' job satisfaction in nursing homes. Gerontologist. 2012;52(4):484-92. 691. Edvardsson D, Fetherstonhaugh D, McAuliffe L, Nay R, Chenco C. Job satisfaction amongst aged care staff: exploring the influence of person-centered care provision. Int Psychogeriatr. 2011;23(8):1205-12. 692. Zwijsen SA, Gerritsen DL, Eefsting JA, Smalbrugge M, Hertogh CM, Pot AM. Coming to grips with challenging behaviour: a cluster randomised controlled trial on the effects of a new care programme for challenging behaviour on burnout, job satisfaction and job demands of care staff on dementia special care units. Int J Nurs Stud. 2015;52(1):68-74. 693. Pringsheim T, Gardner D, Addington D, Martino D, Morgante F, Ricciardi L, et al. The assessment and treatment of antipsychotic-induced akathisia. Can J Psychiatry. 2018;63(11):719-29. 694. Zwijsen SA, Smalbrugge M, Eefsting JA, Twisk JW, Gerritsen DL, Pot AM, et al. Coming to grips with challenging behavior: a cluster randomized controlled trial on the effects of a multidisciplinary care program for challenging behavior in dementia. J Am Med Dir Assoc. 2014;15(7):531.e1-.e10. 695. Faria R, Barbieri M, Light K, Elliott RA, Sculpher M. The economics of medicines optimization: policy developments, remaining challenges and research priorities. Br Med Bull. 2014;111(1):45-61. 696. Hasan SS, Thiruchelvam K, Kow CS, Ghori MU, Babar Z-U-D. Economic evaluation of pharmacist-led medication reviews in residential aged care facilities. Expert Rev Pharmacoecon Outcomes Res. 2017;17(5):431-9. 697. Patterson SM, Hughes CM, Cardwell C, Lapane KL, Murray AM, Crealey GE. A cluster randomized controlled trial of an adapted U.S. model of pharmaceutical care for nursing home residents in Northern Ireland (Fleetwood Northern Ireland study): a cost- effectiveness analysis. J Am Geriatr Soc. 2011;59(4):586-93. 698. Pharmaceutical Benefits Scheme. DPMQ [Internet]. Canberra (AU): Australian Government Department of Health; 2011 [updated 2015 Aug 24; cited 2019 Apr 24]. Available from: https://dev.pbs.gov.au/docs/glossary/DPMQ.html. 699. Department of Health. Medicare Benefits Schedule Book- operating from 01 July 2016. Canberra (AU): Commonwealth of Australia; 2015. 700. Herzig SJ, LaSalvia MT, Naidus E, Rothberg MB, Zhou W, Gurwitz JH, et al. Antipsychotics and the risk of aspiration pneumonia in individuals hospitalized for nonpsychiatric conditions: a cohort study. J Am Geriatr Soc. 2017;65(12):2580-6. 701. Maidment ID, Barton G, Campbell N, Shaw R, Seare N, Fox C, et al. MEDREV (pharmacy- health psychology intervention in people living with dementia with behaviour that challenges): the feasibility of measuring clinical outcomes and costs of the intervention. BMC Health Serv Res. 2020;20(1):157. 702. Banerjee S, Hellier J, Dewey M, Romeo R, Ballard C, Baldwin R, et al. Sertraline or mirtazapine for depression in dementia (HTA-SADD): a randomised, multicentre, double-blind, placebo-controlled trial. Lancet. 2011;378(9789):403-11. 703. Dhuny S, Foley T, Jennings A. General practitioners’ knowledge of and attitudes towards prescribing psychoactive drugs in dementia care: a cross-sectional questionnaire study. Ir J Med Sci. 2020:1-9.
358 | Page Daniel J. Hoyle 704. Curtain CM, Williams M, Cousins JM, Peterson GM, Winzenberg T. Vitamin D Supplementation in Tasmanian Nursing Home Residents. Drugs Aging. 2016;33(10):747-54. 705. Australian Government. Aged Care Royal Commission signed letters [Internet]. Canberra (AU): Commonwealth of Australia; 2018 [cited 2018 Oct 09]. Available from: https://www.ag.gov.au/About/RoyalCommissions/Documents/Aged-Care-Royal- Commission-Signed-Letters-Patent-8-Oct-2018.pdf. 706. Hasham N. PM calls royal commission into aged care after inexcusable 'failures' [Internet]. Pyrmont (AU): The Sydney Morning Herald; 2018 [cited 2018 Sep 16]. Available from: https://www.smh.com.au/politics/federal/pm-calls-royal-commission- into-aged-care-after-inexcusable-failures-20180915-p5040n.html. 707. Royal Commission into Aged Care Quality and Safety. Interim report: neglect. Canberra (AU): Commonwealth of Australia; 2019. 708. Holland R, Lenaghan E, Harvey I, Smith R, Shepstone L, Lipp A, et al. Does home based medication review keep older people out of hospital? The HOMER randomised controlled trial. BMJ. 2005;330(7486):293. 709. Evaluation of the Quality Use of Medicines program. [place unknown]: URBIS; 2018. Report No.: Final report. 710. Center for Clinical Standards and Quality/Survey & Certification Group. Update report on the national partnership to improve dementia care in nursing homes. Baltimore, Maryland: Centers for Medicare & Medicaid Services; 2016. Report No.:S&C: 16-28-NH 711. French SD, Green SE, O’Connor DA, McKenzie JE, Francis JJ, Michie S, et al. Developing theory-informed behaviour change interventions to implement evidence into practice: a systematic approach using the Theoretical Domains Framework. Implement Sci. 2012;7(1):38. 712. Finkel SI, Lyons JS, Anderson RL. Reliability and validity of the Cohen–Mansfield agitation inventory in institutionalized elderly. Int J Geriatr Psychiatry. 1992;7(7):487- 90.
359 | Page Daniel J. Hoyle Appendices
Table of contents
Appendix A Summary of intervention ...... 365 Appendix B Reliability and validity of outcome tools ...... 384 Appendix C Recruitment ...... 388 Appendix D Question booklet ...... 400 Appendix E Data collection workbook ...... 425 Appendix F Register of falls, behavioural episodes, hospitalisations and general practitioner consultations ...... 433 Appendix G Measure of Job Satisfaction survey ...... 439 Appendix H Quality of life assessment for completion by the person responsible for the resident ...... 442 Appendix I Ethics approval letter ...... 448 Appendix J Anatomical Therapeutic Classification codes used to test changes in other medication classes ...... 450 Appendix K Quality of life ...... 451 Appendix L Job satisfaction by occupation ...... 467
360 | Page Daniel J. Hoyle List of figures
Figure 74: Implementation timeline within residential aged care facilities (RACFs). Adapted from
Westbury et al., 2016.217 ...... 371
Figure 75: Sample sedative review plan.217 ...... 375
Figure 76: Reducing Use of Sedatives (RedUSe) guidelines for antipsychotic use in people with
dementia. Reproduced from Westbury et al., 2016.217 ...... 378
Figure 77: Reducing Use of Sedatives (RedUSe) guidelines for benzodiazepine use. Reproduced
from Westbury et al., 2016.217 ...... 379
Figure 78: Front of postcard. Reproduced from Westbury et al., 2016.217 ...... 382
Figure 79: Back of postcard. Reproduced from Westbury et al., 2016.217 ...... 383
Figure 80: Beeswarm plots of the AQoL-4D (A) utility, and (B) independent living, (C)
relationships, (D) physical senses and (E) psychological wellbeing dimension scores at
baseline and four months for all residents taking an antipsychotic at baseline...... 453
Figure 81: Scatterplots showing the relationship between change (%) in the chlorpromazine
(CPZ) daily dose equivalent (DDE) and changes in the AQoL-4D (A) utility, and (B)
independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing
dimension scores...... 455
Figure 82: Beeswarm plots of the AQoL-4D (A) utility, and (B) independent living, (C)
relationships, (D) physical senses and (E) psychological wellbeing dimension scores at
baseline and four months for all residents taking a benzodiazepine at baseline...... 456
Figure 83: Scatterplots showing the relationship between change (%) in the diazepam daily dose
equivalent (DDE) and changes in the AQoL-4D (A) utility, and (B) independent living, (C)
relationships, (D) physical senses and (E) psychological wellbeing dimension scores. .... 458
Figure 84: Beeswarm plots of the AQoL-4D (A) utility, and (B) independent living, (C)
relationships, (D) physical senses and (E) psychological wellbeing dimension scores at
baseline and four months for all residents taking an antipsychotic at baseline...... 460
361 | Page Daniel J. Hoyle Figure 85: Scatterplots showing the relationship between change (%) in the chlorpromazine
(CPZ) daily dose equivalent (DDE) and changes in the AQoL-4D (A) utility, and (B)
independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing
dimension scores...... 462
Figure 86: Beeswarm plots of the AQoL-4D (A) utility, and (B) independent living, (C)
relationships, (D) physical senses and (E) psychological wellbeing dimension scores at
baseline and four months for all residents taking a benzodiazepine at baseline...... 463
Figure 87: Scatterplots showing the relationship between change (%) in the diazepam daily dose
equivalent (DDE) and changes in the AQoL-4D (A) utility, and (B) independent living, (C)
relationships, (D) physical senses and (E) psychological wellbeing dimension scores. .... 466
Figure 88: Beeswarm plots of the adjusted-Measure of Job Satisfaction factor scores at baseline
and four months for registered nurses...... 468
Figure 89: Beeswarm plot of the adjusted-Measure of Job Satisfaction total scores at baseline
and four months for registered nurses...... 469
Figure 90: Beeswarm plots of the adjusted-Measure of Job Satisfaction factor scores at baseline
and four months for enrolled nurses...... 471
Figure 91: Beeswarm plot of the adjusted-Measure of Job Satisfaction total scores at baseline
and four months for enrolled nurses...... 472
Figure 92: Beeswarm plots of the adjusted-Measure of Job Satisfaction factor scores at baseline
and four months for personal care assistants...... 474
Figure 93: Beeswarm plot of the adjusted-Measure of Job Satisfaction total scores at baseline
and four months for personal care assistants...... 475
362 | Page Daniel J. Hoyle List of tables
Table 64: Pre-programmed pharmacist comments available for selection in the sedative review
plan...... 375
Table 65: Pre-programmed champion nurse comments available for selection in the sedative
review plan...... 377
Table 66: Anatomical Therapeutic Classification codes used to identify and subsequently test
for changes in other medication classes that may have variable effects on neuropsychiatric
symptoms...... 450
Table 67: Inter-rater reliability between the resident, ‘person responsible’ for the resident and
residential aged care staff...... 451
Table 68: Mean Assessment of Quality of Life-4D utility and dimension scores at baseline and
four months...... 452
Table 69: Estimated changes in the AQoL-4D utility and dimension scores associated with
changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between
baseline and four months...... 454
Table 70: Mean Assessment of Quality of Life-4D utility and dimension scores at baseline and
four months...... 455
Table 71: Estimated changes in the AQoL-4D utility and dimension scores associated with
changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between
baseline and four months...... 457
Table 72: Mean Assessment of Quality of Life-4D utility and dimension scores at baseline and
four months...... 459
Table 73: Estimated changes in the AQoL-4D utility and dimension scores associated with
changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between
baseline and four months...... 461
363 | Page Daniel J. Hoyle Table 74: Mean Assessment of Quality of Life-4D utility and dimension scores at baseline and
four months...... 463
Table 75: Estimated changes in the AQoL-4D utility and dimension scores associated with
changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between
baseline and four months...... 464
Table 76: Demographic data for registered nurses participating in the job satisfaction survey at
baseline and four months...... 467
Table 77: Mean adjusted-Measure of Job Satisfaction factor scores at baseline and four months
for registered nurses...... 468
Table 78: Demographic data for enrolled nurses participating in the job satisfaction survey at
baseline and four months...... 469
Table 79: Mean adjusted-Measure of Job Satisfaction factor scores at baseline and four months
for enrolled nurses...... 470
Table 80: Demographic data for personal care assistants participating in the job satisfaction
survey at baseline and four months...... 472
Table 81: Mean adjusted-Measure of Job Satisfaction factor scores at baseline and four months
for personal care assistants...... 473
364 | Page Daniel J. Hoyle Appendix A Summary of intervention
The RedUSe project originated in 2008 as the central component of Associate Professor Juanita
Breen’s (previously Westbury) doctoral thesis entitled, ‘Roles for pharmacists in improving the quality use of psychotropic medicines in Residential Aged Care Facilities’.497 The main aim of the
RedUSe project was to reduce the inappropriate use of antipsychotics and benzodiazepines in
RACFs.
In June 2013, Associate Professor Breen and her team at the University of Tasmania received a grant from the Government Department of Health and Ageing to expand RedUSe nationally (as of November 2015, the grant was administered by the National Aged Care Grants
Hub at the Department of Health as part of the ‘Dementia and Aged Care Service Fund’).217
Overall, the national expansion of RedUSe was implemented in 150 RACFs. The national expansion of RedUSe aimed to:
• promote the quality use of antipsychotic and benzodiazepine medication in RACFs;
• deliver a national, co-ordinated, multi-strategic intervention program that targets
inappropriate antipsychotic and benzodiazepine use in RACFs;
• enable nursing staff, prescribers and pharmacists to effectively work together to ensure
the appropriate and optimal use of antipsychotic and benzodiazepine medications in
RACFs;
• promote an overall awareness of the substantial risks and limited benefits associated
with the use of antipsychotic and benzodiazepine medication in older people, and to
encourage the use of non-drug strategies to manage neuropsychiatric symptoms in this
age group;
• equip pharmacists with skills and expertise to perform QUM strategies in RACFs, with
the specific aim to reduce reliance on psycholeptic use; and,
365 | Page Daniel J. Hoyle • take the electronic local audit concept of the original RedUSe trial and scale it to an
enterprise-level e-Health data-mining program capable of interfacing with several
medication packaging software using secure e-Health messaging. Included would be the
ability to provide detailed reporting for RACF managers and health practitioners serving
RACFs.217
Prior to the implementation of the national expansion of RedUSe, enhancements were made to the original project based on the theory of ‘implementation science’.711 This involved the establishment of key personnel.
Key personnel
Appendix A.1.1 Steering group
A steering group was established in August 2013. The group involved key representatives from the RedUSe project staff, the RACF sector (BUPA and Southern Cross Care), RACF advocacy groups (Leading Age Services Australia and Aged and Community Services Australia), consumer groups (Council on the Ageing and Alzheimer’s Australia), the Pharmaceutical Society of
Australia (PSA), NPS MedicineWise, and medical practice including a local GP and a nurse researcher from the School of Nursing, Midwifery and Paramedicine at the Australian Catholic
University. This group was required to:
• provide advice and guidance to assist in the implementation of the project;
• provide advice on matters related to project design, research methods and risk
management for the project;
• provide advice and guidance for the development of training materials related to the
proposed outputs and outcomes of the project;
• assist the project team to ensure that the project activities are carried out in a timely
manner and in accordance with the project plan;
366 | Page Daniel J. Hoyle • contribute to dissemination and knowledge translation activities related to the project;
and,
• provide advice on opportunities to promote the project and communicate outcomes to
relevant parties as they become available.217
Appendix A.1.2 Project team
To facilitate the national expansion of RedUSe, a project team was established from the outset.
The chief investigator, Professor Gregory Peterson, was responsible for the overall governance of the project and assisted with planning and negotiations. The project lead, Associate Professor
Juanita Breen, was responsible for the overall design, coordination, evaluation and leadership of the project. A project manager was responsible for project strategy, resource management and administration. A project officer was appointed to assist with the administration. A research assistance was chosen to assist with evaluation. Lastly, two academic programmers were tasked with designing the e-Health data-mining program and an IT technician was hired on a part-time basis to assist with the installation and enquiries about the e-Health data-mining program.217
Appendix A.1.3 Project pharmacist
Six states and one territory were involved in the national expansion of RedUSe. Each state/territory was appointed a project pharmacist. Their main role was to implement and monitor the project in the states/territory appointed to them. Each project pharmacist provided training to the champion nurses (role described in Appendix A.1.4) and QUM pharmacists (role described in Appendix A.1.5) in their assigned states/territory. The training facilitated the implementation of the QUM strategies (outlined in Appendix A.3). The project pharmacists were also required to liaise between the project team, RACFs, champion nurses, QUM pharmacists and supply pharmacies.217
367 | Page Daniel J. Hoyle Appendix A.1.4 Champion nurse
Each RACF participating in RedUSe were asked to nominate an EN or RN to assist with the delivery of the RedUSe project in their RACF. This role was entitled ‘champion nurse’. The duties of the champion nurse included validating pharmacy data, facilitating the staff training and completing the sedative review plan. The champion nurses were also requested to distribute the plan to the prescriber and liaise with residents and relatives. The champion nurses attended a four-hour educational session provided by a project pharmacist to equip them with the skills required in their role.217
Appendix A.1.5 Quality use of medicine pharmacist
The existing QUM pharmacists for participating RACFs were asked to partake in the RedUSe project. In cases where the QUM pharmacist declined or when the RACF did not contract a QUM pharmacist, alternate QUM pharmacists were sought.217
As part of RedUSe, the QUM pharmacist was required to deliver two staff educational sessions and initiate the sedative review plan. The QUM pharmacists were required to attend a six-hour educational session provided by their project pharmacist to equip them with the necessary skills to deliver these tasks effectively.217
Appendix A.1.6 Supply pharmacy
The supply pharmacies were the community pharmacies that provided medications for RACFs involved in RedUSe. The supply pharmacy was required to install and operate the e-Health data- mining program.217
Armed with these key personnel, the national expansion of RedUSe utilised QUM strategies to improve the review and reduction of antipsychotics and benzodiazepines.
Quality Use of Medicines strategies
Appendix A.2.1 Reducing Use of Sedatives project implementation timeline
Figure 74 describes the implementation timeline used in the national expansion of RedUSe.
368 | Page Daniel J. Hoyle The RedUSe project started with a meeting between the project pharmacists and management at the RACF. After the visit and prior to baseline, the RACF nominated a champion nurse who received a half-day training delivered by their project pharmacist.217
At baseline, prescribing data were extracted from the supply pharmacy. The prescribing data were benchmarked against the national average and presented in a one-hour nursing staff educational session provided by the QUM pharmacist. This information was also used to create sedative review plans where the QUM pharmacist, champion nurse and prescriber were required to review residents who were taking antipsychotics and/or benzodiazepines on a regular basis (e.g., daily).217
At three months, extraction of the prescribing data was repeated. Again, these data were benchmarked and presented to nursing staff in a second one-hour educational session provided by the QUM pharmacist. Similarly, the information was used to create a second sedative review plan.217
At six months, a final extraction of prescribing data was undertaken, and the results were presented to the RACFs.217
Appendix A.2.2 The e-Health data-mining program
A dedicated medication audit data-mining program was installed at each supply pharmacy. The program extracted prescribing information from the pharmacy’s medication package software and uploaded the information to a secure website. Community pharmacies that supply medications to RACFs often use medication packing software to prepare and label resident medications into individualised blistered packs or sachets to facilitate drug administration by nursing staff at the RACF.217
The e-Health data-mining program was used to collect information about medication prescribing in each participating RACF at three-time points (baseline, three months, and six months). The champion nurse at the RACF subsequently verified dosing details for prescribed
369 | Page Daniel J. Hoyle psycholeptic medications against the residents’ medication charts via the RedUSe website. In addition, the champion nurse removed details of residents who had died or left the RACF, indicated whether residents were receiving palliative care or respite care, and added new residents. Psycholeptic medications that were not packed or identified by the packing program
(e.g., clonazepam drops, risperidone wafers) were entered into the database separately.217
Post-intervention evaluations later revealed that the e-Health data-mining program identified over 95% of psycholeptic medications being taken.536
A major feature of the RedUSe data-mining program was the ability to produce an individualised report for each RACF. The report benchmarked the rates of antipsychotic and benzodiazepine prescribing, graphically, against the average antipsychotic and benzodiazepine prevalence rates reported at baseline for that wave of delivery. This report was presented to participating RACFs at baseline, three months and six months as a five-page document and embedded into the educational PowerPoint presentation for delivery during the staff education sessions.217
Dedicated sedative review plans were generated for each resident taking psycholeptic medication on a regular basis (e.g., daily, twice daily, on alternate days, etc). These plans involved input from the QUM pharmacist and champion nurse before being forwarded to the prescriber for comment/action.217
370 | Page Daniel J. Hoyle
Figure 74: Implementation timeline within residential aged care facilities (RACFs). Adapted from Westbury et al., 2016.217 QUM=Quality Use of Medicine, GP=General Practitioner, RedUSe=Reducing Use of Sedatives.
371 | Page Daniel J. Hoyle Education
Educational components were designed by the project lead with assistance from an external consulting firm, Creativeintension®. Separate educational packages were developed for:
• the project pharmacists;
• the QUM pharmacists;
• the champion nurses; and
• the nursing staff (RN, EN and PCAs).217
Open dialogue and debate were promoted from the start. This helped to explore and discuss reservations that participants had about the project.
A 10-item Older Age Psychotropic (OAP) quiz538 was used to assess knowledge about psycholeptic medication before and after training sessions, and training evaluation forms were provided at the end of all sessions.217
Appendix A.3.1 Project pharmacist training
Project pharmacists received two to three days dedicated training in Hobart on the project, change management theory, teaching strategy and team leadership. Components of the nursing staff training sessions were outlined during the training and project pharmacists were taught how to identify and manage barriers to change. Additionally, project pharmacists were instructed on effective ways to teach the program. They were also provided with a custom- made training kit containing all training materials, a facilitator’s guidebook and a USB containing the RedUSe video.217
Appendix A.3.2 Quality use of medicine pharmacist training
QUM pharmacists were required to complete a dedicated continuing education booklet on
RACF psychotropic use, followed by a full day’s dedicated training delivered by their project pharmacist. Components of the nursing staff training sessions were outlined during the training
372 | Page Daniel J. Hoyle and QUM pharmacists were taught how to deliver the training and identify and manage barriers to change. Additionally, QUM pharmacists were instructed on effective ways to teach the program. They were also provided with a custom-made training kit containing all training materials, a facilitator’s guidebook and a USB containing the RedUSe video.217
Appendix A.3.3 Champion nurse training
The champion nurse was provided training on the nursing staff educational session and the project. After agreeing to participate in RedUSe, each champion nurse was required to read a four-page amended article called ‘Psychotropic medication use amongst older adults: what all nurses need to know?’537 and attend a four-hour training session delivered by their project pharmacist. During this training session, champion nurses were taught about the project, theory about psychotropic medication and strategies to identify and manage barriers to change.217
Appendix A.3.4 Nursing staff training
Education provided to the nursing staff was constrained to two one-hour workshops by the
QUM pharmacists; the first session was at baseline and the second at three months.217
The first nursing staff session was designed to generate enthusiasm and willingness toward psycholeptic reduction. The main objective was to challenge the belief that psycholeptic medications improve resident quality of life. The workshop started with didactic education in the form of a seven-minute video and a short PowerPoint presentation. Nursing staff were also shown graphs that benchmarked antipsychotic and benzodiazepine prescribing in their own facility against the average use in other RACFs involved in the project. The workshop involved small-group work and used a case study to promote conversation over whether psycholeptic medications improve the QoL of residents. Additionally, each participant was given a supporting workbook, printed guidelines, and a booklet of the slides.217
The second nursing staff workshop (also delivered by the QUM pharmacist) included further information on non-pharmacological strategies, case studies and provided repeated
373 | Page Daniel J. Hoyle information on the benefits and risks of antipsychotic and benzodiazepine medication for those who attended the first session as well as vital information for new attendees. Additionally, a dedicated handout on common psycholeptic agents and side effects was provided to the attendees, along with a workbook.217
Appendix A.3.5 Sedative review plan
The first data-mining audit produced sedative review plans for residents taking regular doses of antipsychotics and/or benzodiazepines. The plan described the resident’s details, psycholeptic agent(s) taken and doses, and three sections for sequential recommendations by the pharmacist, nurse and prescriber (Figure 75). The pharmacist was required to provide the first recommendation, followed by the nurse, then the prescriber.
The sedative review plan, generated by the RedUSe website, included the option to use pre-programmed comments in the pharmacist and champion nurse sections to facilitate and accelerate the review process. The pre-programmed comments (Table 64 and Table 65) were available to select from a drop-down list on the RedUSe website for residents taking regular doses of medications with the ATC classification codes N05A, N05B (except N05AB04 and
N05AN01), N05C, and N03AE01. The pre-programmed comments were informed through guideline-based use of antipsychotics and benzodiazepines (e.g., if a resident was taking a long- term benzodiazepine, the pharmacist was offered the pre-programmed comment, “Tolerance develops to the hypnotic effects of benzodiazepines within two weeks. May I suggest a gradual reduction in dose”). Whilst the pre-programmed comments were available, the reviewing pharmacist had the opportunity to use free text to create individualised recommendations.217
374 | Page Daniel J. Hoyle
Figure 75: Sample sedative review plan.217
Pre-programmed comments available for selection by the pharmacist and champion nurse are detailed in Table 64 and Table 65, respectively.
Table 64: Pre-programmed pharmacist comments available for selection in the sedative review plan.
Comment Pre-programmed comment number 1 "%s is currently taking %s at a dose of %s daily. This antipsychotic was prescribed for __symptoms__. These symptoms appear to have settled. Non-drug measures are recommended first-line to manage most behaviours associated with dementia. A slow and cautious dose reduction may now be appropriate with monitoring for a return of symptoms. Please note that withdrawal of antipsychotics should be done gradually, for example, by reducing the dose by 50%% every two weeks (RANZCP guidelines 2011)."
375 | Page Daniel J. Hoyle 2 "%s is currently taking %s at a dose of %s daily for __symptoms__. In the clinical notes, %s has ongoing symptoms. Non-drug measures are recommended first-line to manage most behaviours associated with dementia. If you feel the antipsychotic has been ineffective, a cautious dose reduction may now be appropriate. Please note that withdrawal of antipsychotics should be done gradually, for example, by reducing the dose by 50%% every two weeks (RANZCP guidelines 2011)." 3 "%s is currently taking %s at a dose of %s daily for __symptoms__. %s has responded well to antipsychotic treatment. %s has trialled dose reduction twice, and on each occasion, symptoms returned. Accordingly, I would not recommend a dose reduction at this time. " 4 "%s is currently taking %s at a dose of %s daily for __symptoms__. %s is at the end stage of a terminal illness. I would not support a dose reduction of %s antipsychotic medication at this time." 5 "%s is currently taking %s at a dose of %s at night for sleep. Despite this treatment %s continues to have difficulty sleeping. Given the length of usage, it is unlikely that it is providing an ongoing benefit. A slow withdrawal of this agent may now be appropriate. The suggested rate of withdrawal for benzodiazepines is to reduce the daily dose by 25%% per fortnight (British Journal of Psychiatry, 2003). Sleep hygiene measures such as reducing daytime napping, providing settling strategies and activities during the day are recommended before medication." 6 "%s is currently taking %s at a dose of %s at night for sleep. %s sleep problems appear to have settled. Given the length of usage, it is unlikely that it is providing an ongoing benefit. A slow withdrawal of this agent may now be appropriate. The suggested rate of withdrawal for benzodiazepines is to reduce the daily dose by 25%% per fortnight (British Journal of Psychiatry, 2003). Sleep hygiene measures such as reducing daytime napping, providing settling strategies and activities during the day are recommended before medication." 7 "%s is currently taking %s at a dose of %s daily for anxiety/agitation. Despite this treatment, %s continues to have experience these symptoms. Given the length of usage, it is unlikely that %s benzodiazepine is providing an ongoing benefit. A slow withdrawal of this agent may now be appropriate. The suggested rate of withdrawal for benzodiazepines is to reduce the daily dose by 25%% per fortnight (British Journal of Psychiatry, 2003). Antidepressant therapy has been shown to have longer lasting effects on anxiety." 8 "%s is currently taking %s at a dose of %s daily for anxiety/agitation. These symptoms appear to have settled. A slow withdrawal of this agent may now be appropriate. The suggested rate of withdrawal for benzodiazepines is to reduce the daily dose by 25%% per fortnight (British Journal of Psychiatry, 2003). Please note that antidepressant therapy has been shown to have longer lasting effects on anxiety. " 9 "%s continues on %s for __symptoms__ which appears to be effective. %s trialled a dosage reduction of %s benzodiazepine last year and %s symptoms returned. Accordingly, I would not recommend a dose reduction at this time. Antidepressant therapy has been shown to have longer lasting effects on anxiety" 10 "%s is currently taking %s at a dose of %s. %s is at the end stage of a terminal illness. I would not support a dose reduction of %s benzodiazepine medication at this time." % represents fields that are replaced by the resident’s name, drug name, or drug dose.
376 | Page Daniel J. Hoyle Table 65: Pre-programmed champion nurse comments available for selection in the sedative review plan.
Comment Pre-programmed comment number 1 "%s symptoms have been stable for some time. I support the pharmacist's recommendation for a gradual trial withdrawal of %s medication as suggested." 2 "%s continues to have __ despite taking %s at a daily dose of %s. I support the pharmacist's recommendation to reduce %s %s as suggested."; 3 "I do not support the pharmacist's recommendation to reduce %s's %s for the following reason/s: __ "; 4 "I support the pharmacist's recommendation to not reduce %s's %s"; % represents fields that are replaced by the resident’s name or drug name.
Other components of the Reducing Use of Sedatives project expansion
Appendix A.4.1 Guidelines
Guidelines used in the national expansion of RedUSe were produced for the original RedUSe trial. The guidelines were based on recommended best-practice for antipsychotic and benzodiazepine use, from the International Psychogeriatric Association, the Royal Australian
College of Practitioners and the RANZCP.31-33, 77, 106 Information provided in the guidelines were updated for the national expansion of RedUSe but the format remained similar. The guidelines included information on the risk/benefit profile for antipsychotics and benzodiazepines, and guidance on recommended doses and duration of use. To assist with reducing these medications, a sample dosage reduction schedule was included.217 The guidelines were reviewed by NPS MedicineWise, who provided the majority of academic detailing to prescribers, and the RedUSe project’s steering group committee, to ensure the message remained consistent with the original guidelines. Examples of the RedUSe guidelines are shown in Figure
76 and Figure 77. After review, the guidelines were printed professionally on thick card and provided to more than 1,400 nursing staff and 80 pharmacists at the training sessions. The guidelines were also distributed to approximately 600 prescribers. These prescribers were considered by the RACF as one of the top prescribers (i.e., prescribers who provide services to the most residents).217
377 | Page Daniel J. Hoyle
Figure 76: Reducing Use of Sedatives (RedUSe) guidelines for antipsychotic use in people with dementia. Reproduced from Westbury et al., 2016.217
378 | Page Daniel J. Hoyle
Figure 77: Reducing Use of Sedatives (RedUSe) guidelines for benzodiazepine use. Reproduced from Westbury et al., 2016.217
Appendix A.4.2 Prescriber academic detailing
In the original RedUSe trial, a lack of prescriber engagement was identified as a barrier to effective review and reduction of psycholeptic medications.497 To improve the engagement of prescribers in the national expansion of RedUSe, educational outreach visits to prescribers were arranged. These face-to-face visits, commonly referred to as ‘academic detailing’, involved trained facilitators visiting prescribers at their place of practice to provide information on antipsychotic and benzodiazepine use and the RedUSe project to change how they practice.217
379 | Page Daniel J. Hoyle NPS MedicineWise, Australia’s leading organisation that provides academic detailing to prescribers, was contracted to develop an academic detailing program about the quality use of psycholeptic medication. As part of the contract, NPS MedicineWise were required to:
• develop key messages for the academic detailing;
• produce an educational visiting card that included evidence-based information; and,
• develop background materials, training and ongoing support for the NPS MedicineWise
academic detailers delivering the intervention.217
Materials were reviewed by the RedUSe project team, a geriatrician and GP, and were piloted and updated. Background materials outlining the quality use of antipsychotics for neuropsychiatric symptoms and benzodiazepines in RACFs were provided to ten academic detailers from NPS MedicineWise. The academic detailers were also provided with the evidence that specifically supported the key messages of RedUSe and best practice for antipsychotic and benzodiazepine prescribing. Academic detailers attended RedUSe QUM pharmacist training provided by the University of Tasmania as well as an additional day focussed on the delivery of information through academic detailing. NPS MedicineWise provided ongoing support to the academic detailers through an online learning platform, online discussion forum and regular teleconferences led by the NPS MedicineWise clinical lead.
Prescribers were identified for academic detailing by their RACF. Each RACF were asked to provide a list of five to 10 prescribers who serviced their RACF most frequently. Each prescriber was sent an information letter about RedUSe, the RedUSe guidelines, a promotional flier, and a sample sedative review plan. The prescriber’s contact details were sent to NPS
MedicineWise, who subsequently invited each prescriber via fax and phone to an academic detailing session. The NPS also developed their own invitation leaflet for RACF staff to provide to attending prescribers who expressed interest in the RedUSe project.217
380 | Page Daniel J. Hoyle In addition to NPS MedicineWise, Drug And Therapeutics Information Service (DATIS) were asked to provide academic detailing services in South Australia from mid-August 2014 due to a poor uptake of academic detailing provided by NPS MedicineWise during the first implementation wave of RedUSe. DATIS used the same training materials and recruitment process as NPS MedicineWise.217
Appendix A.4.3 Video
An eight-minute video was made to introduce RedUSe, explain key strategies, and outline the benefits of the project for residents, nurse prescribers, doctors and nursing staff. The video was played at each of the training events for the QUM pharmacists, champion nurses, and in the first staff education session. Additionally, the video was uploaded to YouTube and embedded onto the RedUSe website.217
The PSA was contracted to design and produce the video which was filmed over several days in October 2013. The video can be viewed at this link: https://www.youtube.com/watch?v=yIxC3IKu5PU.
Appendix A.4.4 Older age psychotropic quiz
The Older Age Psychotropic (OAP) quiz was developed by Associate Professor Juanita Breen in
2008 as part of her doctoral thesis.497 Initially, the OAP was developed to test the knowledge levels of staff before and after the intervention. However, the OAP proved useful to initiate a discussion regarding psycholeptic use and helped identify key deficits in nursing staff knowledge.
In 2016, the validation of the OAP quiz was published in the Journal of Gerontological Nursing.538
Appendix A.4.5 Resident and relative involvement
Resident and relative pamphlets promoting the appropriate use of benzodiazepines and a relative pamphlet promoting the appropriate use of antipsychotics for neuropsychiatric symptoms of dementia were developed with the assistance from two consumer representatives from the steering group. The leaflets were based on consumer information from the Alzheimer’s
381 | Page Daniel J. Hoyle Association in the United Kingdom but have been simplified and modified to incorporate
Australian drug names. Prototypes of the pamphlets were presented to the consumer advocate group at Alzheimer’s Australia who provided comments on the content. The pamphlets were modified according to the feedback and in early 2015, the Dementia Behaviour Management
Advisory Service (DBMAS) asked if they could distribute the pamphlets.217
In addition to the information pamphlets, small reply-paid postcards were produced at the request of the Council on the Ageing (COTA) representative on the steering group (Figure
78 and Figure 79). The postcard contained website details. However, the postcard could also be sent to the RedUSe project team should further information be required (e.g., relative pamphlets). These postcards were distributed to each RACF by the project pharmacists and sent in recruitment letters to the ‘person responsible’ for residents eligible for participation in the clinical study (focus of this thesis).217
Figure 78: Front of postcard. Reproduced from Westbury et al., 2016.217
382 | Page Daniel J. Hoyle
Figure 79: Back of postcard. Reproduced from Westbury et al., 2016.217
Appendix A.4.6 Newsletters and other printed materials
A newsletter was produced every six months by the RedUSe project team. The newsletter provided news about the project and updates on progress. RedUSe team members were encouraged to provide stories about the staff involved with RedUSe and how residents had responded to psycholeptic reduction.217
Posters were produced for each RACF to promote nursing staff attendance at the
RedUSe educational session and to indicate the relationship between the RACF and RedUSe to visitors of the RACF. A separate information sheet was developed for prescribers and a flier advertising the nursing staff and PCA education was created.217
383 | Page Daniel J. Hoyle Appendix B Reliability and validity of outcome tools
Appendix B.1.1 Neuropsychiatric Inventory-Nursing Home version
The NPI-NH is widely used in both research and clinical care settings, and has been assessed for validity and reliability.545, 555, 556, 558, 560 A Cronbach’s α-value of 0.67 has been reported, indicating acceptable reliability.558 Test-retest reliability has been reported as moderate-to-good for all but one domain with Spearman correlations ranging from 0.44 to 0.78. The night-time behaviour domain had the lowest test-retest reliability (r=0.23).545
The Norwegian version of the NPI-NH has also been assessed against the Behavioural
Pathology in Alzheimer’s disease (BEHAVE-AD) scale; a validated measure of neuropsychiatric symptoms of dementia.556 All subscales of the BEHAVE-AD significantly correlated with corresponding subscales of the NPI-NH (r=0.38 to 0.72).
Furthermore, the five-factor model (Table 18) has been validated against a range of assessment tools.558 The NPI-NH ‘agitation’ factor was moderately correlated with the CMAI
(r=0.55), and the Aggression (r=0.47) and Compliance (r=0.48) items from the Geri-SNAP tool.558
The NPI-NH ‘mood’ factor correlates moderately with the Cornell Depression Scale (r=0.59), the
Brief Psychiatric Rating Scale (r=0.53), and the Anxiety item (r=0.34) from the Geri-SNAP tool.558
The NPI-NH ‘psychosis’ factor is correlated with the Psychotic Features item (r=0.54) from the
Geri-Snap tool.558 The NPI-NH ‘sleep/motor activity’ factor correlated with the Aberrant Motor
(r=0.26) and Sleep (r=0.42) items from the Geri-SNAP tool.558 Lastly, the NPI-NH ‘elevated behaviour’ factor demonstrated correlations with the Mania Rating Scale (r=0.37).558
Appendix B.1.2 Cohen-Mansfield Agitation Inventory
The CMAI has been widely used in geriatric research, and the reliability and validity for this measure are well established in RACFs.545, 561 High internal consistency has been reported with
Cronbach’s alpha scores ≥0.70 for all factors.562 Reliability is also good with test-retest
384 | Page Daniel J. Hoyle correlations ranging from 0.82 (physically aggressive behaviour) to 0.86 (verbally agitated behaviour).545
Correlations between the CMAI and corresponding components of BEHAVE-AD (r=0.28 to 0.43) and Behavioural Syndromes Scale for Dementia (r=0.40 to 0.65) indicate adequate validity.712
Appendix B.1.3 Assessment of Quality of Life-4D
The reliability and validity of the AQoL instruments have been demonstrated across several publications.541, 565-569
The AQoL instruments are the only QoL utility instruments to be constructed using psychometric methods to achieve content and construct validity.567 In 1999, Hawthorne et al. described the construction of the AQoL tool.541 Using a sample of 255 people from hospital
(n=143) and community (n=112), the researchers used exploratory factor analysis to show that the dimensions were orthogonal (i.e., statistically independent) and unidimensional.541
Additionally, internal consistency was relatively good, ranging from α=0.52 for the psychological wellbeing dimension to α=0.77 for the independent living dimension.541
Similar utility scores have been reported between the AQoL, Canadian Health Utilities
Index 3 and EQ-5D among 976 Australians.565 On the other hand, the Finnish 15D and Short
Form-6D had higher utilities with smaller differences between individuals.565 AQoL also had relatively greater sensitivity to health states.565 However, the researchers emphasised that no single QoL utility instrument is considered ‘gold standard’ and researchers should select instruments that are sensitive to the health state being investigated.565
It should be noted that the AQoL-4D has not been specifically validated in residents of
RACFs. However, the AQoL-4D has been used to assess QoL in hospital patients awaiting transfer to a RACF.548 Furthermore, Osborne et al. have validated the AQoL-4D in community-dwelling older people (n=1,056) and compared the results to the Short Form (SF)-36.573 Overall, the
385 | Page Daniel J. Hoyle AQoL-4D was found to be a valid and reliable scale in this group with an internal consistency of
0.73 for the AQoL-4D utility score. Compared to the SF-36, the AQoL-4D appeared to be more suitable for assessing health-related QoL in chronically ill older people given the lack of floor effects.573 Further, the AQoL-4D had greater sensitivity when detecting longitudinal changes in
QoL between baseline and follow-up for those people who entered a RACF.573
Appendix B.1.4 Multidimensional Observation Scale for Elderly Subjects-withdrawal subscale
The MOSES-withdrawal subscale has been shown to be valid and reliable in the RACF setting.583,
585
In 1987, Helmes et al. assessed social engagement in 2,542 individuals across seven psychiatric facilities, 22 RACFs, nine homes for the aged, and seven continuing-care hospitals in
Ontario, Canada, by interviewing nursing staff. Inter-rater reliability was assessed by asking a second staff member to independently rate the participants. Within the RACF setting, the
MOSES-withdrawal subscale had an inter-rater reliability of 0.71 (among 54 residents) and internal consistency of 0.77 (n=970).583 The MOSES-withdrawal subscale correlated highly with the London Psychogeriatric Rating Scale Disengagement subscale (r=0.73) and the Physical and
Mental Impairment-of-function Evaluation withdrawal/apathetic subscale (r=0.78).583
Pruchno et al. also assessed the reliability of the MOSES-withdrawal subscale in 536
RACF residents.585 The MOSES-withdrawal subscale had acceptable internal consistency with a
Cronbach α=0.84.585
Appendix B.1.5 Physical Self-Maintenance Scale
The PSMS has been shown to be a valid and reliable tool to assess the function of older adults.587,
588
In 1999, 25 consecutive outpatients with Alzheimer’s disease were assessed with the
PSMS.588 The patients were assessed by a neuropsychiatrist, neurologist, nurse, occupational therapist and a clinical psychologist. No significant differences in ratings were found between
386 | Page Daniel J. Hoyle these professions, with inter-rater correlations >0.86 indicating good inter-rater reliability of the PSMS.588
The validity of the PSMS has been assessed in 180 institutionalised older people.587 The
PSMS had moderate correlations with the Physical classification, Mental status questionnaire,
Instrumental activities of daily living, and Behavior and adjustment questionnaires.587
387 | Page Daniel J. Hoyle Appendix C Recruitment
Residential aged care facility - letter of invitation
388 | Page Daniel J. Hoyle Residential aged care facility - information sheet
389 | Page Daniel J. Hoyle 390 | Page Daniel J. Hoyle Residential aged care facility - consent form
391 | Page Daniel J. Hoyle Resident - letter of invitation
392 | Page Daniel J. Hoyle Resident - information sheet
393 | Page Daniel J. Hoyle
394 | Page Daniel J. Hoyle Resident - consent form
395 | Page Daniel J. Hoyle Person responsible for resident - letter of invitation
396 | Page Daniel J. Hoyle Person responsible for resident - information sheet
397 | Page Daniel J. Hoyle
398 | Page Daniel J. Hoyle Person responsible for resident - consent form
399 | Page Daniel J. Hoyle Appendix D Question booklet
400 | Page Daniel J. Hoyle
401 | Page Daniel J. Hoyle
402 | Page Daniel J. Hoyle
403 | Page Daniel J. Hoyle
404 | Page Daniel J. Hoyle
405 | Page Daniel J. Hoyle
406 | Page Daniel J. Hoyle
407 | Page Daniel J. Hoyle
408 | Page Daniel J. Hoyle
409 | Page Daniel J. Hoyle
410 | Page Daniel J. Hoyle
411 | Page Daniel J. Hoyle
412 | Page Daniel J. Hoyle
413 | Page Daniel J. Hoyle
414 | Page Daniel J. Hoyle
415 | Page Daniel J. Hoyle
416 | Page Daniel J. Hoyle
417 | Page Daniel J. Hoyle
418 | Page Daniel J. Hoyle
419 | Page Daniel J. Hoyle
420 | Page Daniel J. Hoyle
421 | Page Daniel J. Hoyle
422 | Page Daniel J. Hoyle
423 | Page Daniel J. Hoyle
424 | Page Daniel J. Hoyle Appendix E Data collection workbook
425 | Page Daniel J. Hoyle
426 | Page Daniel J. Hoyle
427 | Page Daniel J. Hoyle
428 | Page Daniel J. Hoyle
429 | Page Daniel J. Hoyle
430 | Page Daniel J. Hoyle
431 | Page Daniel J. Hoyle
432 | Page Daniel J. Hoyle Appendix F Register of falls, behavioural episodes, hospitalisations and general practitioner consultations
433 | Page Daniel J. Hoyle 434 | Page Daniel J. Hoyle 435 | Page Daniel J. Hoyle 436 | Page Daniel J. Hoyle 437 | Page Daniel J. Hoyle 438 | Page Daniel J. Hoyle Appendix G Measure of Job Satisfaction survey
439 | Page Daniel J. Hoyle
440 | Page Daniel J. Hoyle
441 | Page Daniel J. Hoyle Appendix H Quality of life assessment for completion by the person responsible for the resident
Baseline letter and quality of life tool
442 | Page Daniel J. Hoyle
443 | Page Daniel J. Hoyle
444 | Page Daniel J. Hoyle
445 | Page Daniel J. Hoyle Four-month letter
446 | Page Daniel J. Hoyle Reminder letter
447 | Page Daniel J. Hoyle
Appendix I Ethics approval letter
448 | Page Daniel J. Hoyle
449 | Page Daniel J. Hoyle Appendix J Anatomical Therapeutic Classification codes used to test changes in other medication classes
Table 66 describes the ATC codes used to identify and, subsequently, test for changes in other medication classes that may have beneficial or negative effects on neuropsychiatric symptoms in Chapter 6.3.
Table 66: Anatomical Therapeutic Classification codes used to identify and subsequently test for changes in other medication classes that may have variable effects on neuropsychiatric symptoms.
Medication class Anatomical Therapeutic Classification codes used Cholinesterase inhibitors N06DA* Memantine N06DX01 Anticonvulsants N03AA*, N03AB*, N03AC*, N03AD*, N03AF*, N03AG*, N03AX* Adrenergic agents C07*, C02AC01 Non-benzodiazepine anxiolytics N05BB*, N05BC*, N05BD*, N05BE*, N05BX* Z-drugs N05CF* Antidepressants N06A* Tricyclic antidepressants N06AA* Mirtazapine N06AX11 Opioids N02A* Non-benzodiazepine sedatives N05CA*, N05CB*, N05CC*, N05CE*, N05CF*, N05CH*, N05CM*, N05CX* Antihistamines R06A* Anticholinergics C01BA03, N06AA09, N06AA04, N06AA01, N06AA12, N06AA03, N06AA10, N06AB05, N06AA11, N06AA06, N05AB04, R06AD02, R06AB01, R06AA08, R06AB04, R06AA04, R06AX02, R06AB06, R06AB02, R06AA02, R06AA09, N05BB01, A03BB06, A03BA03, A03AB05, G04D10, G04BD13, G04BD02, G04BD04, G04BD08, G04BD07, G04BD09, N04AA01, A03BA01, A03BA*, A04AD01
450 | Page Daniel J. Hoyle Appendix K Quality of life
Inter-rater reliability
A poor to moderate level of inter-rater reliability was found between the different raters of QoL
(RACF staff, ‘person responsible’ for the residents and the residents themselves) using the
AQoL-4D psychometric tool (Table 67). Considering this, changes in QoL were investigated separately according to the rater; RACF staff-rated QoL is reported in Chapter 6.4 and QoL rated by the ‘person responsible’ for the resident and the residents themselves is reported in
Appendix K.
Table 67: Inter-rater reliability between the resident, ‘person responsible’ for the resident and residential aged care staff.
Outcome Subjects ICC coefficient 95%CI Utility 13 0.534 0.207, 0.804 Independent living 13 0.722 0.457, 0.894 Relationships 15 0.323 0.013, 0.653 Physical senses 15 0.648 0.374, 0.849 Psychological wellbeing 15 0.400 0.086, 0.706 ICC=Intraclass correlation, 95%CI=95% confidence interval Person responsible-rated quality of life
Appendix K.2.1 Feasibility of measuring quality of life associated with antipsychotic and/or benzodiazepine dose change
QoL rated by the person responsible for the resident did not prove feasible as the majority of participants did not have data for both baseline and four months. Only 63 of 179 (35%) participants had their QoL assessed at baseline and four months by the person responsible for them. Reminder surveys were mailed to non-responders. However, the impact that the reminders had is unclear.
Appendix K.2.2 Antipsychotic users
Thirty-two residents initially taking an antipsychotic had their QoL assessed at baseline and four months by the ‘person responsible’ for them (e.g., a family member). The utility score, and independent living, relationship and psychological wellbeing dimension scores remained similar
451 | Page Daniel J. Hoyle between baseline and four months (Table 68, Figure 80). On the other hand, there was a clinically significant trend8 towards deterioration in the physical senses dimension between baseline and four months (Table 68).
Table 68: Mean Assessment of Quality of Life-4D utility and dimension scores at baseline and four months.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Utility 0.08 (-0.00, 0.16) 0.07 (0.00, 0.13) 0.802 31 0.429 Independent living 0.15 (0.06, 0.25 0.16 (0.05, 0.27) -0.222 31 0.826 Relationships 0.33 (0.21, 0.44) 0.37 (0.25, 0.48) -1.009 31 0.321 Physical senses 0.62 (0.54, 0.71) 0.56 (0.46, 0.66) 1.950 31 0.060 Psychological wellbeing 0.83 (0.77, 0.88) 0.85 (0.81, 0.89) -1.176 31 0.249 df=degrees of freedom, 95%CI=95% confidence interval
8 A change in the AQoL-4D score by ≥0.06 points is considered clinically significant.
452 | Page Daniel J. Hoyle
Figure 80: Beeswarm plots of the AQoL-4D (A) utility, and (B) independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing dimension scores at baseline and four months for all residents taking an antipsychotic at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval. AQoL-4D=Assessment of Quality of Life-4D tool.
Using multiple linear regression to control for the CMAI factor 2 and 3 baseline scores, we found no evidence of an association between antipsychotic dose change and the AQoL-4D utility score (β=-0.000; p=0.411, 95%CI=-0.001, 0.000; Table 69 and Figure 81).
Investigation into the changes in the AQoL-4D dimension scores also indicated that antipsychotic dose reduction was not associated with deteriorations in the independent living, relationships, physical senses, or psychological wellbeing dimensions (Table 69).
453 | Page Daniel J. Hoyle Similarly, the AQoL-4D utility score remained relatively constant between baseline
(mean=0.003, 95%CI=-0.067, 0.072) and four months (mean=-0.022, 95%CI=-0.034, -0.010) among residents who had their antipsychotic ceased (t(6)=0.954, p=0.377, n=7 residents).
Table 69: Estimated changes in the AQoL-4D utility and dimension scores associated with changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value Intercept -0.067 -0.135, 0.002 -2.001 0.055 Antipsychotic dose change (%) -0.000 -0.001, 0.000 -0.835 0.411
Baseline CMAI factor 2 score 0.005 -0.001, 0.010 1.801 0.082 Baseline CMAI factor 3 score -0.001 -0.009, 0.006 -0.356 0.0725 Utility Adjusted R2 0.102 Multiple R2 0.189 Intercept -0.074 -0.174, 0.026 -1.515 0.141
Antipsychotic dose change (%) -0.000 -0.001, 0.000 -0.980 0.336 Baseline CMAI factor 2 score -0.002 -0.010, 0.005 -0.634 0.531 Baseline CMAI factor 3 score 0.010 -0.001, 0.021 1.842 0.076 dimension Adjusted R2 0.062 Independent living living Independent Multiple R2 0.153 Intercept -0.089 -0.256, 0.079 -1.088 0.286
Antipsychotic dose change (%) -0.000 -0.002, 0.001 -0.752 0.458 Baseline CMAI factor 2 score 0.004 -0.009, 0.017 0.652 0.519 Baseline CMAI factor 3 score 0.006 -0.012, 0.025 0.693 0.494 dimension Relationships Adjusted R2 0.030 Multiple R2 0.124 Intercept -0.046 -0.187, 0.095 -0.665 0.511
Antipsychotic dose change (%) -0.000 -0.001, 0.001 -0.455 0.652 Baseline CMAI factor 2 score 0.002 -0.009, 0.013 0.370 0.714 Baseline CMAI factor 3 score -0.005 -0.020, 0.011 -0.595 0.556 dimension 2 Physical senses Adjusted R -0.082 Multiple R2 0.023
Intercept -0.086 -0.166, -0.006 -2.201 0.036 Antipsychotic dose change (%) -0.000 -0.001, 0.000 -0.518 0.609 Baseline CMAI factor 2 score 0.007 0.000, 0.013 2.148 0.041 Baseline CMAI factor 3 score 0.002 -0.007, 0.011 0.451 0.655
Psychological Adjusted R2 0.255 wellbeing dimension Multiple R2 0.327 AQoL-4D=Assessment of Quality of Life-4D tool, 95%CI=95% confidence interval
454 | Page Daniel J. Hoyle
Figure 81: Scatterplots showing the relationship between change (%) in the chlorpromazine (CPZ) daily dose equivalent (DDE) and changes in the AQoL-4D (A) utility, and (B) independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing dimension scores. AQoL-4D=Assessment of Quality of Life-4D tool.
Appendix K.2.3 Benzodiazepine users
Thirty-one residents initially taking benzodiazepines had their QoL assessed at baseline and four months by the ‘person responsible’ for them (e.g., family member). The utility score and dimension scores remained similar between baseline and four months (Table 70, Figure 82).
Table 70: Mean Assessment of Quality of Life-4D utility and dimension scores at baseline and four months.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Utility 0.06 (0.02, 0.10) 0.04 (0.00, 0.07) 1.493 30 0.146 Independent living 0.14 (0.06, 0.22) 0.11 (0.03, 0.20) 0.927 30 0.362 Relationships 0.43 (0.33, 0.52) 0.43 (0.32, 0.53) 0.002 30 0.999 Physical senses 0.68 (0.60, 0.77) 0.65 (0.56, 0.74) 1.042 30 0.306 Psychological wellbeing 0.77 (0.71, 0.83) 0.77 (0.71, 0.83) 0.219 30 0.828 df=degrees of freedom, 95%CI=95% confidence interval
455 | Page Daniel J. Hoyle
Figure 82: Beeswarm plots of the AQoL-4D (A) utility, and (B) independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing dimension scores at baseline and four months for all residents taking a benzodiazepine at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval. AQoL- 4D=Assessment of Quality of Life-4D tool.
Using linear regression modelling to control for the baseline CMAI factor 2 score and having a documented diagnosis of dementia, we found no evidence of an association between changes in the AQoL-4D utility score and benzodiazepine dose reduction (β=0.000; i.e., AQoL-
4D utility score was 0.00 points higher, on average, for each 10% reduction in the diazepam daily dose equivalent, p=0.869, 95%CI=-0.001, 0.001; Table 71 and Figure 83). Similarly, investigation into the changes in the AQoL-4D dimension scores also indicated that benzodiazepine dose reduction was not associated with deteriorations in the independent living, physical senses, or psychological wellbeing dimensions (Table 71). However, there was a slight, albeit non-significant, deterioration in the AQoL-4D relationships dimension (β=0.001; i.e.,
456 | Page Daniel J. Hoyle AQoL-4D relationships dimension score was 0.01 points lower, on average, for each 10% reduction in the diazepam daily dose equivalent, p=0.561, 95%CI=-0.002, 0.003; Table 71).
The AQoL-4D utility score remained relatively constant between baseline (mean=0.044,
95%CI=-0.028, 0.116) and four months (mean=0.034, 95%CI=-0.033, 0.100) among residents who had their benzodiazepine ceased (t(8)=0.787, p=0.454, n=9 residents).
Table 71: Estimated changes in the AQoL-4D utility and dimension scores associated with changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between baseline and four months.
Outcome Estimate 95%CI t-statistic p-value (β) Intercept 0.006 -0.093, 0.042 -0.784 0.440 Benzodiazepine dose change (%) 0.000 -0.001, 0.001 0.166 0.869
Documented diagnosis of 0.003 -0.060, 0.067 0.108 0.915 dementia
Utility Baseline CMAI factor 2 score 0.001 -0.004, 0.006 0.229 0.820 Adjusted R2 -0.106 Multiple R2 0.005 Intercept -0.081 -0.244, 0.081 -1.025 0.314 Benzodiazepine dose change (%) 0.000 -0.001, 0.002 0.330 0.744
Documented diagnosis of 0.061 -0.091, 0.214 0.824 0.417 dementia Baseline CMAI factor 2 score 0.003 -0.009, 0.015 0.506 0.617 dimension 2
Independent living living Independent Adjusted R -0.062 Multiple R2 0.044 Intercept 0.103 -0.120, 0.326 0.949 0.351 Benzodiazepine dose change (%) 0.001 -0.002, 0.003 0.589 0.561
Documented diagnosis of -0.011 -0.220, 0.199 -0.105 0.917 dementia Baseline CMAI factor 2 score -0.007 -0.024, 0.009 -0.905 0.374 dimension Relationships Adjusted R2 -0.067 Multiple R2 0.040 Intercept 0.006 -0.163, 0.175 0.072 0.943 Benzodiazepine dose change (%) 0.000 -0.002, 0.002 0.210 0.835
Documented diagnosis of -0.074 -0.233, 0.085 -0.954 0.349 dementia Baseline CMAI factor 2 score -0.000 -0.013, 0.012 -0.045 0.964 dimension
Physical senses Adjusted R2 -0.061 Multiple R2 0.045
457 | Page Daniel J. Hoyle Intercept 0.039 -0.072, 0.149 0.715 0.481 Benzodiazepine dose change (%) 0.000 -0.001, 0.001 -0.100 0.921
Documented diagnosis of -0.042 -0.146, 0.062 -0.834 0.412 dementia Baseline CMAI factor 2 score 0.003 -0.011, 0.006 -0.621 0.540 dimension Adjusted R2 -0.057
Psychological wellbeing Multiple R2 0.049 AQoL-4D=Assessment of Quality of Life-4D tool, 95%CI=95% confidence interval
Figure 83: Scatterplots showing the relationship between change (%) in the diazepam daily dose equivalent (DDE) and changes in the AQoL-4D (A) utility, and (B) independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing dimension scores. AQoL-4D=Assessment of Quality of Life-4D tool.
458 | Page Daniel J. Hoyle Resident-rated quality of life (antipsychotic users)
Appendix K.3.1 Feasibility of measuring quality of life associated with antipsychotic and/or benzodiazepine dose change
Only 49 of the 179 residents (27%), present at baseline and four months, had the cognitive capacity to provide informed consent to participate in the study. Of these, 41 participants (84%) had their QoL rated through an interview at baseline and four months. Whilst it may be feasible to assess QoL using the AQoL-4D in residents who have the cognitive capacity to provide informed consent, the proportion of residents in the overall sample able to rate their own QoL was only 23%.
Appendix K.3.2 Antipsychotic users
A total of four residents initially taking antipsychotics underwent psychometric testing with the
AQoL-4D tool at baseline and four months. The utility score, and independent living, and relationships dimensions decreased by clinically, 9 but not statistically, significant amounts
(Table 72 and Figure 84). On the other hand, the physical senses score increased by a clinically, albeit non-statistically significant, amount. The psychological wellbeing dimension remained relatively the same between baseline and four months.
Table 72: Mean Assessment of Quality of Life-4D utility and dimension scores at baseline and four months.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Utility 0.31 (0.10, 0.52) 0.23 (-0.09, 0.55) 0.734 3 0.516 Independent living 0.54 (0.15, 0.93) 0.45 (0.13, 0.78) 0.570 3 0.609 Relationships 0.81 (0.58, 1.03) 0.52 (0.49, 1.00) 1.856 3 0.161 Physical senses 0.84 (0.64, 1.04) 0.87 (0.72, 1.02) -0.790 3 0.488 Psychological wellbeing 0.80 (0.49, 1.11) 0.82 (0.58, 1.05) -0.241 3 0.825 df=degrees of freedom, 95%CI=95% confidence interval
9 A change in the AQoL-4D score by ≥0.06 points is considered clinically significant.
459 | Page Daniel J. Hoyle
Figure 84: Beeswarm plots of the AQoL-4D (A) utility, and (B) independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing dimension scores at baseline and four months for all residents taking an antipsychotic at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval. AQoL-4D=Assessment of Quality of Life-4D tool.
A slight, albeit non-significant, increase in the resident-rated AQoL-4D utility score was related to antipsychotic dose reduction after controlling for CMAI factor 2 and 3 baseline scores
(β=-0.001; i.e., AQoL-4D utility score was 0.01 higher for each 10% reduction in the chlorpromazine daily dose equivalent, p=0.644, 95%CI=-0.029, 0.026; Table 73 and Figure 85).
Furthermore, there were non-significant improvements in the independent living and relationships dimensions, whereas physical senses remained the same and psychological wellbeing worsened with antipsychotic dose reduction (Figure 85).
None of the residents who rated their own QoL had their antipsychotic ceased.
460 | Page Daniel J. Hoyle Table 73: Estimated changes in the AQoL-4D utility and dimension scores associated with changes (%) in the chlorpromazine daily dose equivalent (antipsychotic dose) between baseline and four months.
Outcome Estimate (β) 95%CI t-statistic p-value Intercept -1.483 -6.072, 3.106 -4.106 0.152 Antipsychotic dose change (%) -0.001 -0.029, 0.026 -0.625 0.644
Baseline CMAI factor 2 score 0.214 -0.468, 0.895 3.984 0.157 Baseline CMAI factor 3 score NA NA NA NA Utility Adjusted R2 0.828 Multiple R2 0.943 Intercept -2.161 -11.847, 7.524 -2.835 0.216
Antipsychotic dose change (%) -0.007 -0.064, 0.051 -1.428 0.389 Baseline CMAI factor 2 score 0.306 -1.132, 1.745 2.706 0.225 Baseline CMAI factor 3 score NA NA NA NA dimension Adjusted R2 0.652 Independent living living Independent Multiple R2 0.884 Intercept -2.159 -14.462, -2.230 0.268 10.144
Antipsychotic dose change (%) -0.007 -0.080, 0.066 -1.176 0.449 Baseline CMAI factor 2 score 0.276 -1.551, 2.103 1.918 0.306 Baseline CMAI factor 3 score NA NA NA NA dimension Relationships Adjusted R2 0.400 Multiple R2 0.800 Intercept 0.432 -4.259, 5.123 1.169 0.450
Antipsychotic dose change (%) 0.000 -0.028, 0.028 0.043 0.973 Baseline CMAI factor 2 score -0.061 -0.758, 0.635 -1.119 0.464 Baseline CMAI factor 3 score NA NA NA NA dimension
Physical senses Adjusted R2 -0.262 Multiple R2 0.579
Intercept -0.086 -7.175, 7.003 -0.154 0.903 Antipsychotic dose change (%) 0.004 -0.038, 0.046 1.185 0.446 Baseline CMAI factor 2 score 0.023 -1.030, 1.076 0.281 0.826 Baseline CMAI factor 3 score NA NA NA NA
Psychological Adjusted R2 -0.028 wellbeing dimension Multiple R2 0.657 AQoL-4D=Assessment of Quality of Life-4D tool, 95%CI=95% confidence interval
461 | Page Daniel J. Hoyle
Figure 85: Scatterplots showing the relationship between change (%) in the chlorpromazine (CPZ) daily dose equivalent (DDE) and changes in the AQoL-4D (A) utility, and (B) independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing dimension scores. AQoL-4D=Assessment of Quality of Life-4D tool.
Appendix K.3.3 Benzodiazepine users
Thirty-eight residents initially taking benzodiazepines were interviewed about their QoL at baseline and four months. The utility score, and independent living and relationships dimension scores decreased by clinically significant amounts. These changes were statistically significant in the case of the utility and the independent living dimension scores (Table 74 and Figure 86).
On the other hand, physical senses remained relatively the same between baseline and four months.
462 | Page Daniel J. Hoyle Table 74: Mean Assessment of Quality of Life-4D utility and dimension scores at baseline and four months.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Utility 0.27 (0.20, 0.35) 0.21 (0.13, 0.29) 3.052 37 0.004 Independent living 0.41 (0.29, 0.52) 0.32 (0.21, 0.43) 2.027 37 0.050 Relationships 0.72 (0.62, 0.82) 0.65 (0.55, 0.75) 1.270 37 0.212 Physical senses 0.87 (0.83, 0.92) 0.88 (0.84, 0.92) -0.259 37 0.797 Psychological wellbeing 0.82 (0.77, 0.87) 0.78 (0.72, 0.84) 1.261 37 0.215 df=degrees of freedom, 95%CI=95% confidence interval
Figure 86: Beeswarm plots of the AQoL-4D (A) utility, and (B) independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing dimension scores at baseline and four months for all residents taking a benzodiazepine at baseline. The blue dot represents mean and error bars indicate the 95% confidence interval. AQoL- 4D=Assessment of Quality of Life-4D tool.
Using multiple linear regression to control for the CMAI factor 2 baseline score and a documented diagnosis of dementia, we found no evidence of an association between changes in the benzodiazepine dose and the resident-rated AQoL-4D utility score (β=-0.000; i.e., AQoL-
463 | Page Daniel J. Hoyle 4D utility score was 0.00 points higher, on average, for each 10% reduction in the diazepam daily dose equivalent, p=0.747, 95%CI=-0.001, 0.001; Table 75 and Figure 87).
Furthermore, no associations between benzodiazepine dose change and changes in the
AQoL-4D dimension scores were detected except for the psychological wellbeing domain.
Estimates (β=-0.001) from the regression model indicated that the psychological wellbeing score was, on average, 0.01 points higher for each 10% reduction in the diazepam daily dose equivalent (95%CI=-0.002, -0.000, p=0.010).
There was a small, albeit non-significant, decrease in the AQoL-4D utility score between baseline (mean=0.320, 95%CI=-0.007, 0.647) and four months (mean=0.265, 95%CI=-0.100,
0.629) for residents who had their benzodiazepine completely ceased (t(5)=1.27, p=0.260, n=6 residents)
Table 75: Estimated changes in the AQoL-4D utility and dimension scores associated with changes (%) in the diazepam daily dose equivalent (benzodiazepine dose) between baseline and four months.
Outcome Estimate 95%CI t-statistic p-value (β) Intercept -0.104 -0.236, 0.027 -1.616 0.115 Benzodiazepine dose change (%) -0.000 -0.001, 0.001 -0.326 0.747
Documented diagnosis of -0.081 -0.207, 0.045 -1.307 0.200 dementia
Utility Baseline CMAI factor 2 score 0.006 -0.008, 0.021 0.847 0.403
Adjusted R2 -0.008
Multiple R2 0.074
Intercept -0.031 -0.274, 0.212 -0.261 0.795
Benzodiazepine dose change (%) 0.000 -0.001, 0.001 0.099 0.922
Documented diagnosis of -0.303 -0.536, -0.070 -2.645 0.012 dementia Baseline CMAI factor 2 score -0.001 -0.028, 0.026 -0.071 0.944 dimension
2 Independent living living Independent Adjusted R 0.105
Multiple R2 0.178
464 | Page Daniel J. Hoyle Intercept -0.316 -0.645, 0.012 -1.956 0.059
Benzodiazepine dose change (%) 0.000 -0.002, 0.002 0.321 0.750
Documented diagnosis of 0.210 -0.104, 0.525 1.358 0.183 dementia Baseline CMAI factor 2 score 0.027 -0.009, 0.063 1.516 0.139
Adjusted R2 0.035 Relationships dimension Multiple R2 0.114
Intercept 0.019 -0.092, 0.131 0.353 0.726
Benzodiazepine dose change (%) -0.000 -0.001, 0.000 -1.360 0.183
Documented diagnosis of -0.053 -0.159, 0.054 -0.999 0.325 dementia Baseline CMAI factor 2 score -0.002 -0.014, 0.011 -0.262 0.795
Adjusted R2 0.001
Physical senses dimension Multiple R2 0.082
Intercept -0.184 -0.354, -0.014 -2.196 0.035
Benzodiazepine dose change (%) -0.001 -0.002, -0.000 -2.737 0.010
Documented diagnosis of -0.085 -0.248, 0.078 -1.065 0.294 dementia Baseline CMAI factor 2 score 0.017 -0.001, 0.036 1.877 0.069 dimension Adjusted R2 0.141 Psychological wellbeing Multiple R2 0.210
AQoL-4D=Assessment of Quality of Life-4D tool, 95%CI=95% confidence interval
465 | Page Daniel J. Hoyle
Figure 87: Scatterplots showing the relationship between change (%) in the diazepam daily dose equivalent (DDE) and changes in the AQoL-4D (A) utility, and (B) independent living, (C) relationships, (D) physical senses and (E) psychological wellbeing dimension scores. AQoL-4D=Assessment of Quality of Life-4D tool.
466 | Page Daniel J. Hoyle Appendix L Job satisfaction by occupation
Registered nurses
A total of 96 RNs at baseline and 74 RNs at four months completed the adjusted-Measure of
Job Satisfaction survey. No significant differences in demographic data were detected between baseline and four months (Table 76).
Table 76: Demographic data for registered nurses participating in the job satisfaction survey at baseline and four months.
Demographic variable Baseline (n=96) Four months (n=74) p-value % age groups 0.568 <20 years 0.0% (n=0) 0.0% (n=0) 20 to 29 years 20.8% (n=20) 28.4% (n=21) 30 to 39 years 26.0% (n=25) 20.3% (n=15) 40 to 49 years 24.0% (n=23) 17.6% (n=13) 50 to 59 years 16.7% (n=16) 17.6% (n=13 ≥60 years 12.5% (n=12) 14.9% (n=11) unspecified 0.0% (n=0) 1.4% (n=1) % gender 0.678 female 81.3% (n=78) 75.7% (n=56) male 6.3% (n=6) 8.1% (n=6) unspecified 12.5% (n=12) 16.2% (n=12) Years registered; median (range)* 6 (0, 48) 7 (0.5, 40) 0.839 Years of aged care experience; median 5 (0, 40) 6 (0, 40) 0.171 (range)† % of sample that works in a dementia unit 59.7% (n=46) 40.3% (n=31) 0.434 *data available for 87 nurses at baseline and 67 nurses at four months †data available for 96 nurses at baseline and 72 nurses at four months
There was no substantial difference in any of the job satisfaction factor or total scores between baseline and four months (Table 77). Similarly, the distribution of job satisfaction factor and total scores in the beeswarm plots (Figure 88 and Figure 89) did not differ appreciably.
467 | Page Daniel J. Hoyle Table 77: Mean adjusted-Measure of Job Satisfaction factor scores at baseline and four months for registered nurses.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Personal 4.0 (3.8, 4.1) 4.1 (4.0, 4.3) -1.31 162 0.192 Satisfaction Workload 3.3 (3.1, 3.5) 3.5 (3.3, 3.6) -1.41 158 0.162 Team 4.1 (3.9, 4.2) 4.1 (3.9, 4.2) -0.12 162 0.908 Spirit/Co- workers Training 3.7 (3.6, 3.9) 3.7 (3.6, 3.9) 0.30 167 0.768 Professional 4.0 (3.8, 4.2) 4.0 (3.8, 4.2) 0.25 154 0.801 support Total 83.4 (80.6, 86.2) 84.7 (81.6, 87.7) -0.61 161 0.540 df=degrees of freedom, 95%CI=95% confidence interval
Figure 88: Beeswarm plots of the adjusted-Measure of Job Satisfaction factor scores at baseline and four months for registered nurses. The blue dot represents mean and error bars indicate the 95% confidence interval.
468 | Page Daniel J. Hoyle
Figure 89: Beeswarm plot of the adjusted-Measure of Job Satisfaction total scores at baseline and four months for registered nurses. The blue dot represents mean and error bars indicate the 95% confidence interval.
Enrolled nurses
A total of 61 ENs at baseline and 44 ENs at four months completed the adjusted-Measure of Job
Satisfaction survey. No significant differences in demographic data were detected between baseline and four months (Table 78).
Table 78: Demographic data for enrolled nurses participating in the job satisfaction survey at baseline and four months.
Demographic variable Baseline (n=61) Four months (n=44) p-value % age groups 0.340 <20 years 3.3% (n=2) 0.0% (n=0) 20 to 29 years 23.0% (n=14) 15.9% (n=7) 30 to 39 years 16.4% (n=10) 15.9% (n=7) 40 to 49 years 19.7% (n=12) 36.4% (n=16) 50 to 59 years 31.1% (n=19) 22.7% (n=10) ≥60 years 4.9% (n=3) 9.1% (n=4) unspecified 1.6% (n=1) 0.0% (n=0) % gender 0.417 female 88.5% (n=54) 79.5% (n=35) male 4.9% (n=3) 6.8% (n=3) unspecified 6.6% (n=4) 13.6% (n=6) Years enrolled; median (range)* 4 (0, 43) 5 (1, 36) 0.481 Years of aged care experience; median 7 (0, 33) 9 (2, 30) 0.225 (range) % of sample that works in a dementia unit 62.3% (n=33) 37.7% (n=20) 0.382 *data available for 52 nurses at baseline and 40 nurses at four months
469 | Page Daniel J. Hoyle There were no substantial differences in any of the job satisfaction factor or total scores between baseline and four months (Table 79), except for scores for team spirit/co-workers where there was a trend toward lower job satisfaction. Similarly, the distribution of job satisfaction scores besides the team spirit/co-workers and overall job satisfaction scores did not differ appreciably between baseline and four months (Figure 90 and Figure 91). Both overall job satisfaction and team spirit/co-workers tended to be lower at four months.
Table 79: Mean adjusted-Measure of Job Satisfaction factor scores at baseline and four months for enrolled nurses.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Personal 4.1 (3.9, 4.3) 4.1 (3.9, 4.3) 0.17 86 0.863 Satisfaction Workload 3.3 (3.1, 3.4) 3.2 (2.9, 3.4) 0.59 74 0.555 Team 4.2 (4.0, 4.3) 3.9 (3.7, 4.2) 1.79 78 0.078 Spirit/Co- workers Training 3.8 (3.6, 4.0) 3.6 (3.3, 3.8) 1.32 86 0.193 Professional 4.0 (3.8, 4.2) 3.8 (3.6, 4.1) 1.33 89 0.187 support Total 84.5 (81.5, 87.5) 81.4 (76.8, 85.9) 1.25 79 0.253 df=degrees of freedom, 95%CI=95% confidence interval
470 | Page Daniel J. Hoyle
Figure 90: Beeswarm plots of the adjusted-Measure of Job Satisfaction factor scores at baseline and four months for enrolled nurses. The blue dot represents mean and error bars indicate the 95% confidence interval.
471 | Page Daniel J. Hoyle
Figure 91: Beeswarm plot of the adjusted-Measure of Job Satisfaction total scores at baseline and four months for enrolled nurses. The blue dot represents mean and error bars indicate the 95% confidence interval.
Personal Care Assistants
A total of 83 PCAs at baseline and 106 PCAs at four months completed the adjusted-Measure of
Job Satisfaction survey. No significant differences in demographic data were detected between baseline and four months (Table 80).
Table 80: Demographic data for personal care assistants participating in the job satisfaction survey at baseline and four months.
Demographic variable Baseline (n=83) Four months (n=106) p-value % age groups 0.587 <20 years 4.8% (n=4) 2.8% (n=3) 20 to 29 years 20.5% (n=17) 14.2% (n=15) 30 to 39 years 18.1% (n=15) 23.6% (n=25) 40 to 49 years 18.1% (n=15) 23.6% (n=25) 50 to 59 years 27.7% (n=23) 25.5% (n=27) ≥60 years 6.0% (n=5) 8.5% (n=9) unspecified 4.8% (n=4) 1.9% (n=2) % gender 0.358 female 73.5% (n=61) 69.8% (n=74) male 15.7% (n=13) 12.3% (n=13) unspecified 10.8% (n=9) 17.9% (n=19) Years of aged care experience; median 5 (0, 25) 6 (0, 35) 0.221 (range) † % of sample that works in a dementia unit 44.2% (n=42) 55.8% (n=53) 0.934 *data available for 79 personal care assistants at baseline and 105 personal care assistants at four months
472 | Page Daniel J. Hoyle There were no substantial differences in any of the job satisfaction factor or total scores between baseline and four months (Table 81). Similarly, the distribution of job satisfaction factor and total scores in the beeswarm plots (Figure 92 and Figure 93) did not differ appreciably.
Table 81: Mean adjusted-Measure of Job Satisfaction factor scores at baseline and four months for personal care assistants.
Outcome Mean score at Mean score at four t-statistic df p-value baseline (95%CI) months (95%CI) Personal 4.1 (4.0, 4.2) 4.1 (4.0, 4.2) -0.40 187 0.692 Satisfaction Workload 3.3 (3.1, 3.4) 3.3 (3.2, 3.5) -0.45 186 0.656 Team 4.1 (4.0, 4.2) 4.1 (4.0, 4.2) -0.35 186 0.730 Spirit/Co- workers Training 3.7 (3.6, 3.9) 3.8 (3.7, 3.9) -0.49 183 0.625 Professional 3.9 (3.7, 4.0) 3.8 (3.7, 4.0) 0.36 182 0.722 support Total 83.3 (81.1, 85.6) 83.8 (81.5, 86.2) -0.29 187 0.773 df=degrees of freedom, 95%CI=95% confidence interval
473 | Page Daniel J. Hoyle
Figure 92: Beeswarm plots of the adjusted-Measure of Job Satisfaction factor scores at baseline and four months for personal care assistants. The blue dot represents mean and error bars indicate the 95% confidence interval.
474 | Page Daniel J. Hoyle
Figure 93: Beeswarm plot of the adjusted-Measure of Job Satisfaction total scores at baseline and four months for personal care assistants. The blue dot represents mean and error bars indicate the 95% confidence interval.
475 | Page Daniel J. Hoyle