Download Date 05/10/2021 14:45:35

Medicines Reconciliation: Roles and Process. An examination of the medicines reconciliation process and the involvement of patients and healthcare professionals across a regional healthcare economy, within the United Kingdom.

Item Type Thesis

Authors Urban, Rachel L.

Rights
The University of Bradford theses are licenced under a Creative Commons Licence.

Download date 05/10/2021 14:45:35

Link to Item http://hdl.handle.net/10454/7288

University of Bradford eThesis

This thesis is hosted in Bradford Scholars – The University of Bradford Open Access repository. Visit the repository for full metadata or to contact the repository team

MEDICINES RECONCILIATION: ROLES AND PROCESS

An examination of the medicines reconciliation process and the involvement of patients and healthcare professionals across a regional healthcare economy, within the United Kingdom

Volume 1

Rachel Louise URBAN

submitted for the degree of Doctor of Philosophy

School of Pharmacy

University of Bradford

2014

ABSTRACT Medicines Reconciliation: Roles and Process Rachel Urban

Keywords Medicines Reconciliation; Hospital Admission; Medication History Taking; Hospital Discharge; Health Care Professionals; Patient Involvement; Patient Safety; Human Factors; Medication Safety; Care Transitions

Medication safety and improving communication at care transitions are an international priority. There is vast evidence on the scale of error associated with medicines reconciliation and some evidence of successful interventions to improve reconciliation. However, there is insufficient evidence on the factors that contribute towards medication error at transitions, or the roles of those involved. This thesis examined current UK medicines reconciliation practice within primary and secondary care, and the role of HCPs and patients. Using a mixed-method, multi-centre design, the type and severity of discrepancies at admission to hospital were established and staff undertaking medicines reconciliation across secondary and primary care were observed, using evidence-informed framework, based on a narrative literature review.

The overall processes used to reconcile medicines were similar; however, there was considerable inter and intra-organisational variation within primary and secondary care practice. Patients were not routinely involved in discussions about their medication, despite their capacity to do so. Various human factors in reconciliation-related errors were apparent; predominantly inadequate communication, individual factors e.g. variation in approach by HCP, and patient factors e.g. lack of capacity. Areas of good practice which could reduce medicines reconciliation-related errors/discrepancies were identified. There is a need for increased consistency and standardisation of medicines reconciliation- related policy, procedures and documentation, alongside communication optimisation. This could be achieved through a standardised definition and taxonomy of error, the development of a medicines reconciliation quality assessment framework, increased undergraduate and post-graduate education, improved patient engagement, better utilisation of information technology and improved safety culture.

ACKNOWLEDGEMENTS

‘It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness, it was the epoch of belief, it was the epoch of incredulity, it was the season of light, it was the season of darkness, it was the spring of hope, it was the winter of despair…’

A Tale of Two Cities, Charles Dickens

My PhD has felt like an incredible journey through which I have learned a great deal, especially about myself. This is the result of the fantastic people which I have had supporting me. I have had the most knowledgeable, kind and caring supervisors I could have ever wished for - Julie, Gerry, Alison and Kay. (A special thank you to Julie for being a great friend who has had confidence in me since my undergraduate days and provided me with many opportunities throughout my professional career.) Thanks also to David Alldred for keeping me realistic and grounded, and sharing the taxonomy of medication discrepancies developed by the Universities of Sydney, Leeds and Wisconsin. I must also thank the rest of the Pharmacy Department at the University of Bradford, for your help and support, for keeping me sane with hugs and chocolate when I needed it the most; you are a special group of people and I feel privileged to have spent four years with you. I often affectionately refer to you as ‘brothers and sisters’.

It would not have been achieved without the participation of the hospitals, general medical practices, HCPs and support staff involved, for which I am grateful. I would additionally like to thank Bradford Institute for Health Research (BIHR) and Bradford Teaching Hospitals NHS Foundation Trust (BTHFT) for funding the studentship, especially David Smith, Neill McDonald and the rest of my team, who have supported me throughout. Likewise, thanks to Community Pharmacy West Yorkshire for providing a role through which I can continue to research.

I have the most amazing group of friends anyone could ever wish for. They have shared my ‘ups and downs’ and encouraged me over the last four years. Most specifically (in alphabetical order) Gemma, Ingrid, Jill, Rachel, Ruth, Sue

ii and Zoe. Also, Nicola, Ashleigh, Steve and the rest of North Leeds Dance Academy, for helping me win my first competition during the most difficult stage of my PhD and showing an interest, week on week, by asking ‘Isn’t it done yet?’. I must also thank Jan who has been with me every step of my journey. I appreciate everything you do for me.

To Oz, for always believing in me and teaching me that ‘Verily, after hardship comes ease’ (Quran 94:5), which I have used as my mantra throughout.

To Chris, for being patient, caring for me and providing excellent company, friendship and meals when they were most needed.

To my Mum for painstakingly proof reading (especially my references), my Dad for just being ‘Dad’ and my little sister Sarah, I love you all lots. Lastly, to Grandma for providing my brains and Grandad, my biggest champion; this is for you.

iii

TABLE OF CONTENTS

VOLUME 1

Abstract ...... I

Acknowledgements ...... ii

Table of Figures ...... xii

Table of Tables ...... xiv

Glossary of Terms ...... xvii

Abbreviations ...... xxv

CHAPTER 1 INTRODUCTION ...... 1

1.1 Background ...... 1

1.2 Patient Safety ...... 2 1.2.1 Human factors ...... 4

1.3 Medication Safety ...... 7

1.4 Medicines Reconciliation ...... 8 1.4.1 Measuring error associated with medicines reconciliation ...... 10

1.5 Rationale for this Study ...... 13

1.6 Thesis Outline ...... 16

CHAPTER 2 LITERATURE REVIEW ...... 18

2.1 Background ...... 18

2.2 Methods ...... 18

Findings ...... 21

2.3 Part 1 - Definition, Scale and Severity of Discrepancies Found ...... 22 2.3.1 UK and international context ...... 22 2.3.2 Definition of medicines reconciliation ...... 25 2.3.3 Models of medicines reconciliation ...... 30 2.3.4 Classification of medicines reconciliation-related error ...... 33 2.3.5 Scale of medication discrepancies at transitions ...... 33 2.3.6 Severity of ADEs and pADEs related to medicines reconciliation35 2.3.7 Risk factors for experiencing medication discrepancy at transition points ...... 35 2.3.8 Time taken to complete medicines reconciliation ...... 39 iv

2.3.9 Sources of information used during medicines reconciliation .... 39 2.3.10 Impact of medicines reconciliation on readmissions to hospital ...... 45 2.3.11 Improving patient safety and measuring quality ...... 46 2.3.12 Human Factors and their role in Medicines Reconciliation ...... 48 2.3.13 Summary Part 1 ...... 48

2.4 Part 2 - Medicines Reconciliation Process and HCP Involvement ... 50 2.4.1 The medicines reconciliation process ...... 51 2.4.2 Admission to hospital ...... 54 2.4.2.1 Background ...... 54 2.4.2.2 The role of the doctor at admission to hospital ...... 58 Hospital doctor ...... 58 General practitioner (GP) ...... 59 2.4.2.3 The role of hospital pharmacy staff at admission to hospital ...... 60 2.4.2.4 The role of nurses at admission to hospital ...... 62 2.4.2.5 The role of the community pharmacist at admission to hospital ...... 62 2.4.2.6 The role of ambulance staff at admission to hospital ..... 63 2.4.3 During the patient’s stay ...... 63 2.4.3.1 The role of the hospital doctor during the patient’s stay in hospital ...... 63 2.4.3.2 The role of pharmacy staff during the patient’s stay in hospital ...... 64 2.4.3.3 The role of the nurse during the patient’s stay in hospital ...... 65 2.4.4 At discharge from hospital ...... 65 2.4.4.1 The role of the hospital doctor at discharge from hospital ...... 69 2.4.4.2 The role of the pharmacist at discharge from hospital ... 69 2.4.4.3 The role of the nurse at discharge from hospital ...... 73 2.4.4.4 The role of the community pharmacist at discharge ...... 74 2.4.5 Post-discharge (in primary care) ...... 75 2.4.5.1 The role of the GP post-discharge ...... 75 2.4.5.2 The role of the community pharmacist post-discharge .. 77 2.4.5.3. Role of other staff within primary care ...... 77 2.4.6 Standardisation of the medicines reconciliation process ...... 78 2.4.6.1 Minimum data set for medicines information provided at discharge ...... 79 2.4.7 The role of information technology (IT) ...... 80 At Admission ...... 82 During Stay ...... 84 At Discharge ...... 84 2.4.8 Summary Part 2 ...... 84

2.5 Part 3 The Role of the Patient in Medicines Reconciliation ...... 86 2.5.1 Overview ...... 86 2.5.2 The potential of collaborating with the patient and/or patient representatives to reconcile medication ...... 86 2.5.3 Potential limitations of collaborating with the patient to reconcile medication ...... 88

2.5.4 Role of patient representatives/caregivers in the provision of information ...... 90 2.5.5 Collaborative interventions which support accurate medicines reconciliation ...... 91 2.5.5.1 Bringing medication into hospital ...... 91 2.5.5.2 Self-administration ...... 92 2.5.5.3 Communication and patient counselling ...... 92 2.5.5.4 Hand-held information ...... 94 2.5.5.5 Self-completed questionnaires ...... 94 2.5.5.6 Access to personal electronic health records (EHR) ..... 95 2.5.6 Summary Part 3 ...... 95

2.6 Overall Summary Chapter 2 ...... 96

CHAPTER 3 METHODOLOGY ...... 97

3.1 Introduction ...... 97

3.2 Rationale for Methodology ...... 98 3.2.1 Theoretical perspectives ...... 98 3.2.2 Justifying mixed-methods ...... 100 3.2.3 Exploring quantitative and qualitative approaches ...... 102 3.2.3.1 Quantitative approaches ...... 102 3.2.3.2 Qualitative approaches ...... 103

3.3 Data Collection Methods ...... 105 3.3.1 Survey methods ...... 105 3.3.2 Observation ...... 105 3.3.2.1 Deduction and induction ...... 107 Deductive approach ...... 108 Inductive approach ...... 109 3.3.2.2 Participant and non-participant observation ...... 110 3.3.3 Possible alternative methods ...... 111

3.4 Recording Data ...... 111

3.5 Ensuring Rigour - Concepts of Reliability and Validity ...... 112 3.5.1 Reflexivity ...... 116 3.5.2 Inter-rater reliability and minimising observer bias ...... 116

3.6 Sampling and Sample Size ...... 117 3.6.1 Quantitative sampling ...... 117 3.6.2 Qualitative sampling ...... 117 3.6.3 Patient specialties observed during the study ...... 118

3.7 Access and Recruitment ...... 119 3.7.1 Relationships in the research setting ...... 120 3.7.2 Informed consent ...... 120

3.8 Data Analysis ...... 120 3.8.1 Quantitative data analysis ...... 121 3.8.2 Qualitative data analysis ...... 121

3.9 Ethical Considerations ...... 124

CHAPTER 4 STUDY 1: OBSERVATIONAL SURVEY OF MEDICINES RECONCILIATION FOLLOWING ADMISSION TO SECONDARY CARE .... 127

4.1 Introduction ...... 127

4.2 Aims and Objectives ...... 128 4.2.1 Aim ...... 128 4.2.2 Objectives ...... 128

4.3 Method ...... 128 4.3.1 Participants ...... 128 4.3.2 Data collection ...... 129 4.3.2.1 Development and piloting of data collection proforma . 129 4.3.2.2 Collecting the data ...... 129 4.3.3 Classifying severity ...... 130 4.3.4 Sample size estimation and data analysis ...... 132

4.4 Findings ...... 132 4.4.1 Length of time before medicines reconciliation is undertaken post- admission ...... 132 4.4.2 Time taken to undertake medicines reconciliation ...... 134 4.4.3 Sources used to conduct medicines reconciliation ...... 135 4.4.4 Number and type of discrepancies identified ...... 138 4.4.5 Severity of discrepancies identified ...... 141

4.5 Discussion ...... 143 4.5.1 Time taken to complete medicines reconciliation ...... 143 4.5.2 Sources of information used in medicines reconciliation ...... 144 4.5.3 Patient involvement in medicines reconciliation ...... 145 4.5.4 Medication discrepancies identified ...... 146 4.5.5 Severity of discrepancies identified ...... 147 4.5.6 Strengths and limitations of Study 1 ...... 147

4.6 Conclusion ...... 148

CHAPTER 5 STUDY 2: OBSERVATION OF MEDICINES RECONCILIATION IN SECONDARY CARE...... 150

5.1 Introduction ...... 150

5.2 Aims and Objectives ...... 151

5.3 Method ...... 151 5.3.1 Overview ...... 151 5.3.2 Sample ...... 152 5.3.3 Access to hospitals and recruitment of healthcare professionals (HCPs) ...... 152 5.3.3.1 Participants ...... 153 5.3.4 Observation schedule development and pilot ...... 153 5.3.4.1 Deductive approach ...... 153

vii

5.3.4.2 Inductive approach ...... 154 5.3.5 Observation process ...... 155 5.3.6 Recording of data ...... 157 5.3.7 Data analysis ...... 157 5.3.7.1 Communication systems - Verbal and written information analysis ...... 160 5.3.8 Validity and reliability of methods used ...... 166 5.3.9 Ethical considerations for Study 2 ...... 168

5.4 Results ...... 169 5.4.1 Overview: hospitals and participants ...... 169 5.4.1.1 Organisational information ...... 170 5.4.2 Analysis using secondary care observation schedule (SCOS) .. 170 5.4.2.1 Organisational policy ...... 170 5.4.2.2 Training and education ...... 171 5.4.2.3 Audit ...... 174 5.4.2.4 Overview of medicines reconciliation process ...... 174 5.4.2.5 Taking a medication history ...... 177 Sources ...... 177 5.4.2.6 Prescribing medication ...... 180 5.4.2.7 Verification of the medication history by pharmacy staff ...... 181 5.4.2.8 Transfer of patient care ...... 182 5.4.2.9 Ward round ...... 182 Stop, starting, changing and withholding medication .. 182 5.4.2.10 Medication administration ...... 184 5.4.2.11 Discharge ...... 185 5.4.2.12 Counselling patients ...... 187 5.4.3 Analysis using Yorkshire Contributory Factors Framework (YCFF) ...... 188 5.4.3.1 Active failures – Mistakes, lapses and violations ...... 190 5.4.3.2 Situational factors ...... 191 a) Individual factors ...... 191 b) Team factors ...... 197 c) Patient factors ...... 198 d) Task characteristics ...... 201 5.4.3.3 Local working conditions ...... 202 a) Lines of responsibility ...... 202 b) Staff workload ...... 203 c) Supervision and leadership ...... 204 d) Management of staff and staffing levels ...... 205 e) Equipment and supplies ...... 207 5.4.3.4 Latent organisational factors ...... 210 a) Training and education ...... 210 b) Support from central functions ...... 214 c) Policies and procedures ...... 215 d) Physical environment ...... 216 e) Scheduling and bed management...... 219 5.4.3.5 Latent external factors ...... 220 a) External policy content ...... 220 b) Design of equipment and supplies ...... 222 5.4.3.6 Safety culture ...... 223 viii

5.4.3.7 Communication systems ...... 227 5.4.3.7.1 HCP to patient interaction – the consultation .... 227 5.4.3.7.2 HCP to HCP communication ...... 233 5.4.4 Errors and discrepancies observed ...... 238 5.4.5 Researcher interventions ...... 244

5.5 Discussion ...... 244 5.5.1 Policy, process and practice ...... 244 5.5.2 The impact of human factors ...... 246 Individual and Team Factors ...... 246 Patient Factors and Patient Involvement ...... 252 Training and Education ...... 252 IT ...... 252 Safety Culture ...... 252 5.5.3 Communication ...... 253 5.5.4 Error and discrepancy ...... 254 Outliers ...... 255 5.5.5 Frameworks...... 255 5.5.6 Strengths and limitations of Study 2 ...... 257 Researcher intervention ...... 257 Co-rating ...... 258

5.6 Conclusion of Study 2 ...... 259

CHAPTER 6 STUDY 3: FINDINGS FROM PRIMARY CARE OBSERVATION ...... 260

6.1 Introduction ...... 260

6.2 Aims and Objectives ...... 262

6.3 Method ...... 262 6.3.1 Overview ...... 262 6.3.2 Access to general medical practices and recruitment of staff .... 262 6.3.3 Ethical considerations in Study 3...... 263 6.3.4 Development of primary care observation schedule (PCOS) .... 263 6.3.5 Data collection ...... 264 6.3.6 Data analysis ...... 265

6.4 Results ...... 265 6.4.1 Demographics of general medical practices ...... 265 6.4.2 Policy ...... 265 6.4.3 Training on medicines reconciliation related processes ...... 267 6.4.4 Audit ...... 267 6.4.5 Receipt of information ...... 269 6.4.6 Scanning of information ...... 271 6.4.7 Dissemination of information received ...... 272 6.4.8 Review of information ...... 273 Who ...... 273 How ...... 273 Error ...... 275 Clarification of information ...... 277

6.4.9 Amendment ...... 278 6.4.10 Patient information ...... 279 6.4.11 Repeat prescribing ...... 279

6.5 Discussion ...... 280 6.5.1 Overall process ...... 280 6.5.2 Review of information ...... 281 6.5.3 Discrepancies identified ...... 283 6.5.4 Involvement of the patient in medicines reconciliation ...... 283 6.5.5 Role of the hospital ...... 284 6.5.6 Strengths and Limitations of Study 3 and overall observation ... 285 Observer bias ...... 285 Data recording ...... 285 Reflexivity ...... 286 Framework ...... 286 6.5.7 Conclusion of study 3...... 287

CHAPTER 7 GENERAL DISCUSSION AND CONCLUSIONS ...... 288

7.1 Definition and Taxonomy of Medicines Reconciliation ...... 288 7.1.1 Defining Medication Reconciliation Error ...... 290

7.2 Scale and Severity of Medicines Reconciliation Related Error and Discrepancy ...... 291

7.3 Sources Used to Conduct Medicines Reconciliation ...... 293

7.4 Patient Safety and Human Factors ...... 293

7.5 Process ...... 295 7.5.1 Overview ...... 295 Systems, processes and procedures ...... 295 Healthcare professionals role in process ...... 297 7.5.2 At admission ...... 298 Doctors ...... 298 Pharmacy staff ...... 299 Other staff ...... 300 7.5.3 During stay ...... 301 Nursing staff ...... 301 Doctors ...... 302 7.5.4 At discharge ...... 302 7.5.5 In primary care ...... 303 Community pharmacy ...... 304 7.5.6 Closing the loop ...... 304

7.6 Strategies for Improvement ...... 304 7.6.1 Patient involvement in medicines reconciliation ...... 305 7.6.2 Role of information technology (IT) in medicines reconciliation . 307 7.6.3 Standardisation of medicines reconciliation-related processes . 309 7.6.4 Optimising communication ...... 311 7.6.5 Training and education ...... 314 7.6.6 Improving organisational safety culture ...... 317

7.7 Recommendations for Practice, Policy and Theory ...... 319

7.8 Further Research ...... 322

7.9 Strengths and Limitations of the Overall Study ...... 324

7.9 Conclusion ...... 325

REFERENCES ...... 328

VOLUME 2

APPENDICES ...... 382 APPENDIX A LREC APPROVAL FOR OBSERVATION OF HCPS ...... 383

APPENDIX B NHS BRADFORD AND AIREDALE RESEARCH AND DEVELOPMENT APPROVAL LETTER ...... 386

APPENDIX C CALDERDALE AND HUDDERSFIELD NHS FOUNDATION TRUST RESEARCH AND DEVELOPMENT APPROVAL LETTER ...... 388

APPENDIX D BRADFORD TEACHING HOSPITALS NHS FOUNDATION TRUST RESEARCH AND DEVELOPMENT APPROVAL LETTER ...... 390

APPENDIX E AIREDALE NHS FOUNDATION TRUST RESEARCH AND DEVELOPMENT APPROVAL LETTER ...... 393

APPENDIX F THE LEEDS TEACHING HOSPITALS NHS TRUST RESEARCH AND DEVELOPMENT APPROVAL LETTER ...... 395

APPENDIX G LREC WITHDRAWAL OF INVOLVEMENT LETTER ...... 400

APPENDIX H CHAPTER 4 FINDINGS PUBLISHED IN RESEARCH IN SOCIAL & ADMINISTRATIVE PHARMACY ...... 401

APPENDIX I AUDIT TOOL AND GUIDANCE ...... 415

APPENDIX J CONSENT INFORMATION AND CONSENT FORMS FOR OBSERVATIONAL STUDIES ...... 430

APPENDIX K SECONDARY CARE OBSERVATION SCHEDULE (SCOS) ... 434

APPENDIX L EXAMPLE EMAIL REQUEST SENT TO ORGANISATIONS REQUESTING PROCEDURAL DOCUMENTS ASSOCIATED WITH MEDICINES RECONCILIATION ...... 475

APPENDIX M LETTER TO GP PRACTICES INVITING THEIR PARTICIPATION IN THE PRIMARY CARE OBSERVATION STUDY ...... 476

APPENDIX N PRIMARY CARE OBSERVATION SCHEDULE (PCOS) ...... 478

TABLE OF FIGURES

Figure Title Page 1.1 Stages in the development of an organisational accident 6 1.2 The Swiss cheese model of how defences, barriers, and 6 safeguards may be penetrated by an accident trajectory 1.3 Relationship between medication error and adverse drug 11 events (ADEs) according to Aronsen et al 1.4 Relationship between medication error and adverse drug 11 events (ADEs) according to Morimoto et al 1.5 Relationship between discrepancy and adverse drug event 13 2.1 Flow diagram of search and retrieval process 21 2.2 Process map of verification of medication history 32 2.3 Algorithm for medicines reconciliation from admission through 52 discharge Discharge 2.4 Diagram adapted from the WHO High 5s campaign by 53 Fernandes et al 2.5 Workflow for medicines reconciliation in A&E 55 2.6 Elements of transitional care models 67 3.1 Sequential study design 98 3.2 Relationship between epistemology, theoretical perspectives 99 methodology and research methods. Gray adapted from Crotty 3.3 Deductive reasoning 108 3.4 Inductive reasoning 110 3.5 Dey’s model of qualitative analysis 122 4.1 Percentage patients with medication history verified by a 134 pharmacist 4.2 Number of sources used to reconcile medication within each 136 hospital 4.3 Type of source used to reconcile medication in each hospital 137 4.4 Number of discrepancies experienced per patient within each 139 hospital

xii

Figure Title Page 4.5 Percentage of patients experiencing at least one discrepancy 139 on their inpatient medication chart (IMC) 4.6 Categorisation of severity of discrepancies 141 4.7 Omissions categorised by drug group 142 5.1 The Yorkshire contributory factors framework 158 5.2 Berlos’s SMCR Model of communication 166 5.3 Process of ensuring validity and reliability within the 168 observation data 5.4 Patient medication journey 176 5.5 Percentage of medications omitted from inpatient medication 240 charts (IMC) within Hospitals A-D 5.6 Comparison between type of medication omitted during 243 prospective data collection and observational study 6.1 Information processing model 261 7.1 Gerbner’s General Model 313

xiii

TABLE OF TABLES

Table Title Page

1.1 Errors of action and planning: summary definitions (originally) 7 modified from reason 1990 , P.9 by Armitage

2.1 Medicines reconciliation studies by country 23

2.2 Definitions of ‘medication reconciliation’ used by government 27 and professional organisations

2.3 Definitions of ‘medication reconciliation’ used in published 29 papers

2.4 Models showing the stages of medicines reconciliation 31

2.5 Time taken to conduct medicines reconciliation per patient in 41 hospital

2.6 Summary of interventions relating to the patient’s admission to 56 hospital categorised by HCP and including those interventions carried out by the whole team and the role of the patient

2.7 Interventions which lead to successful transitions in care 66

2.8 Summary of interventions relating to the patient’s discharge 70 from hospital categorised by HCP and including those interventions carried out by the whole team and the role of the patient

2.9 Summary of Information Technology (IT) interventions relating 80 to the different stages of the patient’s journey through hospital

3.1 The application of quantitative and qualitative methods within 104 the study and their relevance – Adapted from Bryman and Gray 3.2 Seven stages in the deduction process applied to this study 109

xiv

Table Title Page

3.3 Types of triangulation described by Denzin (1978) and their 115 application to this study

3.4 Six phases of thematic analysis 123

3.5 Summary of ethics approval for study 126

4.1 Hospital demographics 133

4.2 Descriptive statistics for total time taken to conduct medicines 135 reconciliation (minutes)

4.3 Descriptive statistics for actual time taken to conduct medicines 135 reconciliation (minutes)

4.4 Other sources used to reconcile medication across all four 137 hospitals

4.5 Number of discrepancies identified and categorised by type 140

5.1 Yorkshire Contributory Factors Framework (YCFF) criterion 159 definitions

5.2 Communication during medicines reconciliation 162

5.3 COT assessment criteria adapted for medicines reconciliation 163

5.4 Essential communication skills for a medication history 164 Interview

5.5 Breadth of doctor interactions outside the consultation room 165 (adapted from Thistlethwaite and Spencer)

5.6 Summary of co-rater characteristics 167

5.7 Number and type of participants 170

5.8 Content summary of hospital medicines reconciliation policies 172

Table Title Page

5.9 Reasons for stopping, starting, changing or withholding 183

5.10 Inter-rater reliability of categorisation using YCFF categories 189

5.11 Staff allocation on Medical Admissions Unit (MAU), Hospital D 206

5.12 Level of ward technician training 214

5.13 Summary of ward layout in each hospital 217

5.14 Number of discrepancies identified and categorised by type in 239 all hospitals

5.15 Comparison of the types of discrepancies determined during 242 the prospective observational survey and the observational study

5.16 Conscious competence learning model 247

5.17 Summary of factors which contributed towards or prevented 249 medicines reconciliation related discrepancy or error

6.1 Practice demographics 266

6.2 Number of participants in study categorised by role 266

6.3 Content of policies and protocols for processing 268 correspondence within general medical practice

7.1 At risk behaviours in medication processes relating to 296 medicines reconciliation

7.2 Overview of when training may or may not be appropriate as a 316 human factor approach to patient safety

7.3 Medicines reconciliation in varying levels of intensity as seen in 322 published studies

xvi

GLOSSARY OF TERMS

Active Failure An active failure is an unsafe act (error or violation) committed by those with direct contact with the patient.

Addition (also known as commission) The accidental prescription of a medication to the patient’s current treatment regimen following inaccurate medication history taking.

Adverse Drug Event (ADE) An injury resulting from the use of a drug.

Caldicott Guardian A senior person responsible for protecting the confidentiality of patient and service-user information and enabling appropriate information-sharing within an organisation.

Care Home A home for people who need continuous medical treatment or who are unable to look after themselves.

Care Setting A setting where people receive care. It may include hospitals, nursing homes, residential homes, a patient’s own home or intermediate care.

Care Transition The movement of patients between healthcare locations, providers or different levels of care within the same location, as their condition and care needs change.

Clerking To take a full medical history, perform a physical examination, record the findings in the patient’s notes, then write a care plan for the patient.

xvii

Clinical Check A clinical review of the inpatient medication chart (IMC) to determine problems such as over and under dosing, interactions and missing or innappropriate information such as the form, strength or dosage of the medication. The clinical check can be conducted to varying degrees. A basic chart review would usually be conducted without the use of extra information such as the presence of the clinical notes, the patient or test results compared with a full chart review where these items would be present.

Clinical Commissioning Groups (CCGs) NHS organisations set up by the Health and Social Care Act 2012 to organise the delivery of NHS services in England. They replace some of the functions of primary care trusts (PCTs), which ceased to exist in April 2013.

Commissioning for Quality and Innovation (CQUIN) A framework to support improvements in the quality of services and the creation of new, improved patterns of care, within the NHS.

Complementary and Alternative Medicines (CAMs) Non-conventional approaches to healing, beyond traditional medicine. Complementary medicine is any form of therapy used in combination with other alternative treatments or standard/conventional medicine. Alternative medicine is used alone without recommended standard treatment.

Counsel To provide the patient with information and advice on how to take/use their medication.

Core Training Year 1 (CT1) & Core Training Year 2 (CT2) Following two years of foundation year training, doctors who wish to specialise within hospital may progress on to a further two years of generic training before entering their specialty.

Delphi Approach A process for merging opinion on a specific topic or issue. xviii

Discharge Summary A clinical report prepared by a physician or other health professional at the end of a hospital stay. It outlines the patient's reason for admission, the diagnostic findings, the treatment received, and recommendations on discharge including medication.

E-discharge Summary An electronically produced discharge summary. In some instances this is transmitted electronically to the recipient.

Electronic Health Record (EHR) An electronic version of a patient’s medical history, maintained by a provider or several providers over time. It includes clinical data relevant to the persons healthcare, including demographics, progress notes, consultations, medications, immunisations and laboratory reports.

Executive Walk Round Visits by hospital executives to patient care areas to discuss patient safety issues with providers.

Foundation Year 1 (FY1) & Foundation Year 2 (FY2) The two-year, general postgraduate medical training programme, which bridges medical school and specialist/general practice training. It is compulsory for all newly qualified medical practitioners in the UK.

Grandparenting The ‘grandparenting’ arrangements for pharmacy technicians were put in place to ease the transition to regulation for pharmacy technicians. Up until midnight on the 30th June 2011 an individual with the relevant work experience and qualifications could apply to register as a pharmacy technician without undertaking specific technician training.

Guidelines A rule, principle or recommendation of good practice, which is often based on evidence obtained from research and/or from a broad body of experience from xix multiple contributors, often a body of experts in a particular discipline e.g. a professional body.

Handover The transfer of responsibility and accountability between one or more clinicians on a temporary or permanent basis.

Healthcare Professional (HCP) A person who by education, training, or certification is qualified to and is engaged in providing healthcare, for example, a doctor, pharmacist, nurse.

High Reliability Organisation (HRO) An organisation that has succeeded in avoiding catastrophes in an environment where normal accidents can be expected due to risk factors and complexity.

Inpatient Medication Chart (IMC) An A4-sized document/ booklet which provides a running log of the medications ordered by a prescriber. It is usually kept at the bedside of an inpatient or retained with the patient’s medical records at the nursing station.

Intentional round To implement hourly nursing rounds, to check on patients and ensure their fundamental care needs are met.

Interdisciplinary Rounding The planning and evaluation of patient care with health professionals from a variety of other health disciplines.

Intermediate Care The term used for a range of services that deliver short term care, support or rehabilitation to patients. It bridges the gap between hospital and home, helping patients recover more quickly, so they can cope in their own home, preventing either unnecessary admission to hospital or allowing earlier discharge to patients who are medically fit but need some rehabilitation.

Latent Failure A latent failure is an organisational or managerial failure for example lack of policy and/or procedure.

Lean Methodology Lean is an improvement approach to improve flow and eliminate waste

Liverpool Care pathway (LCP) An integrated care pathway recommended as a best practice model by the UK Department of Health, used in the last hours and days of life. It is to be replaced this year with the ‘End of Life Care Plan’.

Medication Error A failure in the treatment process that leads to, or has the potential to lead to, harm to the patient.

Medication formulary A list of drugs which is evidence based and cost-effective that enables health practitioners to prescribe treatment that is medically appropriate.

Medicines Use Review (MUR) An advanced service offered by pharmacies in the United Kingdom. It is part of the contract pharmacies hold with the National Health Service (NHS). The patient is provided with an opportunity to discuss their medicines use with a qualified pharmacist.

National Institute for Health and Care Excellence (NICE) Provides national guidance and advice to improve health and social care in the UK.

National Patient Safety Agency (NPSA) An Arm’s Length Body of the Department of Health which improves safe patient care by informing, supporting and influencing organisations and people working in the health sector in the UK.

xxi

Outlier Hospital inpatients who are allocated a bed on a hospital ward which is not clinically appropriate for their illness due to a lack of inpatient beds. Also commonly named ‘sleep-outs’ or ‘boarders’.

Over the Counter (OTC) medication Medicines sold directly to a patient without a prescription, usually from a community pharmacy.

Potential Adverse Drug Event (PADE) A ‘serious error or mishap that has the potential to cause an adverse event but fails to do so because of chance or because it is intercepted (also called ‘near miss’ or ‘close call’)’.

Primary Care Primary care is the local healthcare received from GPs, NHS-walk-in centres, dentists, pharmacists and optometrists in the UK. They were provided and managed by PCTs up until April 2013. They are now managed by NHS England.

Primary Care Trust Primary care trusts (PCTs) were, up until April 2013, the statutory bodies in England responsible for ensuring NHS services were available in a defined geographical area and for improving the health of people living in that area. They have been superseded by NHS England.

Process A sequence of interdependent and linked procedures which, at every stage, consume one or more resources (e.g. employee time, energy, machines, money) to convert inputs (data, material, parts, etc.) into outputs. These outputs then serve as inputs for the next stage until a known goal or end result is reached.

Procedure A series of steps taken to accomplish an end.

xxii

Protocol The accepted or established code of procedure or behaviour in any group, organisation, or situation.

Policy A statement of intent by an organisation to follow a particular course of action.

Post-take ward round (PTWR) The first ward round where patients are seen by doctors following emergency admission.

Read Codes A catalogue of codes which represent clinical terms for example diagnoses and investigations. They are used to record patient findings and procedures in UK health and social care Information Technology (IT) systems across primary and secondary care e.g. General Practice and pathology reporting. The codes enable efficient electronic communication and support activity reporting, payments, audit, and research.

SBAR (Situation, Background, Assessment, Recommendation) A structured method for communicating critical information that requires immediate attention and action contributing to effective escalation and increased patient safety.

Secondary Care Secondary (or 'acute') care is the healthcare received in hospital. It may be unplanned emergency care or surgery, or planned specialist medical care or surgery.

Shared Care A model of integrated healthcare delivery where practitioners of different disciplines or with different skills and knowledge collaborate to allow the most appropriate healthcare practitioner to deliver patient care.

xxiii

Six Sigma A service improvement technique which seeks to improve the quality of process outputs by identifying and removing the causes of errors and minimise variability in processes.

Specialist Registrar (SpR) A doctor who has completed their foundation and core training and has begun to specialise

Specials 'Specials' are special-order unlicensed medicines made to meet the needs of an individual patient.

System A set of principles or procedures according to which something is done; an organised scheme or method.

SystmOne An electronic health record (EHR) predominantly used in primary care, mainly general medical practices. Its unique feature is that it allows sharing of the patient’s EHR between HCPs and organisations.

Workforce The people engaged in or available for work, either in a country or area or in a particular firm or industry.

xxiv

ABBREVIATIONS

A&E Accident and Emergency

ACE Angiotensin Converting Enzyme

ADE Adverse Drug Event

ADR Adverse Drug Reaction

AHRQ Agency for Healthcare Research and Quality

APhA American Pharmacist Association

APHID Automated Patient History Intake Device

ASHP American Society of Hospital Pharmacists

ASHSP American Society of Health System Pharmacists

BIHR Bradford Institute for Health Research

BMJ British Medical Journal

BNF British National Formulary

BP Blood Pressure

BPMH Best Possible Medication History

BTHFT Bradford Teaching Hospitals NHS Foundation Trust

CAM Complementary and Alternative Medicine

CBG Capillary Blood Glucose

CCG Clinical Commissioning Groups

CDSS Computer Decision Support Software

CfH Connecting for Health

CINAHL Cumulative Index to Nursing and Allied Health

COT Clinical Observation Tool

CP Community Pharmacy

CPOE Computer Physician Order Entry

CQUIN Commissioning for Quality and Innovation

CT Core Trainee

xxv

DN District Nurse

DOH Department of Health

DRP Drug Related Problem

ECG Echo Cardiogram

EHR Electronic Health Record

FMEA Failure Modes and Effects Analysis

FY1 Foundation Year 1

FY2 Foundation Year 2

GI Gastrointestinal

GP General Practitioner

GSML Gold Standard Medication List

GSPAML Gold Standard Pre-Admission Medication List

HCA Healthcare Assistant

HCP Healthcare Professional

HMPS Harvard Medical Practice Study

HMR Home Medicine Review

HRO High Reliability Organisations

HSCIC Health and Social Care Information Centre

HSRPP Health Services Research and Pharmacy Practice

ICU Intensive Care Unit

ID Intentional Discrepancy

IHI Institute for Healthcare Improvement

IJPP International Journal of Pharmacy Practice

IMC Inpatient Medication Chart

INR International Normalised Ratio

IPA Interpretive Phenomenological Analysis

IQR Inter-quartile Range

xxvi

ISMP Institute for Safe Medication Practices

IT Information Technology

JCAHO Joint Commission on Accreditation of Healthcare Organisations

KPI Key Performance Indicator

LCP Liverpool Care Pathway

LFT Liver Function Test

LMWH Low Molecular Weight Heparin

LREC Local Research Ethics Committee

MAR Medicines Administration Record

MAU Medical Admissions Unit

MDS Monitored Dosage Systems

MDT Multi-disciplinary Team

MI Myocardial Infarction

MUR Medicines Use Review

NHS National Health Service

NICE National Institute for Health and Clinical Excellence

NIHR National Institute for Health Research

NMC Nursing and Midwifery Council

NPC National Prescribing Centre

NPSA National Patient Safety Agency

NPSG National Patient Safety Goals

OTC Over the Counter

PADE Potential Adverse Drug Event

PAM Pre-Admission Medication

PAML Pre-admission Medication List

PCOS Primary Care Observation Schedule

PCT Primary Care Trust

xxvii

PDSA Plan, Do, Study, Act

PE Pulmonary Embolism

PGD Patient Group Direction

PMR Patient Medication Record

POD Patient’s Own Drug

PPI Proton Pump Inhibitor

R&D Research and Development

RPS Royal Pharmaceutical Society (post September 2010)

RPSGB Royal Pharmaceutical Society of Great Britain (prior to September 2010)

SBAR Situation, Background, Assessment, Recommendation

SCOS Secondary Care Observation Schedule

SCR Summary Care Record

SpR Specialist Registrar

TB Tuberculosis

U&E Urea and Electrolytes

UID Unintentional Discrepancy

UK United Kingdom

US United States

VCR Verification, Clarification, Reconciliation

VTE Venous Thromboembolism

WHO World Health Organisation

WTE Whole Time Equivalent

YCFF Yorkshire Contributory Factors Framework

xxviii

CHAPTER 1 INTRODUCTION

1.1 Background

‘The single biggest problem in communication is the illusion that it has taken place.‘ George Bernard Shaw

‘If you can't describe what you are doing as a process, you don't know what you're doing.‘ William Edwards Deming

Within the current United Kingdom (UK) healthcare system many patients will repeatedly transition from one healthcare organisation to another. The general lack of one unified patient healthcare record within the National Health Service (NHS) means that at each transition point information regarding a patient’s previous medical history must be ascertained, including information about their medication. This is most apparent during transitions between hospital and primary care. At admission, the hospital relies on information and communication from various sources to pull together a picture of previous health and treatment. Similarly, at discharge, primary care relies on the hospital providing relevant information on treatment and changes made to treatment during the patients stay, including any changes to medication. Without relevant, comprehensive communication there is an increased risk of error and patient harm. This can be due to inappropriate omission, addition or alteration of treatment. Throughout the patient’s journey, as they transition between healthcare settings, different healthcare professionals (HCPs) undertake a series of tasks to try to ensure accurate ‘reconciliation’ of information regarding medication. This thesis aims to examine, and critically appraise the medicines reconciliation processes at transition points in care in order to better understand the roles of HCPs, the potential reasons why medication error occurs at transitions and reveal good practices which have been implemented to prevent this type of error occurring. The analysis and subsequent recommendations will take into account current thinking in patient safety, including a human factors and systems perspective.

1.2 Patient Safety

Quality and patient safety are high on the international1-5 and UK agenda,6 with many agencies dedicated to co-ordinating improvement efforts (e.g. National Institute for Health and Care Excellence & National Patient Safety Agency [NICE & NPSA]),6 United States (US) Agency for Healthcare Research and Quality (AHRQ)7 and the Australian Commission for Safety and Quality in Healthcare).3 The drive for improvement started in the 1990s following a large scale epidemiological study (The Harvard Medical Practice Study [HMPS]), conducted by Brennan and Leape8,9 which found that adverse events occurred in 3.7% of hospitalisations; adverse drug events (ADEs) being the most common (19%).10 This signified the start of the identification of a substantial series of preventable events. The nature of the adverse events demonstrated key factors which might increase the likelihood of such an event, including inadequate communication. These findings stimulated the first influential policy document which centred on medical error, published by the US Institute of Medicine; To Err is Human.11 This policy made two overarching recommendations. The first around leadership and the second around the creation of safety systems in health through the implementation of known medication safety practices. Since then, patient safety has gained considerable momentum, with many healthcare organisations placing greater emphasis on improvements for their patients.

Meanwhile, in December 1997 in the UK, the Government published a White Paper The New NHS: Modern, Dependable,12 which set out a ten year modernisation strategy for the NHS. The main aim of this paper was to improve the quality of clinical care delivered to patients within the NHS. Each organisation had a new statutory duty to improve quality through the implementation of clinical governance. They were obliged to translate the principle of clinical governance into locally appropriate structures, processes, roles and responsibilities. From 1999, NHS Boards would assume a legal responsibility for quality of care within their organisation. Soon after, An organisation with a memory (2000),13 the British equivalent to To Err is Human (1999) was released. This highlighted the scale of the problem within the UK and recognised the human cost and organisational cost to the NHS. It was

2 decreed that the NHS must learn from its failures in a new culture of openness. This signified the start of prioritising patient safety within the UK.

Like many terms a variety of definitions exist to describe patient safety. The most widely used is that of Vincent et al (2006) who describes it as:

‘The avoidance, prevention and amelioration of adverse outcomes or injuries stemming from the process of healthcare’.14

To understand patient safety, it is necessary to understand the causes of error and the barriers to ensuring safe patient care. The National Patient Safety Foundation,15 US, elaborates further describing safety as emerging from the components of a system. Thus, improving safety depends on learning how safety emerges from the interaction of the components. This is supported by the work of Reason who proposes that human error can be understood either by taking the person approach or the systems approach.16 It is the latter which squarely acknowledges the human factor perspective. The system approach recognises that humans are fallible and errors must be expected, even within the most eminent organisations. Reason claims that:

‘We cannot change the human condition, but we can change the conditions under which humans work.’

(Reason, 2000, p76916)

In 2006, the Department of Health (DOH) identified, through their publication of Safety First: a report for patients, clinicians and healthcare managers, that despite a high level of awareness, patient safety continued to be a challenge.17 It recommended the implementation of a campaign, Patient Safety First. Their aim was to focus on the safety culture in the NHS and to engage clinical staff as well as enable behavioural change leading to safer, better healthcare.18 The campaign ran from 2008 to 2010 and emphasised the role of human factors when reviewing patient safety.19

1.2.1 Human factors

Human factors can be defined as:

‘the scientific discipline concerned with the understanding of interactions among humans and other elements of a system, and the profession that applies theory, principles, data and methods to design in order to optimize human well-being and overall system performance.’

International Ergonomics Association20

It takes into account psychology and engineering principles to support human performance and safety in an attempt to eliminate human error.21 It examines human interactions and their characteristics within the work environment to design systems which support the cognitive and physical abilities of humans. Work systems that are not designed with human factors in mind can create opportunity for error and compromise patient safety. This is important given that many hospital systems were not designed or planned in the light of human factors.

Multiple elements impinge on human factors and thus affect performance. Humans consistently interact with systems and vice versa, especially in complex organisation such as healthcare. Reason proposes that by recognising and accepting human variability, organisations can devote management resources in a more comprehensive approach to error which takes into account individuals, teams, local conditions and latent organisational failuresa.16 Therefore, if we can understand how organisational systems work and their interrelations with other external systems and people we go some way to understanding and potentially resolving the issues associated with medicines reconciliation; a process that typically includes many different personnel and organisational systems.

It was not until the publication of To err is human11 in 1999 that human factors was recognised as essential to understanding patient safety and why errors

a Latent organisational failures can be described as ‘resident pathogens’ which arise from factors such as organisational culture, management decisions, the design of procedures, or deficiencies in training (Reason, 2000).

4 occur in the healthcare setting. Since then, the emphasis on understanding human factors and human error in healthcare has grown and is becoming more common place. The NPSA believes understanding the role of human factors is vital to reducing human error and harm to patients.22

Reason states that the human error problem can be viewed in two ways: the person approach and the system approach.16 The person approach views ‘unsafe acts’ as arising primarily from atypical mental processes such as forgetfulness, inattention, poor motivation, carelessness, negligence, and recklessness. The system approach concentrates on the conditions under which individuals work and tries to build defences to avert errors or mitigate their effects. Reason hypothesises that most accidents can be traced to one or more of four levels of failure, as described by the organisational accident model (see figure 1.1): organisational influences, unsafe supervision, preconditions for unsafe acts, and the unsafe acts themselves. In this model, an organisation's defences against failure are depicted as a series of barriers, with individual weaknesses in specific parts of the system, which are continually varying in size and position. The system as a whole produces failures when all individual barrier weaknesses align, permitting ‘a trajectory of accident opportunity’, so that a hazard passes through all of the holes in all of the defences, leading to a failure. Reason describes a ‘Swiss cheese’ model (see figure 1.2). It demonstrates a succession of defensive layers (barriers and safeguards) each represented by a slice of cheese with holes which continuously open and shut. The presence of holes in any one ‘slice’ does not normally cause a harmful outcome. Only when the series of holes line up can an error pass through and cause harm. In a similar way, not all medication errors lead to Adverse Drug Events (ADEs) (see section 1.4.1).23,24 The holes in the defences arise for three reasons: active failures, local (or error producing) conditions, and latent conditions. Nearly all adverse events involve a combination of these three sets of factors. The active failures are unsafe practices of the people who are in direct contact with the patient or system, for example the prescriber failing to double check a prescription, or the pharmacist failing to identify an incorrect dose on a prescription. They take a variety of forms: slips, lapses, fumbles, mistakes, and procedural violations (see table 1.1). Local conditions can give rise to increased workload and supervision; latent conditions reflect the

5 structure of the organisation, its resources, management and processes which, either alone or in combination with an active failure, can result in error.16 Researchers have tried to identify these various factors in relation to patient safety events.

Figure 1.1 Stages in the development of an organisational accident25 (reprinted with kind permission)

Figure 1.2 The Swiss cheese model of how defences, barriers, and safeguards may be penetrated by an accident trajectory16 (reprinted with kind permission)

Table 1.1 Errors of action and planning: summary definitions (originally from Reason 1990, modified by Armitage26 (reprinted with kind permission) Slip: a potentially observable error which results from failure in the execution and/or storage stage of an action, regardless of the original plan’s adequacy

Lapse: predominantly related to memory failure, a less observable error which may only be apparent to the protagonist which also results from failure of the execution and/or storage stage of an action, regardless of the original plan’s adequacy

Mistake: a deficiency or failure in the judgmental or inferential processes involved in selecting an objective or means of achieving it, regardless of the outcome of any actions

1.3 Medication Safety

Building a safer NHS for patients: implementing an organisation with a memory (2001)27 led to a supporting publication (Building a safer NHS for patients, improving medication safety 2004) which focused on improving medication safety.28 This aimed to reduce the number of medication safety incidents within the NHS. Medication incidents have been defined by the NPSA as ‘any unintended or unexpected incident which could have or did lead to harm for one or more patients.’29 The majority of incidents do not result in patient harm, however, some go on to cause severe harm or even death.29 Medication safety incidents are known to lead to readmission, increased hospital stay, disability and mortality.28 It is estimated that this preventable harm from medicines could cost England up to 750 million pounds per year.30 Medication errors can occur at any stage in the process of prescribing, ordering, and administration;29 and with so many steps within each stage, there is considerable opportunity for error. Boling suggests that although there are many individual factors which contribute to errors at transition and re-admissions, the complex relationship between medication, communication and care coordination continues to be a problem.31

Several national documents and multiple papers have previously highlighted the need for improving safety and seamless care across interfaces.32-35 It is well

7 known that lack of effective communication poses a threat to both patient safety36,37 and medication safety.38 Communication is one of the main contributory factors discussed within the human factors approach.39 Research illustrating issues with communication about medication at transition points in care dates back over 25 years within the UK.40-42 In response, various interventions and initiatives have been suggested and implemented in some hospitals, such as patients bringing their own drugs into hospital, one stop dispensing, self-administration schemes and electronic prescribing.32-35

1.4 Medicines Reconciliation

In 1998, the Mayo Clinic, Minnesota (a not-for-profit medical and research group) sent a team to an Institute for Healthcare Improvement (IHI) sponsored collaborative that focused on patient safety related to the use of medication. One member of the team (a nurse) highlighted transitions of care as one of the most dangerous times for patients. This included transfers within the hospital, transfers to home or to other long term care facilities. Although there was some previous published evidence that medication issues occurred at transition points, the doctors decided to undertake a retrospective review of 20 charts revealing a high number of medication-related errors. The errors were directly related to misinformation regarding medication at admission to secondary care. The team designed a new process with a specific name; ‘medication reconciliation’ (referred to in the UK as medicines reconciliation). The idea gathered momentum and in early 2000 the US Joint Commission on Accreditation of Healthcare Organizations (JCAHO) became interested. Finally, in 2005 the JCAHO issued a standard which required all hospitals in the US to have a process in place to accomplish accurate medicines reconciliation.43 The goal spanned the continuum of care, applicable to all settings including ambulatory, emergency, urgent, long-term and home care as well as in the inpatient setting (Resar R, Senior IHI fellow, IHI US. Personal communication. 29th November 2013).

Over the last seven years, the profile of medicines reconciliation has been further raised internationally by the World Health Organisation (WHO),1 the JCAHO,43,44 and the IHI2 - all of whom have issued relevant guidance on reducing the number of medication errors attributed to medicines management

8 when patients move care settings. Within the UK, there has been a national drive to improve communication regarding medication at transitions in care through the publication of various national guidelines. These laid out core principles, standards and recommendations regarding the transfer of information at care interfaces and the reconciliation of medication at admission and discharge.

More recently, the notion of medicines reconciliation has been encompassed into the UK medicines optimisation agenda to ensure patients get the right medicine at the right time.45 There is international agreement that organisations should implement appropriate protocols and policies to standardise medicines reconciliation processes and that any existing procedures should be reviewed.1,2,6,44 The difficulties in implementation of standards associated with medicines reconciliation were recognised through the review and refinement of the medicines reconciliation National Patient Safety Goals (NPSG) in the US in 2009, however, medicines reconciliation still remains a NPSG in the US.46

Between November 2003 and March 2007, 1,372,062 incidents were reported to the National Patient Safety Agency,47 117,332 (8.5%) of which were associated with medication and 7070 (0.05%) of those specifically related to medicines reconciliation (Khunpha S, NPSA. Personal communication by e- mail. 3rd November, 2010). A systematic review commissioned by NICE concluded that the scale of the current problems associated with medicines reconciliation, exacerbated by failures in existing systems, contributed to medication error.48 This review estimated that medication error associated with medicines reconciliation at admission varied between 30 to 70% medicines taken per patient. It gained these figures and based its recommendations on an American49 and Canadian50 study making its applicability to the UK unclear. This led to the joint publication between NICE and NPSA which suggested hospitals focus their efforts on improving medicines reconciliation at admission. There has been considerable variation on the implementation of medicines reconciliation across NHS hospitals and a range of different systems have been employed to improve the process.

1.4.1 Measuring error associated with medicines reconciliation

Terms used when describing error and outcome associated with medicines reconciliation include medication error, ADEs and medication discrepancy. A medication error is defined as ‘a failure in the treatment process that leads to, or has the potential to lead to, harm to the patient’.51 This acknowledges that error may or may not lead to patient harm. When a medication error does lead to harm, it can be described as an ADE which is defined as ‘an injury due to medication’.24 The term ADE includes harm caused by the drug (adverse drug reactions (ADRs) and overdoses) and harm from the under-use of the drug (including dose reductions and inappropriate discontinuations of drug therapy)’.24 Aronsen and Morimoto both show the relationship between medication error, ADEs and potential adverse drug events (pADEs) (see figures 1.3 and 1.4).23,24 Medication errors can be further classified to include various other types of error e.g. prescribing error, administration error etc. There is a difference between a prescribing error and medicines reconciliation-related error. A prescribing error is ‘a failure in the prescribing [decision-making] process that leads to, or has the potential to lead to, harm to the patient’.23 Dean et al specify that ‘a clinically meaningful prescribing error occurs when, as a result of a prescribing decision or prescription writing process, there is an unintentional significant (i) reduction in the probability of treatment being timely and effective; or (ii) increase in the risk of harm when compared with generally accepted practice’.52 In comparison, a prescription error is ‘a failure in the prescription writing process that results in a wrong instruction about one or more of the normal features of a prescription’.23 Medicines reconciliation error is not well defined although one could argue that medicines reconciliation error in part is ‘a failure in the medication history taking process’, although this does not encompass the errors in communication due to poor documentation when medication is altered.

Figure 1.3 Relationship between medication error and adverse drug events (ADEs) according to Aronsen et al23 (reprinted with kind permission)

Figure 1.4 Relationship between medication error and Adverse Drug Events (ADEs) according to Morimoto et al 24 (reprinted with kind permission)

Bates et al highlights that medication errors are common, although relatively few result in ADEs.53 ADEs are most common in the first three weeks post discharge54 and are often preventable.55,56 Where medication errors are active failures, they can be classified according to whether they are mistakes, slips, or lapses. During medicines reconciliation, there can be a mismatch between two sources of information or two medication lists e.g. the general medical practice electronic health record (EHR) and the inpatient medication chart (IMC). At the point of comparison it is unclear whether the mismatch is an error (mistake, slip or lapse) or an intentional, conscious change to the patient’s medication regimen. These mismatches in information are referred to as medication discrepancies. The Collins dictionary defines ‘discrepancy’ as ‘a conflict or variation, as between facts, figures, or claims’.57 The term medication discrepancy, therefore implies a lack of agreement between different medication regimens, and provides a more precise term for capturing the potential medication errors that can occur at transition points in care, mainly admission and discharge. The discovery of a medication discrepancy does not necessarily infer error nor does it imply potential harm.

Medication discrepancies can be further classified into those which are intentional discrepancies (IDs) (e.g. stopping aspirin in a patient with a gastrointestinal bleed) and those which are unintentional discrepancies (UIDs) (e.g. accidental omission of medication at admission) (see figure 1.5). Where medication discrepancies are unintentional this has the potential to result in an ADE and patient harm although not all UIDs will lead to harm. This is sometimes termed potential adverse drug event (pADE). The WHO defines a pADE as a ‘serious error or mishap that has the potential to cause an adverse event but fails to do so because of chance or because it is intercepted (also called ‘near miss’ or ‘close call’)’.58 In the case of medicines reconciliation (at admission) it is often intercepted by the pharmacist during the verification of the medication history, thus does not result in an ADE.

Figure 1.5 Relationship between discrepancy and adverse drug event (ADE)

Intentional

Discrepancy No Harm (pADE)

Unintentional

Harm (ADE)

1.5 Rationale for this Study

The majority of published studies associated with medicines reconciliation have been conducted in the US making some aspects of generalisability to the UK difficult and creating a need for further studies in the UK. The systematic review which underpins the NICE & NPSA guidance,6,48 also highlighted the need for better designed studies, which include adequate methods to minimise bias. This is reinforced in further systematic reviews by Mueller et al and Chabbra et al who highlight the need for more rigorously designed studies to compare different medicines reconciliation practices and their effectiveness especially on clinical outcomes.59,60 The American Pharmacists Association and American Society of Health-Systems Pharmacists (APhA & ASHSP) have emphasised the need for more studies on effective medicines reconciliation processes which can identify best practice’,61 with Campbell et al requesting better controlled studies.48 Before relevant interventions can be designed and implemented it is important to understand where things go wrong and how errors associated with medicines reconciliation arise through exploration of the process.

There has been much research published over the last 15 years highlighting the problems associated with different aspects of medicines reconciliation,40,48,62 together with many recommendations for good practice some of which demonstrate a decrease in error.59,63-73 Despite these recommendations, there is little to suggest that practice has changed. There is continued need for

13 process improvement to enable a smoother transition between care settingsb, a significant reduction in medication errors and discrepancies and the eradication of adverse events. Within the UK, the most common source of error associated with medicines reconciliation is the omission of clinically relevant medication from in-patient charts.74-79 This often goes unnoticed, highlighting the importance of medicines reconciliation processes at transitions in care.78

The National Prescribing Centre (NPC) stipulates that there needs to be enough competent people available to undertake medicines reconciliation at all times and that restricting the responsibility for medicines reconciliation to one professional group could make it difficult for some organisations to provide a reliable service, especially ‘out-of-hours’.80 The WHO goes further to suggest that healthcare organisations assign personal, professional and corporate levels of responsibility, and allocate clear roles and responsibilities to all HCPs for all steps in the medicines reconciliation process to ensure accountability.1 There are recommendations that the responsibility for carrying out medicines reconciliation should be included in job descriptions and personal development, endorsing it as a multidisciplinary effort. However, there is no clear indication within these documents or other supporting evidence to suggest what these roles and responsibilities may be. This is supported by an analysis of clinicians perceptions which revealed two questions: What does medicines reconciliation really mean and who is actually responsible for the process?81

Many descriptive papers (predominantly international) emphasise the rates and severity of discrepancies with fewer looking at the solutions to the errors that arise and the barriers that may be encountered. These descriptive papers often report on a specific specialty or department e.g. care of the elderly or accident & emergency (A&E) and isolated parts of the process rather than the process as a whole, across a healthcare economy. This was previously identified as an issue by Pickrell et al82 in 2001 with no signs of improvement since. Pharmacist-led interventions have been shown to have greatest benefit in comparison to those delivered by other HCPs. However, the role of the HCP and indeed the patient within the process is not well defined. Recent advances in technology provide b A care setting may include hospitals, nursing homes, residential homes, a patient’s own home or intermediate care.

14 opportunity to improve the medicines reconciliation process and add safeguards to the system, however, the extent to which they add benefit and the role of IT is not clearly determined.

The timeliness of this research is significant as the UK starts to implement the medicines optimisation agenda; a holistic patient-focused approach to getting the best from medicines which requires an enhanced level of professionalism, and partnership between clinicians and the patient.45 Medicines optimisation looks at how patients use medicines over time. It may involve stopping some medicines as well as starting others both of which must be communicated accurately to other healthcare providers and indeed the patient. Medicines optimisation covers all aspects of medicines usage, including safe processes and systems, and effective communication between professionals. This research as well as exploring the role of the patient in medicines reconciliation might also inform the implementation of the ‘Professional Standards for Hospital Pharmacy’ which provides a framework through which to improve the quality of hospital pharmacy services and includes medication accuracy at admission and provision of information to patients.83

Mindful that much of the work on the scale of reconciliation errors and discrepancies has been in North America, this research aims to determine whether the epidemiology of medicines reconciliation indicators within the study population mirror those of international studies. This research will then identify through observation, both good practice and the reasons why medicines reconciliation errors occur, whereabouts in the process they occur and the role of both the patient and healthcare HCP. A human factors approach is employed to explore medicines reconciliation following the patient journey from primary to secondary care and back again, transitioning through the hospital setting.

In summary, there is lack of effective medicines reconciliation process studies especially those which look at the process as a whole. There is also need to define roles and responsibilities surrounding medicines reconciliation. This study addresses these deficiencies and makes recommendations about standards of good practice which should be incorporated into the medicines reconciliation process together with the roles and responsibilities of HCPs.

1.6 Thesis Outline

The purpose of this thesis is to examine current medicines reconciliation practice within primary and secondary care together with the role of different HCPs and patients within the process. The objectives are: a) To determine a picture of medicines reconciliation within the West Yorkshire region and consider where this fits nationally and internationally, demonstrating the difference between desired practice, guidance and current practice. b) To establish the role of different HCPs including pharmacists, doctors and nurses within the medicines reconciliation process. c) To explore the role of patients and patient engagement in the reconciliation of medicines. d) To examine the overall process by which hospital staff undertake medicines reconciliation. e) To examine how general practitioner (GP) practices within primary care process information and/ or receive communication about medication from other care settings. f) To determine the role of IT and, in particular, the EHR in improving communication around Medicines Reconciliation.

The thesis has been divided into four stages which scope current practice and ways of working then recommend standards. Using a mixed methods, exploratory design, data were collected through: (1) a literature review, (2) a descriptive prospective study which collected data on the type and severity of discrepancies related to medicines reconciliation; (3) a structured observation of the reconciliation process within primary and secondary care (including documentary analysis); qualitative and quantitative data were collected and analysed to illustrate current medicines reconciliation practice and highlight areas of quality and concern. The analysis was informed by a human factors perspective and communication theory. The reviewed literature and data collected was then used to (4) identify good practice recommendations for 16 medicines reconciliation. Potential interventions and appropriate further work are also identified. This thesis comprises seven Chapters outlined below:

Chapter 2 explores the literature, first looking at definitions of medicines reconciliation and the scale of the problem, secondly reviewing the process of medicines reconciliation and the role of the HCP and lastly looking at the patient’s role within the medicines reconciliation process.

Chapter 3 describes the methodologies used, justifies the use of a mixed methodology and offers the rationale behind each method and how the methods relate to underpinning theory.

Chapter 4 applies a quantitative methodology to identify the type and severity of discrepancies observed in four acute UK hospitals.

Chapter 5 uses an evidence-informed framework and qualitative methods to contextualise the findings from the prior quantitative study (Chapter 4). Through the systematic observation of the reconciliation process and the role of various HCPs, the tasks relating to medicines reconciliation within secondary care are examined and considered in the light of human factors. This includes nurse handover, medicines administration, ward rounds, pharmacy staff verification of the medication history, patient counsellingc, prescribing & transcribing and the writing of discharge advice notes.

Chapter 6 uses a second evidence-informed framework and qualitative methods to observe the receipt and processing of discharge information within primary care.

Chapter 7 discusses the research findings with reference to the existing literature to explain both concerns and good in the medicines reconciliation process. This chapter also discusses potential implications and recommendations for practice, strengths and limitations of the research and proposals for further work.

c Counsel – To provide the patient with information and advice on how to take their medication

CHAPTER 2 LITERATURE REVIEW

2.1 Background

This chapter explores medicines reconciliation by presenting the findings from a narrative review to determine what is already known about medicines reconciliation, the processes associated with medicines reconciliation and the roles of those involved in the process. It will outline significant findings and inconsistencies, highlighting any gaps in the current literature to support and demonstrate this study’s contribution. The review will first set the context of medicines reconciliation including existing definitions and the scale of the current problems concerning reconciliation (Part 1); secondly review the process of medicines reconciliation and the role of different HCPs, establishing the evidence for specific interventions (Part 2); and lastly, look at the patient’s role within the medicines reconciliation process (Part 3). The inclusive approach of a narrative review enables comprehensive coverage of both roles and process and includes a range of empirical studies with varied measures and methods, which is unlikely to be covered in a systematic review. The complexity of medicines reconciliation, the sheer range of studies, and the variation in national guidance also makes the topic more suited to a narrative review. Additionally, the narrative review provides detailed knowledge of the existing context, revealing problems, weaknesses and contradictions within the literature. A traditional systematic review would diminish the number of studies included and not allow opportunity for full discussion and description of roles and process, ultimately limiting the full picture of current practice. The literature review reports studies from 2000, up to and including October 2013. Some studies from before this date are cited which provide background to the medicines reconciliation story.

2.2 Methods

A literature review was conducted using a systematic approach to determine the current UK picture of medicines reconciliation, within an international context. Studies included in the review were identified by keyword searches of Medline, Cumulative Index to Nursing and Allied Health (CINAHL), Cochrane, Pubmed and Embase databases. Keywords included medic* reconc*, medic* discrep*, patient discharge and community pharmacy, medication error and patient 18 history taking, seamless care, and interface. A search of publications by relevant government bodies and health organisations in the UK was also undertaken (including the DOH, NICE & NPSA). Search terms were amended to match keywords used in specific databases.

Results were merged using reference-management software (Endnote X7®), and duplicate references removed. Articles were excluded if they were not published in English; published in peer-reviewed journals; empirical studies or primary studies. An information scientist assured the search strategy to ensure the capture of all relevant studies and ensure the validity of the literature review. All studies published before October 2013 were sought. Studies of all ages were initially included. The reference lists of all relevant articles identified were hand searched and relevant papers obtained.

The electronic searches yielded 14,329 results from 1974 onwards. Following elimination of irrelevant papers such as Dahlkvist et al84 and Zhang et al85 and duplicates, the inclusion and exclusion criteria were applied to 1128 papers. With the inclusion of 32 relevant papers found through hand searching the total number of empirical studies which were deemed relevant was 452. Due to the high volume of studies and their age an exclusion criteria of papers published before 2000 was applied. The current relevance of studies prior to 2000 is unclear due to the length of time since publication. This year also coincides with the introduction of clinical governance and the increased drive for patient safety (including medication safety) outlined within the NHS Plan86 and An Organisation with a Memory.13 The application of these criteria eliminated 67 studies leaving 385 papers, 60 of which were from the UK and 325 international. The literature review focused on those studies from the UK, using the international studies to put them in context (see figure 2.1).

The studies were reviewed to determine whether they were intervention or non- intervention studies. The intervention studies were then assessed for quality using a simple quality assessment tool adapted by Spinewhine et al, which reviewed the type of study design and whether the study was controlled, randomised or blinded.87 Also, whether a power calculation had been conducted, statistical tests used and confidence intervals reported. Overall, from the 167 intervention studies, 60 were controlled of 19 which 23 were randomised and 7 blinded. Inferential statistics, used to demonstrate significance, were reported in 98 studies, however, only 29 (1 UK) reported power calculations and 46 confidence intervals. Of the 22 UK intervention studies, nine were controlled of which two were randomised and none blinded.

The remainder of studies consisted of a mixture of prospective descriptive comparative studies and non-controlled before and after studies, some of which were evaluated, for example, through opinion88-91, vignettes92 or competency assessment93-97 of the HCP rather than more quantitative outcome measures e.g. rates of discrepancy. The way in which the authors described and explained the methods used within the study was not always clear.

From this analysis it is clear that intervention studies published in relation to medicines reconciliation are heterogeneous in their design with a small number of randomised control trials (RCTs) and many which did not include a specific description of their design. This supports the conclusion of the systematic review commissioned by NICE which concluded that 'better designed studies that include adequate methods to control for bias such as randomisation, allocation concealment, blinding at outcome assessment and intention to treat analysis' are needed. Due to the nature of some of the interventions it would be difficult to blind the patient and in some cases the HCP providing the intervention, for example the provision of hand-held information or follow-up visits post discharge. The investigator or statistician could however be blinded when undertaking analysis. This is more important for those studies which have subjective measures rather than objective measures.98

Each study identified was also categorised into topic area within Endnote®, for example scale, severity, admission, discharge, counselling etc. These topic areas form the sub-headings within the literature review. Each paper was then reviewed methodically and relevant information extracted and tabulated where relevant. The results presented are predominantly findings rather than the study author opinion.

Figure 2.1 Flow diagram of search and retrieval process

Potentially relevant studies identified through searches n = 14329

Elimination of irrelevant studies and duplicates n = 1128

32 documents Number of empirical found through studies Hand Searched n = 452

Number of studies published ≥ 2000 n = 385

Total Number UK Total Number studiess International studies n = 60 n = 325

Findings

The results of the literature review will be presented in 3 parts:

Part 1 – Describes the context of medicines reconciliation within the UK including definition, scale, severity and impact of discrepancies and/or error and places them in an international context.

Part 2 – Looks at interventions introduced and tested at the different stages within the medicines reconciliation process and HCPs’ roles within the process.

Part 3 – Describes the role of the patient within medicines reconciliation.

2.3 Part 1 - Definition, Scale and Severity of Discrepancies Found

2.3.1 UK and international context

There is an international drive to reduce the number of medication errors associated with medicines reconciliation at all transition points in care. The UK is no exception, with relevant guidance and standards being published by several organisations. Significant numbers of errors associated with medication and more specifically medicines reconciliation are cited in the literature and also reported to the NPSA (Khunpha S, NPSA. Personal communication. 3rd November, 2010). A systematic review commissioned by NICE concluded that the scale of the current problems associated with medicines reconciliation and failures in existing systems is a major contribution to medication error.48 This first section of the literature review will describe and define medicines reconciliation, and provide a picture of the scale and severity of discrepancy rates at admission and discharge from secondary care.

Studies reporting the rates of error in medication histories detected at admission and discharge have been published worldwide (see table 2.1), however, the majority of studies published which highlight the scale of error have been conducted within the US, making their applicability to the UK unclear due to the difference in healthcare systems. There has also been a higher number of non- intervention studies published compared with intervention studies both in the UK and internationally. Several of the studies involve complex interventions i.e. interventions with several interacting components. Complex interventions are widely used in the health service and present a number of problems for evaluation as it can be difficult to determine exactly which aspect has contributed to an observed outcome.99 This is particularly relevant as several of the UK intervention studies involve multiple interventions with multiple outcome measures.74,75,78,82,100-107

Table 2.1 Medicines reconciliation studies by country

Total Intervention Non-intervention Country number of Studies studies studies

UK 22 38 60

Australia 17 17 34 Austria 1 0 1 Belgium 2 2 4 Canada 12 17 29 Denmark 3 6 9 France 3 0 3 Germany 1 2 3 Ireland 1 5 6 Israel 2 1 3 Jordan 0 1 1 Malaysia 0 1 1 Netherlands 7 10 17 New Zealand 0 1 1 Nigeria 1 3 4 Norway 0 1 1 Pakistan 0 1 1 Saudi Arabia 0 1 1 Singapore 0 1 1 Slovenia 0 1 1 Spain 2 5 7 Sweden 5 6 11 Switzerland 1 1 2 Taiwan 0 2 2 USA 87 95 182 Total Non-UK 145 180 325 Overall Total 167 218 385

Many studies focus on a specific clinical specialty, most commonly elderly care67,71,92,108-115 or general medicine,63,89,116-123 with fewer studying surgery,124,125 psychiatry126,127 or paediatrics.128-130 There are few studies which demonstrate

23 discrepancy rates within the UK. Where they exist they have been published in abstracts rather than full studies.131-133 Nevertheless, it is clear that there is a problem internationally regarding the transfer of information about medication between care settings. Different countries are at varying stages with the implementation of safe systems, processes and policies to ensure medicines reconciliation is conducted appropriately. In more developed countries, such as the UK, US, Canada and Australia, aims at improvement have been implemented over the last 10 years,2-6,80,134,135 however, in more developing countries such as Malaysia the concept of medicines reconciliation is just beginning to emerge.136 However, within the developed countries there is a lack of uniformity around the implementation of improvement processes and outcome measures to support medicines reconciliation.137

Internationally, a number of terms and acronyms are used to describe the complete list of medicines a patient was taking prior to hospital admission. These include ‘best possible medication history’ (BPMH),138,139 ‘pre-admission medication’ (PAM), ‘pre-admissions’ medication list’ (PAML),69,140-142 ‘gold standard medication history’,143 ‘gold standard medication list’ (GSML)140 and ‘gold standard pre-admission medication list’ (GSPAML).144 There is no consensus as to which term should be used to best describe the most current and complete theoretical list of medication the patient is taking. The term used is perhaps not as important as the methods used to determine the list. Although the literature describes the overarching stages within a medication history, there is lack of evidence around how to elicit a ‘gold standard’ to support systematic and consistent collection. Pippins et al provide a protocol for collecting a ‘gold standard’ medication history, although it is not clear on what this is based.145 There is consensus, however, that a medication history should contain prescription medications, over the counter (OTC) medications, Complementary and Alternative Medicines (CAMs), and allergies and ADRs/intolerances including the reaction suffered.146-150

The studies which have investigated discrepancies associated with medicines reconciliation fall into two main categories; those which look at the recorded medication history in comparison with the best possible medication history available and those which look at the discrepancies between the different

24 sources of information available which contribute to the BPMH. These will be discussed in the sections 2.1.5 Scale of the Problem and 2.1.8 Sources of Information respectively.

2.3.2 Definition of medicines reconciliation

(This section was informed by personal communication with Dr.David Alldred; Senior Lecturer in Pharmacy Practice, University of Bradford, 1st May 2013)

Although the concept of medicines reconciliation was introduced in 1998 (Roger Resar. Senior IHI fellow. IHI, US. Personal communication. 29th November 2013), it became more widespread within the literature around 2003,151-154 especially in the US. Whilst it is commonly used in practice in the UK, there are few UK studies which utilise the term, referring more often to terms such as medication history taking and discharge communication etc. In practice, the term is often used interchangeably to mean different things, sometimes, specifically to refer to the verification step undertaken by a pharmacist to ensure that the medication history has been taken accurately and medicines prescribed appropriately on admission to hospital. Alternatively, the term can refer to the creation of an accurate account of a patient’s medicines at all transition points in care.

There is no universally accepted definition used for medicines reconciliation, either internationally or within the UK. The various definitions which do exist have been most commonly developed by either governments or professional bodies to try and improve standards of practice surrounding medicines reconciliation (see table 2.2). One commonly cited definition of medicines reconciliation was produced by the IHI. It describes medicines reconciliation as:

‘the process of creating the most accurate list possible of all medications a patient is taking — including drug name, dosage, frequency, and route — and comparing that list against the physician’s admission, transfer, and/or discharge orders, with the goal of providing correct medications to the patient at all transition points within the hospital’.2

This restricts the idea of medicines reconciliation purely to a secondary care setting, disregarding the interface between primary and secondary care. An

25 alternative description is provided by Cambridge et al2 who believe medicines reconciliation results in an ‘accurately communicated complete list of medications’. Similarily, the WHO describe it as prevention of miscommunication; this leaves room for interpretation of finer detail.155 Although the NICE & NPSA technical appraisal does not cite a definition, they specifically describe the aim of medicines reconciliation at hospital admission:

‘to ensure that medicines prescribed on admission correspond to those that the patient was taking before admission, highlighting specifically the details to be recorded, i.e. the name of the medicine(s), dosage, frequency, and route of administration’.6

Whilst the NPSA guidance specifically refers to admission, it also recognises that medicines reconciliation can occur when patients are transferred to other units within the hospital, to another hospital, or at discharge from hospital. Others have defined medicines reconciliation for the purpose of their research (see table 2.3). Without a common definition for medicines reconciliation, it is unsurprising that the term can be misinterpreted.

Often medicines reconciliation is described as the process of matching one list of medicines to another and resolving any discrepancies that arise e.g. medication taken before admission compared with the IMC. Cited definitions include several general components:

1. Prevention of medication errors 2. Identification of an accurate list of medication used prior to transfer 3. Identification and clarification of ‘discrepancies’ (between medicine lists) 4. Verification/ Resolution of medication discrepancies 5. Communication of changes to the patient and other HCPs

Table 2.2 Definitions of ‘medication reconciliation’ used by government and professional organisations

Organisation Definition Australian ‘Medication reconciliation is a formal process of obtaining and Commission on verifying a complete and accurate list of each patient’s current Safety and medicines. Matching the medicines the patient should be Quality in prescribed to those they are actually prescribed. Where there are Healthcare3 discrepancies, these are discussed with the prescriber and reasons for changes to therapy are documented. When care is transferred (e.g. between wards, hospitals or home), a current and accurate list of medicines, including reasons for change is provided to the person taking over the patient’s care. Points of transition that require special attention are:

• Admission to hospital

• Transfer from the Emergency Department to other care areas (wards, Intensive Care, or home)

• Transfer from the Intensive Care Unit to the ward

• From the hospital to home, residential aged care facilities or to another hospital.’

National Patient ‘The aim of medicines reconciliation on hospital admission is to Safety Agency ensure that medicines prescribed on admission correspond to (NPSA) (UK) 6 those that the patient was taking before admission. Details to be recorded include the name of the medicine(s), dosage, frequency, and route of administration. Establishing these details may involve discussion with the patient and/or carers and the use of records from primary care. This does not include medicines review.’ Institute for ‘The process of creating the most accurate list possible of all Health medications a patient is taking — including drug name, dosage, Improvement frequency, and route — and comparing that list against the (USA) 2 physician’s admission, transfer, and/or discharge orders, with the goal of providing correct medications to the patient at all transition points within the hospital.’ The Joint ‘Medication reconciliation is the process of comparing a patient's Commission medication orders to all of the medications that the patient has (USA)43 been taking. This reconciliation is done to avoid medication errors such as omissions, duplications, dosing errors, or drug

Organisation Definition interactions. It should be done at every transition of care in which new medications are ordered or existing orders are rewritten.’ Institute for ‘Medication reconciliation is a formal process in which healthcare Safe Medication providers work together with patients, families and care providers Practices (ISMP) to ensure accurate and comprehensive medication information is (Canada)4 communicated consistently across transitions of care. Medication reconciliation requires a systematic and comprehensive review of all the medications a patient is taking (known as a BPMH) to ensure that medications being added, changed or discontinued are carefully evaluated. It is a component of medication management and will inform and enable prescribers to make the most appropriate prescribing decisions for the patient.’

High 5’s Project ‘[Medication reconciliation] addresses the prevention of (WHO)155 medication errors resulting from incomplete or mis-communicated information at points of transition in the patient care process, such as admission to the hospital, transfers within the hospital, and discharge from the hospital.’ Agency for ‘Medication reconciliation is a process to decrease medication Healthcare errors and patient harm by comparing the patients current Research and medication regimen against the physicians admission, transfer, Quality and/or discharge orders to identify discrepancies.’ (AHRQ)156 American ‘Medication reconciliation is the comprehensive evaluation of a Pharmacists patient’s medication regimen any time there is a change in Association therapy in an effort to avoid medication errors such as omissions, (APhA) & duplications, dosing errors or drug interactions, as well as to American observe compliance and adherence patterns. This process Society of should include a comparison of existing and previous medication Health System regimens and should occur at every transition of care in which Pharmacists new medications are ordered or existing orders are rewritten or (ASHSP)157 adjusted or if the patient has added non-prescription medications to (his or her) self-care.’

Table 2.3 Definitions of ‘medication reconciliation’ used in published papers

Author Definition of medication reconciliation Bandres et al158 ‘The formal, standardised process of reviewing patients’ complete previous medication regimen, comparing it with current prescriptions, and analysing and resolving any discrepancies found.’ Boockvar et al159 ‘Medication Reconciliation involves creating a complete and accurate prior medication use list, identifying discrepancies between current and prior medication use, and ensuring prescriber awareness of current and prior medication use to inform prescribing decisions.’ Cua and ‘The process of comparing the patient’s medication list to Kripilani160 what he or she should be receiving, and resolving any discrepancies.’ Feldman et al70 ‘An active process that occurs at transitions in care (admissions, transfers in level of care, and discharge) and is designed to prevent medication errors as the patient moves across the continuum of care.’ Schwarz and ‘The process by which an accurate list of a patient’s home Wyskiel161 medications is compiled and compared with admission, transfer and discharge orders.’ Mergenhagen et ‘The practice of reviewing and comparing a patient’s al162 medication regimen throughout care transitions to prevent unintentional discrepancies.’ Meyer et al163 ‘The process of comparing a patient's medication orders to all the medications that he or she has been taking.’ Miller et al164 ‘A process to reduce errors and harm associated with loss of medication information as the patient enters and moves through a new healthcare system.’ Paton et al10 ‘To ensure that medicines prescribed on admission to hospital do not differ unintentionally from those that the patient was taking immediately prior to admission.’

Author Definition of medication reconciliation Rogers et al165 ‘Reconciling medications is a formal process for creating the most complete and accurate list possible of all pre- admission medications for each patient and comparing the physician’s admission, transfer and/or discharge orders against that list. Discrepancies are brought to the attention of the physician and, if appropriate, changes are made to the orders. Any resulting changes are documented.’ Varkey et al64 ‘The process of identifying the most accurate list of all medications a patient is taking, including the name, dosage, frequency, and route of each medication, and using this list to provide correct medications for the patient anywhere within the healthcare system.’ Vira et al166 ‘A three step process of verifying medication use, identifying variances, and rectifying medication errors at interfaces of care.’ Yi et al167 ‘Medicines reconciliation accurately identifies all medications patients are taking – including drug name, dosage, frequency and route, and uses this list to provide correct medications for patients anywhere within the healthcare system.’ Zeigler et al168 ‘The process of comparing medications that the patient was taking prior to admission or entry to a new setting with those that the organisation will provide and prescribe at discharge.’

2.3.3 Models of medicines reconciliation

Internationally various models and acronyms have been used to describe the medicines reconciliation process. Originally it was described by the IHI in the US as VCR (Verification, Clarification, Reconciliation).2 Subsequently other organisations have tried to define the stages within the process and produce simplified models to assist medicines reconciliation (see table 2.4). All involve the collection of the medication history, verification to ensure it matches what is

30 prescribed, documentation and communication, mirroring the components found within the cited definitions. However, all describe it in slightly different ways, with two different organisations using the term reconciliation to describe completely different stages, one the verification of the history and the other the documentation.2,169

Table 2.4 Models showing the stages of medicines reconciliation

Description of Institute for National Joint Society of Hospital stage Healthcare Prescribing Commission Pharmacists of Improvement Centre (NPC) on Australia 169 (IHI) (VCR)2 Accreditation (3Cs)80 of Healthcare (4 step process) Organisations (JCAHO)

(5 step process)43 Collection of Verification Collection Develop a list Medication history medication of current history medications Develop a list Confirmation (with of medications as many other to be sources as deemed prescribed appropriate)

Verification of Clarification Checking Compare the Reconciliation the medication medications on prescribed to the two lists ensure it matches Make clinical decisions based on the comparison Documentation Reconciliation of changes Communication Communicating Communicate Transfer of verified to next provider the new list to information appropriate caregivers and to the patient

In the current study, medicines reconciliation is conceptualised as a process designed to ensure an accurate, complete list of medications is attained at transition points in patient care. This incorporates the components found within the published definitions. The transition points may occur within secondary care or primary care, or alternatively, across the interface. It is based on the UK NPC’s 3Cs approach; collection (collection of the medication history by doctors 31 at admission), checking (subsequent verification of the medication history by pharmacy staff) followed by communication to primary care.80

Figure 2.2 Process map of verification of medication history taking 5 (reprinted with kind permission)

The stages of verification of the medication history are outlined in figure 2.2 and involve the review of a number of sources of information to create a BPMH which is compared to a prescription and any discrepancies identified deemed intentional or unintentional through reviewing the notes or discussing with the prescriber.

The APhA & ASHSP state that medicines reconciliation should be performed when the patient undergoes a transition of care between care settings or transfer of care within a setting. A transition in care can be described as:

‘the movement of patients between healthcare locations, providers or different levels of care within the same location, as their condition and care needs change.’170

2.3.4 Classification of medicines reconciliation-related error

In addition to the lack of definition, there is no universal taxonomy or classification system for discrepancies or errors associated with medicines reconciliation, making it difficult to standardise reporting of the types of error which occur. Studies have used varying categories to classify the discrepancies, with some more detailed than others. All classifications include the omission and addition (sometimes referred to as commission) of medication.77,109,125,145,166,171-183 Apart from these common items, the classification schemes vary with some amalgamating UIDs such as alterations in strength form and dose into one category176,177 and others leaving them separate.145,184 These variations also make it difficult to compare discrepancy rates between studies.

2.3.5 Scale of medication discrepancies at transitions

Studies have used different measures, denominators and terminology to highlight the scale of error due to inadequate medicines reconciliation. This makes it difficult to compare error rates between studies. Most commonly the number of UIDs is measured, with some studies using the number of interventions as an outcome measure. These are sometimes cited per medicines reconciliation,74 per patient185 or per number of drugs prescribed.6 Other studies have used ‘drug related problems’ (DRPs).72,186,187 The lack of consistency with the measures used makes comparisons between studies difficult.

There are few studies within the UK which record the proportion of patients who have their medication verified by a pharmacist post-hospital admission133 or the number and type of discrepancies that can occur, making it difficult to compare studies. The largest published UK study which looked at unintended discrepancies at admission was conducted in 50 NHS hospitals by Dodds,77 it identified 11,366 unintended discrepancies over 8621 medicines reconciliations

33 averaging 1.32 unintended discrepancies per medicines reconciliation. Results from other studies varied from 1.06 UID per patient to 1.83 UID per 6 medications.48,74,172,185 Fertleman et al188 reported an even higher error rate of two interventions per patient, but it is unclear whether these were just associated with medicines reconciliation or included other clinical interventions. Internationally the number of UID per patient at admission varied from 0.62 per patient to 3.5 discrepancies per patient.145,175,189,190 Other studies presented the percentage of patients experiencing a UID which varied from 32 to 90% of patients.117,145,166,173,176,178,179,189,191-197 It is difficult to compare rates of discrepancies as the way in which they have been calculated and presented differs (see section 2.3.4). Linsky and Simon report a rate of 3.7 discrepancies per patient, but this is of those patients who had a discrepancy rather than the total within the study.173 Similarly, Forjuoh report a rate of 2.0 discrepancies per patient, however this also includes omitted OTC medication and CAMs.196 Some international studies have also looked at the rate of omission at discharge, showing percentage of patients or discharge summaries with an UID to vary from 23.3% to 88.5%.109,111,126,174,177,183,198-200

Both nationally and internationally, the single most common discrepancy found when reconciling medication is omission.77,109,125,145,166,171-183,201 A number of studies have looked at the rate of omission of medication at admission and discharge with figures ranging in the UK from 56% to 73%74-79 and internationally from 17% to 95% per patient.109,125,145,166,171-183,193,201 Most studies, however, report figures greater than 50%.70,145,167,180,181,194,195,202 Missed information from the medication history has been shown to be clinically relevant and can include potentially fatal errors such as the omission of anticonvulsants and insulin.78 These errors would have been difficult to identify during a basic chart review, highlighting the importance of the medicines reconciliation process.78 Within the UK, studies associated with medicines reconciliation have been conducted in specialties, including paediatrics,88,128,203 renal medicine,204 the intensive care unit (ICU),205 A&E and surgery,76 however, most medicines reconciliation studies have been conducted in medical admissions units (MAU). This range of UK specialty settings mirrors the international picture in terms of study settings.124,206-215 Transitions to care homes are not well studied within the UK or internationally in terms of medicines

34 management. An Australian study showed that where patients transfer to residential care facilities doses can be missed or significantly delayed post discharge.216

2.3.6 Severity of ADEs and pADEs related to medicines reconciliation

A number of different scales are used within the literature to categorise the potential severity of discrepancies discovered during medicines reconciliation. Severity scales are historically derived from actual harm events. It is difficult to predict the actual severity of unintended discrepancies as often the discrepancies are discovered before they result in harm (see figure 1.5). This provides no tangible or measureable outcome on which to measure severity, leading to a prediction rather than occurrence. Irrespective of the scale used, all studies reported that the potential severity of the discrepancies found would either not cause harm or would cause minimal to moderate harm with few discrepancies potentially causing death or major harm.50,70,113,129,145,164,178,180,181,191,192,194,197,217 Only one study has looked at actual ADEs occurring from medication discrepancies. It determined that of 1350 discrepancies; 65 (4.8%) resulted in ADEs.159

The scales vary in their content and number of points on the scale (from 3 to 5). Some studies use degrees of severity when classifying ADE e.g. lethal, serious, significant, minor.65,210,218,219 Some look at discomfort and monitoring e.g. Climente-Martí et al189 and Arora et al.217 Others use the likelihood to cause harm e.g. unlikely, possible, and probable.183 For accurate measurement and comparison between studies it is important that the severity scales used reflect the potentiality of the harm and use the same point scale with common descriptors.

2.3.7 Risk factors for experiencing medication discrepancy at transition points

Numerous risk factors have been shown to contribute to an increased number of discrepancies at admission and discharge, although not within the UK. The most common risk factor is the number of medicines a person is taking. The higher the number of medicines the more likely the patient is to experience an

35 unintended discrepancy,70,109-111,117,145,173,177,180,181,183,197,212 with the odds increasing by 10% for every additional medication taken prior to admission (OR 1.10; 95% CI 1.06-1.14).117 Some studies found that where patients are taking greater than four197 or five109,111,212 medications their risk increases, especially where these are for gout, muscle relaxants, lipid-lowering medication, antidepressants or respiratory medicines, however, this is a single study.145 Cornu et al demonstrated using binary logistic regression demonstrated that if the patient was taking ≥ five medicines then their risk was 2.85 times higher than fewer medicines (P=0.01). For every additional medication listed in the discharge letter the likelihood of experiencing one or more discrepancies increased by 20%.109 Using multivariate analysis they demonstrated that more than five drugs increased the likelihood of experiencing more than one discrepancy by a factor of 3.22 (95% CI 1.40-7.42; P=0006). Manias et al212 demonstrated 12.22 greater odds for the same phenomenon and Feldman et al found that each additional medication increased the odds by 8.7%.70 Similarly Wong et al183 found that if taking between one and nine medicines UID averaged one per patient and where taking between 10 and 19 medicines they averaged two discrepancies per patient. Linsky and Simon found that it was specifically commissions and duplications which increasingly occurred the more medications the patient is taking.173 In contrast, Nelson et al,126 Cornish et al50 and Climente-Marti et al189 found no link between the number of medicines taken and the number of discrepancies found. The reasons for this are unclear.

As well as the number of medicines, Pippins et al145 found the number of medication changes during the patients stay had an impact on the number of discrepancies; greater than or equal to six changes making an impact at discharge. Coffey et al197 also looked at the types of medication which, if a person is prescribed prior to admission, may increase the chances of a discrepancy; they found that both narcotics and antiepileptics increased the risk. As the number of medications a person is taking is likely to increase with age, 220 this would suggest a potential positive correlation between the number of discrepancies and the patient’s age, as well as the number of medications a person is taking. Several studies found no link with age.70,109,117,125,126,177,221 but others did.111,180,189,212 In the US, Linsky and Simon specifically found that patients aged over 65 were more likely to experience an omission.173 The

36 frequency of outpatient visits within the three months prior to admission also increased the risk of discrepancies with greater or equal to 13 outpatient visits having an impact and non-locally dispensed medication increasing the risk of duplication.173 No studies found a link between gender and the number or types of discrepancy,109,117,126,177,221 except Lau et al179 who determined that women were more likely than men to experience a discrepancy. Although this is a single study, the findings may have some validity due to women taking on average more medications than men.220 Nevertheless, further study would be required to demonstrate any clear association between the two variables.

No link between the number of discrepancies and the time50 or day180 of admission was found. But, if the patient was seen by the hospital doctor within an hour of shift change they had 3.7 times greater odds of experiencing a medication discrepancy.145 Moreover, for each additional minute wait to see a physician the odds increased by a factor of 1.01. Manias et al suggest various reasons for this including a potential link to high workload. They conclude that further work needs to be undertaken to try to explain wait times as a risk factor. Hellström determined that if the medication was reconciled more than four days following admission then there was also an increased risk. Although, where patients were reconciled on days four to eleven some admission errors had already been corrected.117

The patient’s usual place of residence was not a risk factor,109 yet, Hellström et al117 found that patients at home with no care support were at a higher risk of discrepancies, suggesting that living at home is not generally a risk factor but lack of carer or family support may have an influence. This mirrors other findings which suggest that living alone increases health risks, potentially meaning an increase in medication.143 Pippins et al found that a medication history taken from a family member or carer increased the risk of a discrepancy,145 which contradicts findings by Porter et al and Morrison et al who determined that carers played a pivotal role, these studies, however, were in paediatrics.88,222

The majority of studies have been conducted within one clinical specialty, therefore, few have been able to look at the relevance of specialty as a risk factor for patients experiencing a discrepancy. One study showed that surgical 37 patients were more likely than general medical patients to suffer a discrepancy,125 in contrast, Grant and Larsen195 found the opposite, however, their study was conducted in paediatric ICU. A further study looking at prescribing within psychiatry discovered that discrepancies occurred equally in psychiatric versus non psychiatric medicines, however, omissions were higher in non-psychiatric medicines 80% versus 20% in psychiatric medicines. Eijsbroek et al223 found that medication started in ICU to control acute symptoms were sometimes not stopped at discharge, including Proton Pump Inhibitors (PPIs), sedatives and psychotropic medication. Paradoxically, chronic therapy is often stopped205 or omitted at admission to ICU224 then not restarted at transfer or discharge. This is not unique to ICU, also surgical patients were at risk of having long-term therapy, especially warfarin, unintentionally discontinued due to not being re-started post-surgery225 and patients with end stage renal disease frequently experience high numbers of medication related problems specifically related to gaps in medication information at transitions in care.226

Prescribing for children at the interface between primary and secondary care, can be more complex than for adults as children often need unlicensed or ‘off- label’ medicines which present challenges with shared care arrangements and obtaining supplies post-discharge,227,228 although, prescribing for children at the interface has not been demonstrated as an increased risk factor for medicines reconciliation related error. Huynh et al found that only 67% of pharmacists in the UK would carry out medicines reconciliation in all children and that only a third had policies for medicines reconciliation in children.130

The evidence for length of patient stay as a risk factor is mixed, with Hu et al111 determining it as a positive predictor for increased likelihood of discrepancy and Nelson et al126 finding no link. The following factors have also been shown to increase the risk of discrepancies: visual deficit,212 the need for an interpreter 212 level of education,70 low patient understanding,145 known allergy,212 medication history taken by an junior doctor,145 and those patients who are eligible for benefits.212 The following do not appear to be risk factors: race,70,126 marital status,70 primary payer (the insurance company who pays the healthcare bill in

38 the US), 70 severity of illness,70 direct admission to the ward without transfer from another ward 117 and destination at discharge.212

2.3.8 Time taken to complete medicines reconciliation

Time is considered within the literature from two different perspectives. The first is the time between admission and the patient having their medication verified by a member of pharmacy staff. The second is the amount of time taken by the HCP to verify the medication history. The WHO1 and NICE & NPSA6 state that patient’s medicines should be verified by a member of pharmacy staff within 24 hours of admission to hospital, although the evidence base for this is unclear. There is little evidence which demonstrates whether the suggested timeframe is being achieved in practice, or studies which suggest interventions to accomplish this target. One UK study by Badham et al132 showed that 53% of patients taking anticonvulsant medication had their medication histories taken within 24 hours of admission. A further study, in 2008 from Kings Hospital London, showed only 55% of patients had medication histories taken, with 79% conducted within 24 hours.133 Similarly, Aljamal et al229 (UK) showed that approximately half of patients had their medicines reconciled, 80% within 24 hours.

The average length of time taken to verify a medication history varies within the literature from between 12.5 minutes up to 45 minutes where counselling is included at an average of 11 minutes (see table 2.5). The variation in time is potentially due to the patient’s knowledge and the complexity of their regimens230 or maybe due to the HCPs approach, although this has not been studied. HCPs have shown concern about the lack of adequate time within their daily routine to gather appropriate medication histories.230,231 Both studies, however, were based on nurse perception and it is unclear whether this is a reflection of other HCP opinion.

2.3.9 Sources of information used during medicines reconciliation

Various sources are used within a secondary care setting to try to establish an accurate medication history. These include GP medication list printouts, repeat prescribing slips, patient’s own drugs (PODs), communication with patients or

39 carers, previous discharge summaries, medicines administration record (MAR) charts from social and care home settings, community pharmacy records and monitored dosage systems (MDS). In 1995, Leape et al identified that the most common cause of medication error was a lack of access to accurate information about either the medicine or the patient.232 It can be difficult, especially out of hours or in an emergency, to obtain appropriate information.78 Previously the receipt of a list of current medications the person is taking from the GP, via fax has been untimely and sometimes has to be requested on more than one occasion,76 there is nothing to suggest this has changed.

Worldwide, many studies have looked at how to elicit appropriate information from patients through medication histories and have identified the sources of information which can contribute to accurate reconciliation, for example EHR or MAR charts.49,50,145,166,233-235 There is no one source more accurate than another. It is acknowledged that a range of sources must be used to verify the list of medication a patient is taking.76 Medication histories should be collected from the most recent and reliable source and where possible, information should be cross-checked and verified against another source.76,80 There is evidence, however, from Denmark to suggest that a second interview with the patient following the initial medication history, using GP records, added to the medication history in two-thirds of cases.121 The extent to which different sources are used is not widely published in the literature.

It is often wrongly assumed that the GP record is the most accurate source of information.76,78 GP referral letters can include incorrect information.78 Dutton et al found that over reliance on the GP letter was responsible for some but not all of the errors in medication histories.78 There were still prescribing errors in 52% of patients who had accurate GP letters, compared with 77% with an inaccurate or no letter, reinforcing the importance of using more than one source to verify a medication history.78 Collins et al found that 5% of medicines the patient was taking were found not to be on either the GP or nursing home record.76 Also GP doses were sometimes prescribed as directed, making it difficult to confirm the intended dose;76 although this was 10 years ago, there is nothing to suggest this has changed. It is also contrary to national prescribing guidance.236

Table 2.5 Time taken to conduct medicines reconciliation per patient in hospital

Author Country Time Medication History and/ or Reconciliation Coffey et al Canada Mean 18 mins (2-75mins) Medication (2009)197 More than 4 medications took History and Reconciliation mean 31 mins (range 10-75mins)

Cornish et al Canada Mean 24 mins Inter-quartile Medication (2005)50 range (IQR) (20-30mins) History and Reconciliation

Feldman et al US Mean 11.2±8 minutes Medication (2012)70 (medication history) History and Reconciliation

29.3± 30.2 mins (Reconciliation)

Geurts et al Netherlands Mean 45 mins (inc 11mins Medication (2013)200 counselling) History and Reconciliation

Kent et al Canada Mean 23 mins Medication (2009)237 History and Reconciliation

Lessard et al US 30 mins per patient including Medication (2006)194 clinical check History Only Miller et al US >80% medication history took Medication (2012) 184 less than 15 minutes History Only Quélennec et al France Mean 36.4±12.1 Medication (2013)238 History Only Remtulla et al Canada Mean 12 minutes (range 2- Medication (2009)239 48minutes) History Only Vira (2006)166 Canada Median 15 mins IQR 10-21 Medication History and Reconciliation

Collins et al concluded that ‘currently there is no ‘gold standard’ medication history available other than a list of medicines taken from the patient who is deemed reliable.’76 Discrepancies in GP EHRs are not unique to the UK. Orrico in the US commonly found medication recorded in the EHR but not taken by the patient or conversely not recorded in the EHR but taken by the patient; most commonly analgesia.240 More recently Schmiemann et al found that 41% patients in Germany were taking at least one medication not on the GP health record, in 90% of cases for one of the following reasons: the GP had knowledge of what the patient was taking but it was not documented, the medication was documented in the history but not in the medication list, or it was an OTC medication therefore not prescribed, the medication was issued by a specialist.241 This mirrors other findings by Tulner et al (Netherlands) and Staroselsky et a (US).242,243

Green et al warns that patients may potentially have up to four lists of medication (including repeat prescription lists, GP letters, GP printout summaries, patients own lists, MAR charts and specialist service records), emphasising that they should be interpreted with caution and assessed in conjunction with other sources of information, especially the patient and/ or carer.244 Green et al suggests an IT solution. In Denmark, Foss et al found that a median of one to three discrepancies per patient were identified in seven paired source comparisons including primary healthcare records, hospital departments, health insurance data and patient interviews. Full agreement was only found in 8% of patients.245

Numerous factors can affect the validity of the information source.246 The age of the source, especially if greater than one year, corroborating the need to ascertain the date of the source. The presence of acute and ‘when required’ medications decreased the accuracy. The researchers also found that discharge information was more likely to be valid than outpatient information.

In many instances it is the patient who provides relevant information,78 however, some authors are of the view that their reliability can be unclear, despite this not having being studied in depth.76 It is almost 30 years ago since Price et al suggested that patients should be encouraged to bring in their medicines on admission to hospital.247 This was reinforced in Pharmacy in the future: 42 implementing the NHS plan.248 Over the years this has improved and there has been a recent drive in many NHS Hospitals to encourage patients to bring their own medication with them on admission.249 The role of the patient in medicines reconciliation will be discussed further in Part 3 of this literature review.

Other sources used to verify the list of medications a patient is taking include the community pharmacy patient medication record (PMR), MAR charts and MDS. Whilst it is unclear the extent to which these sources are used in the UK, Wilcock and Lawrence107 found that the community pharmacy PMR would have added value to the medication history in 4% of cases and some value in up to a third of cases. This study only looked at systemic medication, rather than topical which has been recognised to be prescribed inappropriately during hospital stays.107

Internationally some countries have central repositories which host medication lists e.g. Sweden, Denmark, Canada and Ireland.250-253 These national databases have been shown to be relatively accurate and contemporaneous sources of information which can add value in approximately 10 to 28% of cases (up to 1.1 medicines per patient).251,252,254 Although useful these have also been shown to contain inaccuracies in information in some cases.119,250 The databases are limited where certain medicines are not contained in them, for example, medicines which are prescribed and dispensed in hospital e.g. cancer and tuberculosis (TB) medicines, OTC medicines and CAMs. This can depend on the country as community pharmacy records in Denmark were shown to add value especially for the identification of OTC medicines and CAMs.254 This would not be the case in the UK as OTC and CAMs are not routinely recorded within community pharmacy. Like all EHRs, national databases rely on accurate data entry and being kept up-to date. Inaccuracies in health records can be of significance especially where there are potential interactions.255

In countries where private healthcare is more mainstream the number of different physicians a patient sees may be increased. Balon and Thomas showed that 41% of patients were seeing three or more doctors.256 The advantage of countries with health insurance is that the claims databases often contain complete information. Where they were found to lack information is 43 where the patient did not get reimbursed or bought their own OTC medication or CAMS.147 Conversely, Olsen et al showed incongruence between the claims databases and the primary care provider medication lists.257 This is not as significant an issue in the UK as patients are registered with one general medical practice, however, on occasion they may be seeing multiple hospital consultants.

In a UK study by Oborne and Luzac, 64% of patients used OTC/CAMS medication, however, only 5% of these were recorded in the medication history at admission to hospital.149 This is supported by Cockayne et al who showed that 58% patients took CAMs prior to admission with only 28% being documented in the medical record.258 The recording was improved through education of HCPs. In the US, Holmes et al have gone further to identify a list of herbal medicines and supplements as a useful tool to identify potentially harmful medication (specifically in the hospice setting).259 There is nothing in the literature which looks at the use of illicit substances and whether this should be included as part of the medication history. In a letter to the British Medical Journal (BMJ), Kinney points out that illegal drugs should be considered in the differential diagnosis as they can cause derangement of Liver Function Tests (LFTs) and other symptoms.260

The NPC divides sources into those which are more reliable and those which are less reliable, stating GP medication list printouts, repeat prescribing slips, PODs, patients or carers, and previous discharge summaries to be more reliable than MAR charts from social and care home settings, community pharmacy records, single assessment process documentation and MDS for gaining a true account of a patient’s medication.80 It is unclear how these have been categorised as there is insufficient UK evidence to suggest that any one source is more reliable than another when used in isolation, and although items like the MAR and MDS may not include medication which the patient’s GP thinks they are taking, it is more likely to be a true reflection of what they are actually taking. This is contrary to findings by Fitzimmons et al in Ireland who found that community pharmacy and GP staff were the most available and accurate sources of information compared with the patient/carer, PODs, GP

44 letter, GP surgery, community pharmacy (CP), past inpatient Kardex and discharge summaries, nursing home lists and staff.140

With improved IT systems there are a number of studies from the US which look at the accuracy of computerised PAML builders as a source of information although the accuracy of these can also vary. Peterson et al found that most commonly they included false positive entries i.e. the patient was not taking the medication despite it being listed.141

2.3.10 Impact of medicines reconciliation on readmissions to hospital

The NPC suggests that accurate medicines reconciliation can lead to a reduction in re-admissions caused by harm from medicines, however, there is little evidence to support this.80 No single intervention alone has been shown to reduce the risk of rehospitalisation within 30 days, however, several common interventions have been studied as part of a ‘bundle’ of care, including patient education, medicines reconciliation, discharge planning and follow up.67,100,261 Most commonly these include discharge education and counselling with follow up.67,87,100,105,123 Lack of meta-analysis and inadequate description within the studies does not allow the success of any single intervention to be determined (see also section 2.3.1).

Al-Rashed et al showed twenty fewer readmissions in the study versus control group following pharmacist intervention, which included pharmacist counselling, provision of discharge summary plus reminder chart together with a follow up visit.100 The role of the individual component of the intervention is unclear, or whether it was a combination of all four aspects which produced the effect. Community liaison pharmacists, who facilitate pharmaceutical care during the transition between hospital and home (including medicines reconciliation at admission and communication at discharge), have also been shown to reduce readmission rates from 8.8% to 6.4%, producing cost savings of £83,484 per annum based on an average of 7.7 days stay per patient.75 A similar model has been demonstrated to work in the US262 although the actual cost savings were not highlighted within the study. Such interventions are supported by a model- based review paper showing that they are a cost-effective use of NHS

45 resources, with pharmacist led medicines reconciliation being the most-cost- effective.263 This is especially important since the introduction of the 30 day readmission policy where hospitals may not be reimbursed for readmissions within 30 days of discharge.264 Estimated costs of ADEs range from 900 to 1800 Euros.263 However, it is difficult to know how many discrepancies go on to cause actual harm.125 There are few studies which have undertaken cost analysis on medicines reconciliation or associated interventions to determine their cost effectiveness.

Conversely, in the US Hansen et al found no link between the presence of an accurate medication history at admission and readmission rates nor a link between the timeliness of receipt of discharge and rates of readmission.265 This is supported by a Swedish study that also found no impact on readmission following accurate medicines reconciliation.266 Similarly, in the UK, Nazareth et al found a comprehensive discharge plan had no effect on re-admission rates.105

2.3.11 Improving patient safety and measuring quality

The international drive to improve medicines reconciliation has led to the implementation of several patient safety programs, different suggestions of indicators which may demonstrate quality within the process and a range of standards.6,7,43 It comes at the same time as many countries promote and implement general patient safety programmes to ameliorate avoidable harm.267- 269

The international High 5’s project, a WHO initiative supported by health ministries focuses on five distinct patient safety topics including medicines reconciliation.268 Canada led the development of a guide to implement standard operating procedures for medicines reconciliation which has been trialled in participating countries.5 Through this campaign the Netherlands has shown that inaccuracies in a patient’s current medication history can be reduced by 90% within one to five months of the introduction of a toolkit to improve medicines reconciliation. However, it is unclear exactly how changes were implemented and whether any improvement methodology was used. Similarly, the IHI in the US previously developed a ‘how-to-guide’ to support the reduction 46 of medicines reconciliation error as part of their 5 million lives campaign.2 Its success in terms of medicines reconciliation is unclear, however, it did raise the profile of medicines reconciliation internationally.

Relevant indicators must be used to identify whether the changes to medicines reconciliation being made in practice are leading to improvement.80 This allows providers to feel their efforts are worthwhile and gives commissioners data to support informed decisions. Various audit tools for medicines reconciliation have been developed by the NPSA (UK) and IHI (US) to help support improvement.2,6 Also in the UK, Aljamal et al has generated ideas for quality indicators for medicines reconciliation on admission, using a group of practising pharmacists in Manchester, however, this was based on pharmacist experience, not on evidence and the quality indicators were not detailed.270 The NPC suggests that organisations should use procedural measures, such as the presence of written policies and procedures, and the percentage of patients that have had their medicines reconciled within 24 hours.80 As well as procedural measures, they also suggest a further number of impact measures, such as the number of medication discrepancies in patients whose medicines have been reconciled. This mirrors terminology used within patient safety which classifies measures in terms of structure, process and outcome.141 There is currently little in the way of structure, process and outcome measures for medicines reconciliation. This study will look at defining appropriate standards that potentially could be used as measures within the process.

Many countries have set national targets for medicines reconciliation. Within the UK, NICE & NPSA issued joint guidance which specified that all healthcare organisations must put policies in place for medicines reconciliation at admission to hospital.6 The policies must specify standardised systems for collecting and documenting information about current medications, and should ensure that pharmacists are involved in medicines reconciliation as soon as possible after admission. The responsibilities of pharmacists and other staff in the medicines reconciliation process should also be clearly defined. It also stipulated that strategies to obtain information about medications for people with communication difficulties be incorporated. NICE & NPSA agreed with the WHO that 24 hours is an appropriate target for organisations to include in their

47 policies.1,6 The US have also specified standards on medicines reconciliation within their NPSGs since 2005. Since then the NPSG for medicines reconciliation have slowly evolved. The Institute for Safe Medication Practices (ISMP) in Canada chose medicines reconciliation as one of three medication safety indicators to prioritise for reporting, recognising that in some ways it is different to other medication safety indicators such as venous thromboembolism (VTE) prophylaxis and antibiotic prophylaxis as it focuses on overall system integration rather than a specific clinical practice,therefore not as easily linked to patient outcomes. Methodological difficulties also occur with capturing standardised data through which to compare different organisations e.g. common definitions, and appropriate numerators and denominators.271

2.3.12 Human Factors and their role in Medicines Reconciliation

The study of human factors in relation to medicines reconciliation is not previously well explored. Where it is, it looks at human factors in relation to the design of IT systems to enhance medicines reconciliation.272,273 Monkman et al determined that the advantage of involving IT into medicines reconciliation is that process standardisation and specific constraints can be embedded into the system to ensure steps are performed sequentially, in a timely manner and that all tasks are completed fully to eliminate variability.273 Also important are the workflows and roles and responsibilities associated with the process. Lesselroth et al used human factors specifically to review medicines reconciliation workflow to determine how patient-centred technology should be designed to collect a medication history.272 They identified system themes and design constraints that influence the quality of the medicines reconciliation process to be able to subsequently design new technology to support medicines reconciliation.

2.3.13 Summary Part 1

A vast amount of literature has been published on the subject of medicines reconciliation. These studies thoroughly explore the scale of error and discrepancies associated with inadequate medicines reconciliation, especially in the US; less studies look at the scale within the UK or interventions which have been implemented to reduce medicines reconciliation-related errors and/or 48 discrepancies. The lack of clear definition, taxonomy and measures makes it difficult to compare studies or for organisations to benchmark their progress. Various models exist which summarise the medicines reconciliation process, all of which contain the same main steps; collection of medication history, verification of the medication prescribed, documentation of changes and communication to the next provider.

Although there is consensus that understanding human factors are vital to reducing human error and harm to patients,22 there are few studies which look at the role of human factors in the medicines reconciliation process. Neither are there any empirical studies of the safety of the entire process when undertaken as part of routine practice. It is evident that to reduce the incidence of medicines reconciliation-related errors and/or discrepancies, a structured systematic approach to medicines reconciliation is needed to ensure an accurate and detailed medication history.274,275 The next section of the literature review will go on to explore medicines reconciliation-related processes and interventions which improve the accuracy and safety of medicines reconciliation, taking into account the HCP role in the process.

2.4 Part 2 - Medicines Reconciliation Process and HCP Involvement

Medicines reconciliation relies on the patient being subject to many small interdependent processes, for example, medication history taking on admission, prescribing, transcribing IMCs, counselling at discharge and processing discharge information, all of which feed into a larger overall process. The NPC identifies the movement of the patient through the healthcare system as the ‘patient’s medicines management journey’ (p1) and although the NPC depict the patient as the ‘end-user’ (p8), the commonality between each different stage is clearly the patient, making it logical to describe the process flow in terms of the patient journey through the healthcare system.80 The former Royal Pharmaceutical Society of Great Britain (RPSGB) illustrated the start of the patient’s medicines reconciliation journey as the admission to hospital, whether planned or in an emergency, via A&E or direct to a ward. 276

The WHO suggests that healthcare organisations should assign clear roles and responsibilities to HCPs for all steps in the medicines reconciliation process to ensure accountability, however, the WHO do not make any suggestion as to what these roles and responsibilities should be and to whom they should be assigned.1 The NPC goes further to suggest personal, professional and corporate levels of responsibility, highlighting key skills needed for effective medicines reconciliation which include effective communication, technical knowledge of processes and therapeutic knowledge.80 They also recommend that levels of staff competency should be agreed, and the responsibility for carrying out medicines reconciliation should ideally be included in job descriptions and personal development, endorsing it as a multidisciplinary effort. The NPC stipulates that there needs to be enough competent people available to undertake medicines reconciliation at all times and that restricting the responsibility for medicines reconciliation to one professional group could make it difficult for some organisations to provide a reliable service, especially ‘out-of-hours’ provision.80 The level of staff competence and availability within current practice is unknown. One previous review has looked specifically at pharmacist and IT interventions but no other HCPs nor stages of the process outside hospital.59 The review determined that successful intervention included intensive pharmacy staff involvement and targeting interventions to a high-risk patient population. 50

The earlier section within the literature review (Part 1) identified the scale and severity of the problems inherent in medicines reconciliation. This section (Part 2) goes on to identify the interventions which have been shown to improve medicines reconciliation at different stages of the patient journey within the hospital (admission, during stay, discharge) and primary care, taking into account HCPs’ roles.

2.4.1 The medicines reconciliation process

The complexities associated with the medicines reconciliation process are a result of the number of people and care settings involved in the process.61 There are few papers that provide evidence or even state how medicines reconciliation should be conducted or how a medication history should be taken leaving it to national bodies to provide guidance and organisations to implement local policies and procedures. There are also few papers that study the medicines reconciliation process as a whole or the role of the entire multidisciplinary team (MDT). The majority focus on isolated aspects of the process or individual HCPs. Various diagrammatical representations of the patient’s journey through hospital with respect to medicines reconciliation exist with varying degrees of detail (see figures 2.3 and 2.4). All include three distinct stages: admission, during stay and discharge. This section of the review will discuss the process at each of these stages, continuing to discuss post- discharge, within primary care. In practice, within the UK, many medicines reconciliation tasks are conducted uniquely by one individual HCP, hence the role of each individual HCP will be compartmentalised and considered at each stage. It is sometimes difficult to isolate interventions specifically related to medicines reconciliation from overall packages of care, which can incorporate other interventions such as counselling and follow up (sometimes referred to as bundles of care). Interventions relating to the patient have also been incorporated into the table for comparison. These will be discussed separately in Part 3. There are some areas where medicines reconciliation is important which do not exactly fit the patient journey through hospital into primary care, for example, in outpatient clinics. There are various studies which also demonstrate the value of the pharmacist at specialty outpatient clinics; oncology,277 cardiac,278 and Parkinson’s disease,279 all of which include improvement to medicines reconciliation. 51

Figure 2.3 Algorithm for medicines reconciliation from admission

through discharge160 (reprinted with kind permission)

Some studies have redesigned the medicines reconciliation process to improve quality and decrease medicines reconciliation error103,280-282 or utilised management techniques to improve medicines reconciliation,283-286 although only one within the UK.287 Methods such as Failure Modes and Effects Analysis (FMEA),283,285 Plan, Do, Study, Act (PDSA) cycles,161,287 Lean methodology284 and Six Sigma286 have been used. One quality improvement programme in the UK, led by the prescribing observatory for mental health, to improve the uptake of medicines reconciliation and number of sources used, showed modest improvement increasing the percentage of medicines reconciliations undertaken from 71% to 79% at re-audit with an increase in interventions from 25-31%.10 This involved nationally supported implementation of audit, detailed provision of audit results, action planning and re-audit. All service improvement studies which focus on the use of management techniques have been conducted within secondary care and are lacking within primary care.288 Although this is a deficiency in the research evidence, it is at least clear that successful leadership supplemented by local champions and evidence based tools, are imperative to ensuring a robust medicines reconciliation system.97,281

Figure 2.4 Diagram adapted from the WHO High 5s campaign by

Fernandes et al289 (reprinted with kind permission)

2.4.2 Admission to hospital

2.4.2.1 Background

Patients can be admitted to hospital via one of three main routes; via A&E, GP referral or through an outpatient clinic appointment (elective and non-elective admission). These three paths are different in approach, however, irrespective of the path taken, the patient will be admitted either through an admissions ward or direct to a specialty ward where a medication history will be gathered, usually by a doctor. Approximately 20% of patients presenting at A&E will be admitted to hospital.290 Following arrival at A&E the patient goes through a variety of steps outlined in figure 2.5. Although this model is based on a study from the US, the system is mirrored in the UK. Here, a medication history is usually taken in A&E, although the main goal of emergency medicine is to stabilise the patient. A&E staff members are predominantly interested in determining those medications which may be having an acute impact on the patient’s state, leaving a more complete and accurate medication history to be conducted once the patient is stabilised, usually on an admissions or specialty ward (Wood, T. A&E Matron. Bradford Teaching Hospitals NHS Foundation Trust. Personal communication. 15th April 2011). Where the patient’s admission is planned they will often be seen within a pre-admission clinic where their medication history will be documented. Sagripanti et al suggests that given the large number of changes which can occur in the time between pre-admission assessment and admission the chart should not be written in advance nor the medication supplied; although, it is unclear whether this is happening in practice.291

At some point following admission the patient may have their medication history verified, usually by a member of pharmacy staff. The majority of intervention studies focus on admission. Several interventions have shown improvement in medicines reconciliation at admission which are not specific to a particular HCP; most commonly a medicines reconciliation form to accurately document the medication the patient is taking prior to admission (see table 2.6) 63,64,292,293. They all showed a reduction in discrepancies at admission. Most, however, were undertaken in specific specialties e.g. paediatrics, making their general applicability to other specialties unclear. Educational interventions which either reinforce the process or provide feedback on discrepancies have also been

54 shown useful,63,92 as well as interventions which facilitate the gathering of medication information at admission e.g. the use of pre-visit questionnaires294 and aide-mémoires.295

Figure 2.5 Workflow for medicines reconciliation in A&E284

Patient arrives in A&E and checks in

Patient waits in waiting room

Triage nurse evaluates patient

Patient waits in waiting room

Nurse takes patients to treatment room and clinically assesses

Doctor assesses patient and makes clinical decision

Tests performed/ Treatment administered

Discharge/ transferred

Table 2.6 Summary of interventions relating to the patient’s admission to hospital categorised by HCP and including those interventions carried out by the whole team and the role of the patient

Country of Nurse Hospital Pharmacy Staff CP Hospital Patient Ambulance MDT Study Doctor Personnel

UK None Pharmacist reviews patient at Community Recognition of Bringing PODs None None admission185 pharmacist used performance into as a source of with medicines hospital65,78,298 Medication history verified by information 107 reconciliation 297 Pre-admission hospital pharmacist questionnaire294 73,74,78,82,107,296

Hospital pharmacist writes inpatient chart 66,79

Pharmacist takes medication history 66

International Ward nurse Pharmaceutical care planning Community Allergy form Pt completed Bringing in Standardised handover 299 for hospital stay pharmacist used 150 questionnaire POD 333,334 medicines 73 as a source of 330*331 reconciliation Education information Medical student form 63,64,292,293 and feedback Verification of the medication 327 education on Improving on accuracy history by a pharmacist medication patient Report of medicines 49,71,118,122,144,162,208,277,278,300-320 history taking knowledge medicines reconciliation 93,96,328 about medicines reconciliation 298 Education by the pharmacist at to improve discrepancy92 point of verification118 Real time recall332 feedback and Standardised Surgical pre-admissions education 329 Counselling question list as pharmacist124,321 patients to an aide decrease memoire295 Adherence check312 medication errors82 56

Country of Nurse Hospital Pharmacy Staff CP Hospital Patient Ambulance MDT Study Doctor Personnel

Medication histories taken by Second a technician 138,239,313,322 Bringing PODs interviews and into outpatient GP lists as Pharmacy students taking appointment65 sources121 medication histories71,323-326 Medicines Educational tool to teach reconciliation pharmacy students to take a to reduce medication history 95 hospitalisation 266* Pharmacist writing medication chart210 Medicines reconciliation 335

*no significant effect seen following intervention

2.4.2.2 The role of the doctor at admission to hospital

Hospital doctor

At admission the junior doctor will take an account of the presenting complaint and past medical history from the patient, including any medication they were taking prior to admission. This is known collectively as clerking. Once a medical and medication history is obtained, the doctor makes a decision in the context of the presenting complaint regarding which medication must be prescribed for the patient during their stay. This includes both acute and regular medication. Few UK studies have looked at the role of doctors in the medicines reconciliation process. More frequently medicine reconciliation studies investigate doctors’ accuracy in medication history taking compared with other HCPs, namely pharmacists. These studies, however, have been conducted in the US rather than the UK.68,208,301,302,309 A study conducted in St Mary’s Hospital Portsmouth by Dutton et al78 found that approximately 40% of doctors felt their accuracy when taking medication histories was between 50 and 75%, with 17% feeling they were less than 50% accurate. In Nigeria, Yusuff and Tayo found that the doctors specialty also affected the recording of the medication history in the patients’ notes with doctors recording more frequently and in greater depth those medications that may aid the diagnostic task.311 These findings are similar to those of Ziaeian et al, in the US, who determined that medications relevant to the primary diagnosis were more likely to be accurate.336 Nigerian doctors also felt that more emphasis should be placed on the medical history rather than the medication history.337 It is unclear whether these findings are applicable in the UK and it would be worthwhile conducting similar studies.

In an attempt at improving the accuracy of medical prescribing generally, the British Pharmacological Society have identified learning outcomes which should be incorporated into the undergraduate medical prescribing curriculum using a Delphi approach.338 Many of these outcomes were associated with medicines reconciliation. The evaluation or the extent to which these have been embedded and/or adopted into the curriculum, however, is unclear. Internationally, a variety of different educational interventions have been evaluated to improve medication history taking amongst medics. In the US,

58 curricula which specifically focus on medicines reconciliation at transition in care and medication history taking have been implemented which led to improvements in confidence, knowledge and skills.93,96 Simulations for second year students to improve medication history taking have been piloted in Chicago328 and an educational intervention programme in New Zealand, where reconciliation errors found were fed back in real time.329 The American Medical Association has produced a guide for the role of the doctor in medicines reconciliation, it highlights issues, strategies and safe principles on implementation of good processes, stress that the doctor should take the lead in embedding safety into the medicines reconciliation process.339 This contradicts guidance from the APhA and the UK NPSA which states that pharmacists should take the lead.6,340 One study by Peyton et al showed that although educational intervention of doctors and nurses increased the accuracy of medicines reconciliation, it only improved from 14.4% to 18.9%.118

The lack of interventions which focus on improving medication history taking amongst doctors is surprising given the rate of error reported in studies.77,109,125,145,166,171-183,201 Those interventions which do exist all aim to improve the doctor’s medication history taking accuracy.150,297 One of these interventions focuses on standardisation using an allergy questionnaire to improve the structure of consultation and standardise the information gathered.150 The second demonstrates that feedback, then praise following good performance can reduce the number of discrepancies at admission.297 It would appear that further work is needed to establish the role and responsibilities of doctors plus effective interventions aimed at doctors to improve medication history taking.

General practitioner (GP)

There is little evidence looking at the role of the GP or their practice staff in communicating about medicines at admission to hospital. Often general medical practices are contacted when one of their patients is admitted to hospital to determine which medication they are usually prescribed in primary care. There is evidence to suggest that the information held on the GP EHR is not always up-to-date or accurate (as discussed in section 2.5.11) therefore non-contemporary information is sometimes communicated. Where patients

59 are referred by the GP to hospital and are admitted directly to a ward the GP will send a referral letter with the patient. There are no studies that look at the accuracy of the medicines related information within these letters within the UK, however a study in Zurich shows high numbers of omissions in admissions letters, for example, 12% patients had their medication information missing and less than 50% of patients taking more than four medications had any directions stated.341

2.4.2.3 The role of hospital pharmacy staff at admission to hospital

Currently within the UK pharmacy staff are predominantly involved in verifying medication histories taken by doctors (and occasionally nurses) post admission. (This has previously been described in Part 1 – see figure 2.2). This does not happen in isolation and will usually go hand in hand with other clinical checks of the patient’s IMC. Thus errors discovered may be amongst a number of different error sources outwith reconciliation. The activities undertaken by pharmacists within the hospital setting are not unique to the UK and are similar to that of Ireland,342 the US343 and Australia.344 The number of patients seen by pharmacy staff varies from hospital to hospital, with some potentially not having their medication history verified throughout their whole hospital stay. Articles in the Pharmaceutical Journal dating back to 1982 suggest that the ward pharmacist is the best placed to elicit the medication history from the patient.345 There is a wealth of evidence both within the UK39,66,74,296 and outside49,69,71,118,122,124,162,208,237,239,277,278,300-321,346-348 which demonstrates the effectiveness of pharmacists and pharmacy technicians undertaking medicines reconciliation and improved medication history taking in comparison to doctors 162,208,309,349,350 and nurses.208,301 There are also suggestions from a range of international papers that pharmacists writing IMCs on admission can improve patient care.208,210,211,303,306,308,351 This is more limited within the UK. 66,79

A substantial body of evidence also demonstrates the benefit of pharmacists and pharmacy technicians conducting medicines reconciliation within A&E.69,208,302,303,306,308,314,318,347,348,352 A small survey carried out in 2010, in a specialist workshop assessing the role of pharmacists within A&E found that most UK A&E departments had some level of pharmacy service. Roles were similar across all A&Es, and included guideline development and review,

60 training and preparing Patient Group Directions (PGDs).353 Less than half of these pharmacy roles, however, included medication history taking, with only 8% of respondents writing IMCs.353 Another UK study, in the same year, looking at the role of a pharmacist independent prescriber in A&E, demonstrated decreases in prescribing errors and an associated reduction in potential and actual harm.230 It showed that moving the pharmacists’ work (medication history taking and initiating the in-patient chart) to A&E, transfers their input to the start of the patient journey and applies their knowledge and skills to the patient’s benefit.66 In contrast, nurses have expressed discomfort with eliciting a medication history at triage in A&E.

Kwan et al124 and George et al321 also looked at the role of pharmacists within surgical pre-admissions clinics in the UK, demonstrating increased quality and decreased discrepancies. Pharmacists have also demonstrated skills in other areas which contribute towards improved communication about medication at care transitions including the production of pharmaceutical care plans at admission (US).73 Further work in all specialties to establish the impact of pharmacists taking the initial medication histories and writing IMCs at admission is needed, although both pharmacists and nurses have been shown to make clinically appropriate prescribing decisions.354

The majority of evidence relates to pharmacists although a small number of studies provide evidence which promotes the use of support staff such as pharmacy technicians138,239,313,315,355,356 and pharmacy students.71,323-326 Increasingly, within the UK, pharmacy staff, other than registered pharmacists, play a role in the medicines reconciliation process. They are familiar with the dosage forms and strengths and usual dosing schedules of a wide range of medicines.239 They may undertake roles such as medication history taking and patient counselling. There is no UK evidence which explores the role of these pharmacy staff within the process, however, evidence from outside the UK suggests that pharmacy students and pharmacy technicians can play a valuable role in identifying drug-related problems, allergy information and admission discrepancies which equal that of the pharmacist,138,322,324,325,357 not only on the wards, but also in pre-admission clinics322 and within A&E356 respectively. Through changing the focus of the pharmacy service, pharmacy staff can be

61 better utilised in the medicines reconciliation process.103 Remtulla et al suggests that pharmacist expertise is better utilised in assessing the appropriateness of therapy rather than technical tasks of compiling the BPMH.239 Guidance from NICE & NPSA6 suggest involvement from both pharmacists and technicians. Their guidance is specifically targeted at admission; however, there is no reason to suggest that this is not applicable to other areas. As the roles of pharmacy staff change, more evidence is needed to determine how best to utilise their skills in practice.

2.4.2.4 The role of nurses at admission to hospital

The role of nurses in medicines reconciliation in the UK has not been examined although it is known that, where admission is planned, nurses often undertake medication histories at surgical pre-admissions clinics although the accuracy is unknown.358 Outside the UK, especially in the US, nurses more commonly take medication histories at admission to hospital.208,301 A survey of nurses, conducted in Canada, identified that clearer roles and responsibilities would improve medicines reconciliation together with improved communication at handover.231 In interviews conducted by Candlish et al nurses expressed their desire not to take a medication history at admission suggesting an appreciation for the help they received from pharmacists.230 Good practice guidance on medicines reconciliation in hospital produced by NHS Greater Glasgow and Clyde suggests that nurses maintain an overview of whether patients on their ward have had their medicines reconciled at admission, prompting relevant HCPs to take action as required; administer medication as directed on the admission chart; incorporate medicines reconciliation into handover procedures when patients transfer care and at discharge and where the prescribers are nurses ensure that prescriptions are accurately and legibly recorded.359

2.4.2.5 The role of the community pharmacist at admission to hospital

Studies both within the UK107 and internationally327 have shown that the community pharmacist has a role in providing information about the patient’s medication regimen taken within primary care. In Canada, Dersch-Mills et al determined that 24% paediatric patients could have had their medication clarified by a community pharmacy; the rest were either not on any medication or used multiple pharmacies.139 They also determined that following an 62 informed interview community pharmacy records were the most accurate source followed by the provincial prescription database. However, often the pharmacist is unaware that the patient has been admitted to hospital.360

2.4.2.6 The role of ambulance staff at admission to hospital

For those patients who come to hospital via ambulance the ambulance staff can play an important role in bringing the patient’s medication in with them to hospital. They can be involved in ‘green bag schemes’, where green plastic bags are used by paramedics to bring medication into hospital.361 Although such schemes have not been evaluated in the UK other than locally,249 they have been shown to work in Australia333,334 and have recently been backed internationally by the WHO.1

2.4.3 During the patient’s stay

There are a number of tasks which occur during the patients stay which may impact on medicines reconciliation. These include prescribing, altering medication, transcribing and ward rounds. There are few studies which look at medicines reconciliation interventions during this stage, the only interventions relating to pharmacists.

2.4.3.1 The role of the hospital doctor during the patient’s stay in hospital

During the patients stay the patient’s progress is reviewed by a doctor, usually on a consultant-led ward round. Medication may be withheld, stopped, started or changed. This usually occurs following discussion on the ward round and relevant amendments are made to the prescription chart by a relevant prescriber, usually a doctor although this may be a non-medical prescriber. Any amendments to existing medication regimens must be documented and communicated appropriately to ensure medicines reconciliation can be conducted at transitions in care. There were no studies which investigated the proportion of medications stopped, started and changed or the documentation of medication changes using the search terms.

It is likely that a patient will have their medication transcribed at some point during their hospital stay, usually by a doctor. Within the acute setting, transcribing is undertaken when medication is copied onto an IMC from a 63 patient’s medical record, when the in-patient chart is re-written because the administration records are full or when discharge information is copied from the IMC to the discharge advice summary. There is currently no widely accepted definition of transcribing. Hobson and Sewell have described it in their study as ‘the process where a list of drugs that has been prescribed by a doctor is copied from one chart to another chart or prescription’.362 Within their study they were specifically looking at pharmacist transcribing, however, it is recognised that it may include other HCPs, such as doctors and nurses. It is recognised that transcribing carries a substantial risk of errors,78,177 with a suggested overall error rate of 1% in the UK associated with rewriting charts and writing discharge prescriptions.52 This is significantly lower than a more recent Australian study which demonstrated approximately 12% errors in transcribing, the majority being omissions, although this specifically looked at transcribing at discharge.363 Transcribing errors may go unnoticed, then transfer to the next care setting. A study from Denmark demonstrated that transcription errors affected all main groups of medication equally and were not associated with age, gender, or number of medications. The applicability of these international studies to the UK is unclear. With such a low number of pharmacists transcribing discharge notes and guidance from the Nursing and Midwifery Council (NMC) stating that nurses should only transcribe in exceptional circumstances,364 it is left to the doctor to transcribe in the majority of cases.

2.4.3.2 The role of pharmacy staff during the patient’s stay in hospital

The pharmacist will conduct regular clinical checks on the IMC during the patients stay. The only interventions found which related to the patient’s stay, involved pharmacists either attending ward rounds or transcribing charts. Within the UK, hospital pharmacists have been shown to add value, both to the MDT78,88,102,185 and during the post-take ward round (PTWR),74,188,296 making significant interventions, providing cost savings, reducing errors and preventing major incidents. Improved working relationships have overcome previous impressions of pharmacists being prescribing critics and have led to the pharmacists playing a more proactive role in medicines management.102 This is supported by doctors who have proposed that pharmacists should always verify medication history and should be available on the ward round to advise on

64 medication problems.78 Pharmacist attendance on ward rounds has been shown to save money through interventions made in the US.365

2.4.3.3 The role of the nurse during the patient’s stay in hospital

Medication is administered to the patient during their stay by nursing staff and the IMC annotated to signify that a dose has been given. When medication is accidentally omitted from a patient’s prescription chart there is no indication to the nurse that it is required. There are few studies which look at nurse involvement in medicines reconciliation related processes. One study by Maxson et al299 demonstrated that bed to bed handover increased communication about the patient’s medication regimens. Further work on the knowledge, roles and responsibilities of nurses in the medicines reconciliation process, especially within the UK, is needed.

2.4.4 At discharge from hospital

The decision to discharge a patient is usually made during the consultant ward round. Following this decision, the patient’s discharge medication is prescribed on a discharge advice summary, transcribed from the patient’s medication chart and a copy sent to the patient’s GP. Junior medical staff have previously been reported to prepare nearly all discharge prescriptions with hospital pharmacists adding value to the process.88,366 It is unclear how much discharge prescribing by hospital pharmacists has increased since the introduction of pharmacist prescribing in 2006.367 At discharge patients are counselled on their medication to varying extents. A review by Spinewine et al87 showed that education and patient counselling reduces the risk of ADEs and readmission in some but not all studies.

Discharge communication has been cited as a problem for nearly 35 years.40- 42,368 Despite a number of improvements, for example, the introduction of discharge advice summaries (a discharge note that doubles as a prescription for medication),369,370 there are still a number of outstanding issues. In the 70s and 80s, Price et al247 Parkin et al368 and Harding42 identified that changes to medication during the hospital stay are not communicated effectively to either the GP or the patient. This was reinforced by Munday et al371 and Brackenborough,41 who determined that HCPs, in primary care, would like 65 information on reasons for medication changes to enable good continuity of care, suggesting that the discharge prescription be modified to include reasons for these changes.371 A more recent review, in 2007, by Kripalani et al372 show that communication has not significantly improved and there are still deficits in communication and information transfer at hospital discharge that can adversely affect patient care. There have been on-going suggestions that communication can be improved through the use of technology373 (see section 2.2.6 Role of Information Technology (IT)).

Care transitions are defined by the American Geriatric Society as ‘a set of actions designed to ensure the co-ordination and continuity of healthcare as patients transfer between different locations or different levels of care within the same location.374 Corbett et al375 suggests that successful transitional care models have a focus on medicines management and include improving patient knowledge of medication, providing a list of medication, follow-up post discharge and simplification of the medication regimen (see figure 2.6) Often medication forms part of a larger transition intervention or ‘bundle of care’.376 Laugaland et al categorises interventions which have led to successful transitions of care (see table 2.7),377 with Corbett et al incorporating elements specific to medication (see figure 2.6).375

Table 2.7 - Interventions which lead to successful transitions in care 377

Commence at an early stage of hospitalisation and are maintained throughout the stay and the post discharge period Discharge co-ordinator Patient participation/ and or education Interventions that involve family/caregivers Multidisciplinary approach Curriculum interventions (post graduate and undergraduate) Pharmacy interventions Standardised reports/ discharge summaries Comprehensive transitional care programs with multi – interventional components.

A review by Hesselink et al which looked at handovers between primary and secondary care found that most effective discharge interventions incorporated those activities which focused on reconciling discharge information.378 There is

66 no evidence that one single intervention is regularly associated with positive effects on a single outcome measure. Most studies are for a multicomponent intervention with various outcome measures reported in various ways that makes comparison difficult (see table 2.8).

Figure 2.6 Elements of transitional care models375 (reprinted with kind permission)

Improving patient family knowlge of medication prior to discharge

Providing user- Simplifying post- friendly hospital medication Medication lists/tables at regimen Successful discharge Transitional Care Model

Scheduling Follow-up post follow-up care discharge via appointments home visits &/or prior to telephone calls discharge

The accuracy of the medication history taken at admission can affect the accuracy of discharge communication. If medication is not recorded on admission it may be accidentally omitted by either the patient or GP at discharge.379 It can sometimes be difficult to tell whether medication has been purposefully stopped or whether they were never recorded on admission.379 The WHO specifies that hospitals should have a process in place for ensuring

67 that, at discharge, the patient’s medication list is updated to include all medications the patient is to take following discharge, including new and continuing medications and previously discontinued medications that are to be resumed.1 The NPC states that discharge information should be clear, unambiguous and legible and should be available to the GP (or other primary care prescriber) as soon as possible post-discharge.80 The NPC goes on to suggest that ideally discharge information should be available to the GP within two working days of the patient’s discharge from hospital.80 There is limited evidence to determine whether this is happening in practice.380 Legibility of discharge communication has previously been reported as poor especially if it is the third carbon copy of a discharge prescription or a handwritten document.80 The legibility is improved where pharmacists write the discharge summary.79

The method of delivery of discharge information varies between NHS hospitals and may include posting by the ward or pharmacy, hand-delivery by the patient, faxing and electronic means, or a combination of these.103,366 Postal communication is currently the most popular method of receipt of discharge information in general practice. This is very slowly changing as electronic discharge is being introduced. Sexton et al suggests that through increasing the amount of medicines supplied at discharge, time pressures can be decreased.366 Alternatively, time is saved if the pharmacy sends a copy of the discharge information to the GP rather than the ward, although only few NHS hospitals reported doing this.101

Outside the UK, both Paquette-Lamontagne381 and Bedard et al292 show that the introduction of a medicines reconciliation form at admission improves consistency in data collection which can lead to the improvement in quality of the discharge information. In 2008, Midlov108,114 showed that the addition of specific information highlighting which medication has been stopped, started and changed led to significant reduction in medication errors. The alignment of primary and secondary care formularies can also reduce miscommunication and decreases discrepancies at discharge.382

2.4.4.1 The role of the hospital doctor at discharge from hospital

Junior medical staff are responsible for preparing nearly all discharge prescriptions in the UK.366 There are few interventions, however, that look at reducing the number of discrepancies made at discharge on discharge advice summaries; all come from outside the UK. Structured and standardised interviews by doctors informing the patient of their discharge information have been shown to significantly increase the patients’ knowledge.383 Also a dedicated geriatrician provided discharge package which incorporated medicines reconciliation assessment of ADEs and difficulties managing the regimen and optimisation of the medication regimen reduced readmission at three but not six months post-discharge.67

2.4.4.2 The role of the pharmacist at discharge from hospital

Sexton et al showed wide variation within hospital pharmacy practice within the UK relating to medicines at discharge.366 Thirteen years on this wide variation still exists.384,385

A number of pharmacist-led interventions have demonstrated an improvement in the discharge process within the UK, improving communication with the patient and other HCPs, timeliness of discharge and bed management.82,100,103 This is mirrored within international studies in a variety of specialties.187,219,310,312,386-388 Pharmacists reviewing the discharge summary and providing feedback on any errors has been shown to reduce medication errors at discharge in Sweden and the US.73,389 US pharmacists co-ordinating paediatric discharges have also improved ability to obtain supplies post- discharge.390

In contrast, Shaw et al found that discharge planning in mental health did not lead to any significant improvement in patient knowledge.106 It is difficult to tell whether this was due to the specialty or the type of mental illness within the intervention and control groups which were significantly different as the groups were not stratified for diagnosis. There were fewer medication problems post discharge in the intervention group; however, this was not shown to be statistically significant. Brookes et al suggests that pharmacists should also counsel patients at discharge and provide written information to reinforce

Table 2.8 Summary of interventions relating to the patient’s discharge from hospital categorised by HCP and including those interventions carried out by the whole team and the role of the patient

Country Nurse Hospital Pharmacy Staff Community Hospital GP Patient Multi- Pharmacy (CP) Doctor disciplinary team (MDT)

UK 48 hour post Pharmacist discharge Provision of No Provision of Provision of Patient aide- discharge planning discharge intervention discharge discharge memoire to 105,106 100,392 telephone information to studies summaries v information support patient structured community pharmacy letters370 398* education at interview and Provision of discharge 398 Information discharge116 home visit information 392 Provision of booklet204,400 with Provision of shared shared care Form to indicate medicines Discharge liaison care guidelines399 guidelines399 Reminder medication stop, reconciliation pharmacist 75,101 card/list100,383,401,402 started and 391 Community Explanation of changed during Pharmacy led discharge pharmacy domiciliary stopped/started/ Patient delivers to hospital 74,88,393 108,114 Provision of service visit105 changed GP401 stay discharge medication82 information Reissue of PODs at Discharge discharge394 Provision of planning meeting by information to GP (doctor and community 71 Writing discharge by pharmacist397 nurse) nurses 388 summaries102 Integration of Pharmaceutical care plan hospital with domiciliary visit395 admission medication, Verification of discharge inpatient information122 changes and hospital Pharmacist discharge into counselling82,100,102,396,397 one form381

Country Nurse Hospital Pharmacy Staff Community Hospital GP Patient Multi- Pharmacy (CP) Doctor disciplinary team (MDT)

International Supported Patient Med review and Formulary 382 No intervention No intervention Discharge transition375 education123,277,278,304,312,317,3 counselling416 studies studies planning187 83,388,405,406 Medicines Discharge Intervention identify reconciliation Pharmacist – coordination Counselling cost- drug related as part of nurse-facilitation 403 407 417,418 effectiveness problem discharge of resolution of package with 405 discrepancies Counselling Technician counselling med review 413 using teach phone follow up123,304,312 Communication to and post discharge back CP219 education67 technique403,4 04 Hospital pharmacist Structured verifying accuracy of standardised Discharge discharge information patient survey tool154 73,123,219,388,390,408,409 interview to

Pharmacist checking educate the reconciling medication in a patient 383 care home post discharge 389

Active communication of a discharge plan(addition of drug related problems and resolution to plan)410

Questionnaire to gain feedback from patients on their experience of their hospital stay in relation to their medication at discharge 89

Country Nurse Hospital Pharmacy Staff Community Hospital GP Patient Multi- Pharmacy (CP) Doctor disciplinary team (MDT)

Transitional care pharmacist 262

Production of interim MAR chart for nursing home411,412

Transcription408

Pharmacist–nurse- Facilitation of resolution of discrepancies post discharge413

Transfer summaries with med review by CP and case conferences with doctors and pharmacists414

Discharge planning317

Counselling73

Educational tool to apply pharmaceutical care at transitions415

72 information provided.75 Intensive pharmaceutical care including discharge counselling and explanation of changes to medication regimens has been shown to improve the patients understanding and recall of their medication,82 however, there is conflicting evidence on whether this impacts on readmission rates. In the US, Wilkinson et al showed a significant decrease of approximately 6% in 30 day readmission rates,406 however, in the UK no decrease in readmission rates within six months nor a change in deaths, visits to the outpatient or general practice were seen.105 When combined with medication review and follow up post discharge, counselling also led to decreased ADE within 30 days.123

Campbell et al suggested the widespread utilisation of pharmacists to bridge the continuity gaps at admission and discharge and improve communication between primary and secondary care.48 Various terms are used to describe pharmacists who co-ordinate and bridge the gap between primary and secondary care, including community liaison pharmacists75,101 and transitional care pharmacists.262 These roles are not dissimilar to nurse discharge coordinator roles, however, they focus on medication use. These pharmacists may do more than plan the discharge of patients they may also follow the patient up in primary care. They have demonstrated clear benefits in achieving seamless care within the UK75,88,101 and internationally, 262,414,419 undertaking activities such as preparing discharge information, counselling at discharge and communicating with other HCPs, demonstrating both cost savings and reducing readmissions.75 Discharge plans written by hospital pharmacists in Canada have been shown to decrease a number of drug related problems at discharge and allow community pharmacists to make relevant interventions.410 Both Elliott et al412 and Tran et al411 have demonstrated improved communication and reduced discrepancies through the production of MAR to send to care homes post discharge; neither of these were conducted in the UK.

2.4.4.3 The role of the nurse at discharge from hospital

Within the UK, it usually falls to the nurses to inform patients about discharge medication, including what it is for, when and how to take it and how to obtain further supplies.106,420 Studies from the US show that nurses are well placed to identify reconciliation discrepancies, especially at discharge, and, when working

73 in collaboration with pharmacists, increased discrepancy resolution can be seen.375,413 Nurse co-ordinated discharge and case management including counselling and follow up post hospitalisation have also proven to improve error rates 391 and decrease rehospitalisation within 30 days of discharge.403

2.4.4.4 The role of the community pharmacist at discharge

Community pharmacists can find it difficult to know whether a patient has been in hospital.360,421 Supplying community pharmacists with a copy of the patient’s discharge has been shown within the UK to reduce unintentional medication discrepancies,316 however, in the US Lalonde et al found that supplying the community pharmacist and the GP with the discharge medication did not decrease the amount of medication discrepancies.398 Communication between hospital and community pharmacists can prevent incorrect medication histories, create awareness of current issues or problems at home, prevent unintentional changes and ensure continuity of supply, allowing problems to be resolved before a patient arrives at their community pharmacy,422 however, UK community pharmacists rarely receive discharge information and where they do it is often for a specific reason for example where special formulations for example oral liquids for patients with dysphagia or compliance aids are required.40,360,366,395 Some GPs support the notion of community pharmacists receiving discharge information;423 inclusion of the community pharmacist in communication would contribute towards effective, seamless care.

Community pharmacists report that patients present to them before going to their GP following hospital discharge. This can be especially true in the case of MDS.360 The lack of provision of discharge information to community pharmacists is common. Community pharmacists would like to receive medication information on discharge,75,107,360,424 especially medication changes, special requirements, ADRs and allergies.

It has been claimed that patients do not use the same community pharmacy within the UK,40,425 making it difficult to know where to send discharge information or for the hospital to know who to liaise with. Yet, there is much evidence to suggest that this is not the case with many patients being loyal to a single pharmacy.75 It is unclear, however, whether a patient’s loyalty is

74 proportional to the number of medicines a patient is prescribed, their age or the locality they live e.g. near a town centre versus village.

2.4.5 Post-discharge (in primary care)

At discharge the patient may move to their own home, a care home or intermediate care, with their discharge information being passed to a number of different people, most definitely the GP and in some cases district nurses, community pharmacist, social services, care home staff or intermediate care. The recipient must then process the information they have received. In the UK, Akram and Urban found that 80% of discharges were received within 7 days and took four to nine days to be actioned by the general medical practice.380 This is similar to previous findings by Coleman et al who found that 96% of discharges arrived within eight days, with outpatient letters arriving within 21 days in 92% of cases.373 The delay in receipt of discharges is not unique to the UK; a study in Australia showed that 33% of GP respondents felt the discharge summaries were not always received in a timely manner.426 Also, a study in the Netherlands showed 75% of GPs experienced a delay.427 One Australian study showed that within 30 days post discharge 71% of patients will visit their GP and 86% their community pharmacy, making it important for primary care HCPs to receive the discharge within this time.428 Similar studies have not been conducted in the UK.

2.4.5.1 The role of the GP post-discharge

The GP is the central record holder in primary care within the UK and on receipt of discharge, information should review and update their records appropriately. Ultimately these records then may inform the patient’s medication history if they are readmitted. Crowe et al suggests that general medical practices use an information processing model to process correspondence from secondary care.429 This involves six stages: receipt of information, dissemination, amendment of records, reading and actioning recommendations, scanning information and informing the patient. There is often pressure within primary care to turn prescriptions around in a timely manner,75 especially where patients are requesting further supplies. Delays in processing can be caused by work patterns, working over split sites and prioritisation of other tasks.430 Prior to the

75 re-issue of medication post discharge, the GP may see the patient in primary care, however, this is not routine and often dependent on the practice or individuals within the practice.429 In Sweden, Rahmner et al found that GPs had differing perceptions of their responsibility for the patient’s medication list with some feeling they had responsibility for all the medication the patient was on, others just those that they prescribed, and further GPs a shared or imposed responsibility with other healthcare providers.431

Unintentional changes in medication, especially dose instructions, omissions and additions occur in supplies obtained post-discharge rather than at discharge,379,432 with medication being accidentally discontinued or added in primary care. This can cause further problems if the repeat medication list on the GP system is not updated appropriately.379,430 It is often unclear whether discrepancies post-discharge are a result of error on discharge or poor medication history taking on admission, supporting the cyclical nature of medicines reconciliation and the dependence of subsequent stages on the accuracy of previous stages. Despite the age of these studies there is nothing to suggest that this has changed. Repeatedly, general practices rely on the patient, their relatives or their carers to inform them of information, especially where it is conflicting. GPs also frequently rely on hospital discharge information to indicate how long patients should continue specific medication but this information is often not present on the discharge summary.88,101 One US study suggests that within care homes, nursing staff with varying qualifications perform medicines reconciliation post discharge, with the process they use varying. It is unclear what currently happens in the UK and further work is needed to understand what happens in these areas and whether staff have appropriate skills to accurately process the discharge information.

Crowe et al discuss the amending and updating of records post discharge as a task exclusively carried out by GPs themselves,429 whilst in practice this is sometimes conducted by pharmacists and in some instances administrative staff, who do not necessarily have appropriate competencies.80 Duffin et al found that often changes post-discharge could not be contributed to a conscious clinical decision,433 and, although using administrative staff to amend records post discharge reduces delays, it can leave room for error, especially if there is no requirement for the GP to see the letter before records are updated. 76

The extent to which this is still practice is unclear. The NPC specifies that medication lists must first be reviewed by the GP, the practice nurse, or the practice pharmacist, with the untrained person only taking responsibility for the administrative side of the process.80 A systematic review from 2009 concluded that there was no good quality evidence demonstrating effective medicines reconciliation interventions in primary care and that further studies looking at this area were needed.434

2.4.5.2 The role of the community pharmacist post-discharge

There is evidence from Europe to suggest that community pharmacists can usefully counsel patients following discharge.416,417 Although there is little evidence to support it, some hospitals in the UK are encouraging patients to visit their community pharmacy for a Medicines Use Review (MUR)d post- discharge to inform them about their medicine.435 The national target groups for MURs includes those patients recently discharged from hospital who have had medication changes made.436 This is supported by the Royal Pharmaceutical Society (RPS)437 and has more recently been introduced as an advanced service in Wales.438 One published abstract shows improved Functional Independence Measure Scores in patients receiving a locally commissioned, enhanced domiciliary MUR service within seven days post discharge,439 although this was part of a wider discharge programme. A similar scheme (Home Medicines Review) in Australia has proven successful and popular with patients.440,441

2.4.5.3. Role of other staff within primary care

The literature search did not reveal any studies which discussed the role of the practice pharmacist, practice nurse or administrative staff in relation to medicines reconciliation within primary care, however, it is known in practice that these personnel can be involved in the processing of discharge information.

d MUR is an advanced service offered by pharmacies in the United Kingdom. It is part of the contract pharmacies hold with the National Health Service (NHS). The patient is provided with an opportunity to discuss their medicines use with a qualified pharmacist.

2.4.6 Standardisation of the medicines reconciliation process

Throughout the entire process many interventions describe standardisation as a way of minimising error associated with medicines reconciliation. It is clear that more complete accurate documentation of medication histories is required.76

Standardisation of documentation or proformas have been shown to aid the medication history taking process within the UK101,442 and Canada,292 providing appropriate boxes and prompts to aid information gathering.401 Previously, in 1993, the RPSGB produced an admissions and discharge checklist to support communication between primary and secondary care.443 Although evaluation deemed it useful to indicate when regimens had changed, there was a lack of awareness by pharmacists of its existence,425 highlighting the need for appropriate promotion and dissemination of information. In Australia, Smith et al developed a multidisciplinary form to try to capture information about medication discrepancies occurring across the interface to standardise the reporting.92 Whilst this was only tested using vignettes and not in practice, the level of agreement on completion made it a potential solution to standardising the collection of information on the type of discrepancies occurring at the interface. It was ultimately validated in a study by Corbett et al who used nurses to identify discrepancies at transition from hospital to home.375

It was suggested by Mottram, in 1994, that although discharge letters contained more information than a discharge summary they were often poorly structured and it could be difficult to extract key information; necessitating an improvement in the design of the discharge forms, completion of the forms and administrative systems to dispatch the forms.369 Subsequently, the inclusion of structured cues and standardisation of discharge summaries has improved communication as long as they are completed in full.101 This has the potential to be improved through the use of IT (see section 2.4.7).

Nevertheless, there remains inter-hospital and intra-hospital variability of forms which can cause confusion when employees move areas of work. The WHO have appealed for health-care organisations to put in place clear policies and procedures, plus standardised systems to collect and document information for

78 each patient to provide to the next caregiver at each care transition point, whether at admission, transfer, discharge or outpatient visit.1

2.4.6.1 Minimum data set for medicines information provided at discharge

In 2007, with the aim of standardisation, the WHO suggested a minimum dataset which should be included in discharge information.1 This included details of started, stopped and changed medication including reasons and the intended duration of treatment.1 This was added to by the NPC, who suggested the minimum dataset should also include complete and accurate patient details including NHS number, weight if under 16 years, consultant, ward, dates of admission and discharge, presenting condition plus co-morbidities, procedures carried out, a complete list of all medicines the patient should take following discharge, the dose, frequency, formulation and route of all the medicines listed, allergies and patient information provided such as corticosteroid record cards, and anticoagulant books.80 In 2010, one primary care trust (PCT) liaised with their general medical practice system supplier to be able to extract the relevant data included in the minimum data set to provide to hospitals when one of their patients was admitted.444

More recently the RPS has also published a minimum dataset which mirrors that of the NPC but also includes GP details, usual community pharmacist, specialist nurse, description of allergy reactions, generic and brand name (where relevant), the indication, duration, on-going monitoring requirements, requirements for adherence support, special product information, how information has been given to a representative or carer (where appropriate) plus details of the person completing.445 The extent to which this is embedded in practice is unclear. An audit of discharge information carried out in four general medical practices in Surrey, UK found that the NPC minimum data set were not being met or fully completed.446 This is supported by a further audit in Central Manchester University Hospitals NHS Foundation Trust suggesting that compliance with the standards varied between 19 and 100% with medication changes not being documented routinely.447 Although the majority of these are essential in effective communication, with the amount of information required, it is not surprising that often information gets overlooked. The lack of completion of discharge documentation is not uncommon either within or outside the UK.448- 79

450 Legault et al determined that 95% of discharge summaries failed to include recommendations regarding medications in dysphagia.451 2.4.7 The role of information technology (IT) Table 2.9 Summary of IT interventions relating to the different stages of the patient’s journey through hospital

Admission During Stay At Discharge Primary Care UK studies Using EHR as No Standardised No source during intervention discharge form. intervention medication studies 293 studies history 235 e-discharge452 International Web based No Electronic No Studies system to intervention discharge intervention improve studies summaries363,473 studies information provision from Electronic patient nursing homes at list474 admission453

Computer physician order entry (CPOE) 454,455 456*

Kiosk to collect medication history90,457,458

Using EHR as a source

Using electronic prescription records as a source119,252,459,460

Electronic medicines reconciliation too l461-470

Electronic communication471

e-prescribing472

Pre-admission medication list (PAML) builder 142,463,464 *no significant effect seen following intervention 80

Technology has the ability to improve existing clinical and administrative processes in health475 and mitigate vulnerabilities at transitions in care (see table 2.9).476 In 2010, a review by Bassi et al showed that IT can and has been used to facilitate aspects of the medicines reconciliation process, including decision support software, electronic health records and electronic prescribing.477 Despite these suggestions, eight years on from the survey by Hobson and Sewell,362 it is uncertain how much the use of IT has increased. The NPC suggests that the full implementation of electronic transfer of patient information between healthcare settings will make the reconciliation process much simpler, however, they also acknowledge that until such times as it is fully operational, organisations will need to develop robust procedures for medicines reconciliation and establish them within everyday work patterns.80 Whilst electronic means of communication are likely to reduce errors they will not take away the need for trained personnel to check that medication has been prescribed accurately.6 Data completeness is important for patient safety.475 Once implemented, IT can ensure information is uniformly available and can address active and latent failurese, reducing error for example by using forcing functions and alerts to ensure fields are not left blank.475,478 HCPs have also expressed the importance of a simplified, streamlined computer entry process.230

The systematic review commissioned by NICE & NPSA concluded that improvements in IT were likely to reduce medication errors in medicines reconciliation, going further to suggest insufficient evidence to promote one particular method.6,48 This was supported by a report from the WHO Information Technology for Patient Safety Expert Working Group which examined the role of IT in improving patient safety in healthcare, reinforcing the need for increased use of IT to improve patient safety.475 This report focused on the management of information, especially where there was need for integration of different types of data. This is especially pertinent to medicines reconciliation as information from a variety of sources is assimilated and transferred between care settings. There is, however, a gap in evidence

e An active failure is an unsafe act (error or violation) committed by those with direct contact with the patient. A latent failure is an organisational or managerial failure for example lack of policy and procedures

81 regarding the significance of IT interventions and the associated cost-benefits for medicines reconciliation within the UK. The few studies which exist are international (mainly from the US) and focused at admission, making the role of IT in medicines reconciliation somewhat vague.48,475,479

In contrast, other studies have recognised the significant limitations of IT in the reduction of errors, for example the ability to override warning messages or the complexity of the software.475,480,481 Huckvale et al suggests that the introduction of new tools can disrupt existing processes introducing sources of risk or facilitating unsafe behaviours.475 There can be variance in the level of accuracy of the information due to the accuracy of input and the maintenance of records; Samaranayake et al describes these as socio-technical errors as often they were due to poor interface between the user and the computer.482 The use of technology allows for novel solutions such as timeline displays to allow visualisation of the medication history483 and intelligent systems which can predict the likely omitted medication using collaborative filtering approaches.484,485

At Admission

Various electronic medicines reconciliation tools have been shown to improve and standardise data collection and reduce discrepancies at admission to hospital in the US465,468 and Spain.467 In the US, PAML builders have been evaluated.142,463,486,487 These are applications which create a PAML from several electronic sources and, although not perfect due to limitations in the system, have been shown to decrease the number of unintended discrepancies. Difficulties arise integrating and interfacing the different existing electronic systems and sources available.488

Various US studies have also looked at Computerised Physician Order Entry (CPOE) (also sometimes referred to as Computerised Provider Order Entry or Computerised Provider Order Management).455,456,466,489 CPOE systems are designed to replace a hospital's paper-based system. They allow users to electronically write the full range of orders, maintain an online medication administration record, and review changes made to an order. They can also incorporate elements of e-prescribing systems such as Computer Decision

Support Software (CDSS) and interaction alerts. They have shown to be important in the reduction of medication error through decreasing transcribing, and thereby increasing seamless care.455 There are some studies which integrate PAML builders with CPOE to improve process, information flow and decrease transcribing error.466,470

Lesselroth et al take automation and IT one stage further, allowing the patient to verify their medication history themselves using an automated patient history intake device (APHID).90,457,458,490 This proved cost-effective and allowed the identification of relevant discrepancies. Medication lists are retrieved from an EHR and the patient verifies the information using a kiosk, making additions and amendments to the list where appropriate. Of course, there are still limitations for those who are unable to interact with the system due to e.g. cognitive, locomotor, or visual impairment.

An accurate medication history needs to be easily accessible. The idea of a common shared record between secondary and primary care has been discussed for 25 years.247 Sexton et al suggests that given the wide access to electronic banking it is surprising that there is not access to an up to date patient regimen.366 This has been somewhat addressed in the UK with the national summary care record, although it also has limitations and is not being utilised to its full potential.491 Locally within Bradford Hospitals NHS Foundation Trust, the shared EHR has been shown to reduce medication error when used in conjunction with other sources, primarily the patient or carer.153 In the Netherlands, Van der Kam471 showed that the electronic sharing of records between the GP and community pharmacist resulted in better agreement regarding their current medication, although did not eliminate totally the problem of dissimilar information. More recently the experiences of the shared EHR in Sweden highlighted the differences between access to data and information and access to knowledge.91 Data and information within the EHR must still be reviewed; this relies on HCPs interpreting the information within the EHR accurately. Finally, national electronic prescription records have shown to add value,252,460 as well as computerised outpatient and inpatient dispensing data459 (see section 2.3.11).

During Stay

Electronic prescribing was introduced into general medical practices in the UK in 1990s to try and improve the accuracy and efficiency.492 In hospital, however, progress has not been made at the same pace. A UK survey in 2010 of 63% hospitals showed that just 8% of hospitals had access to electronic prescribing with variability in the systems used and compatibility with other systems (32% of all hospitals for chemotherapy prescribing).385 Four years on it is unclear how this has changed.

At Discharge

Communication of information at discharge can be improved through computer generated discharge summaries and standardisation.293,372,408,409,473 However, Callen et al found that electronic summaries contained more errors than handwritten, concluding that it is not necessarily that electronic discharge summaries are of higher quality than handwritten ones but that information is less likely to be omitted due to forcing functions.493

A survey by Hobson and Sewell, in 2003, concluded that 79% of NHS Hospitals used paper based discharge systems, once again showing the limited use of IT systems within secondary care.362 Traer and Madhaven argue that there is no obvious joined up approach to the implementation of electronic systems within secondary care in the UK.492 The recent extent to which electronic discharge systems have been embedded within the hospitals is unclear.

2.4.8 Summary Part 2

The patient’s journey into hospital starts at admission where the doctor clerks the patient, takes a medication history then prescribes for the patient. Pharmacy staff then verify this medication history. During the patient’s stay the patient may be seen several times by different doctors and medication altered. At discharge a discharge summary is written and transferred to primary care once the patient is discharged. The scale and severity of discrepancies and error seen in Part 1 are a result of the complexities of the health system and the involvement of many different HCPs.

Many interventions lead to a level of reduction in medicines reconciliation error at admission and discharge. By targeting both admission and discharge together, error can further be reduced.122 With evidence that pharmacists are more likely to write legibly, counsel more effectively, take more accurate medication histories and identify both allergy status and OTC medications than other HCPs,78,79 it could be argued that they should undertake many of the processes associated with medicines reconciliation within the healthcare system. There is scope to increase the input of pharmacists and their support staff at admission to hospital,342 undertaking tasks such as taking the initial medication history and prescribing IMC. The role of nurses and primary care staff is less well defined and need further exploration, although it is clear that they need accurate and timely information to be able to effectively continue care post-discharge. There is potential for increased community pharmacy input at admission and discharge.

Varying educational interventions have tried to improve medication history taking accuracy and reduce discrepancies. Although these are usually targeted at one particular HCP there is nothing to suggest that these are not transferable. There is scope for inter-professional education of medical and pharmacy students to take medication histories as previously shown to work by Greene et al.494 Further work on the impact of medicines reconciliation on readmissions is needed together with better quality economic evaluation of medicines reconciliation interventions to enable cost-effective recommendations to be made.495 A range of different IT solutions have been tried and tested, nevertheless, further work is still needed to support the use of IT for medicines reconciliation

2.5 Part 3 The Role of the Patient in Medicines Reconciliation

2.5.1 Overview

Accurate medicines reconciliation at transition points in care is extremely important to ensure patients receive the correct medication. As patients age they are more likely to have at least one chronic disease which requires treatment and to be on several medications.220 As discussed previously, these complex medication regimens can lead to increased medication discrepancies at hospital admission resulting in potential harm to the patient.70,109- 111,117,145,173,177,180,181,183,197,212 Self-management of their medication relies on the individual patient being able to understand their medication. This can involve obtaining supplies, understanding changes and knowing when to ask for help. Various factors have been shown to affect the patient’s ability to manage their medication effectively. These include low health literacy, management of multiple medications and increased frequency of administration.496-500 Ultimately, this poor medication management can affect recall of medication history, leading to discrepancies at admission.

Over the last five years there has been an international drive to empower patients, make them central to healthcare2,475,501 and include them in improving patient safety;502,503 medicines reconciliation being no exception. The WHO suggests that to be optimally effective, patients and their families must be involved in the medicines reconciliation process and their participation encouraged.1 Yet, patients are not aware of the medicines reconciliation process and that they have a clear role within it,332,504,505 nor do HCPs appear to encourage active partnership with patients.506

2.5.2 The potential of collaborating with the patient and/or patient representatives to reconcile medication

The patient is the common factor within the medicines reconciliation process and can play a pivotal role in accurate transition of information about their medicines.507,508 The NICE & NPSA,6 IHI,2 WHO1 and the NPC80 all suggest that the patient is utilised as a source of information during medication history taking at admission to hospital. This is not surprising as patients are in the best

86 position to provide information on precisely what medication they take/use and how they take/use it, as well as being aware of the medications prescribed.

Discussions with the patient are important as some patients may decide to stop a medication or not to take it as intended by the prescriber, often without consultation with a HCP.509 This could lead to the addition or omission of medication on prescription charts at admission if the patient is not consulted. It is thought that between 33 and 50% of people do not take their long-term medicines as recommended.510 This can lead to increased medication discrepancies.497-499 One US study showed disagreement in 86% of cases between what the doctor thought the patient was taking and what they were actually taking.511 A further US study in 2008 reported 97% discrepancy rate between the medical record and what they were actually taking.512 Thirty-two per-cent had an allergy or ADR discrepancy.512 Also, some patients admit that they would fail to tell their GP if they chose not to have their prescription dispensed.513 Conversely, an Australian study demonstrated that only 20% of patient-completed medication histories were accurate in surgical patients.514

The NPC describes two categories of source, those which are more reliable and those which are less reliable;80 classifying patients as a more reliable source, however, their rationale is unclear. There is no evidence to suggest that the patient is more accurate than other sources, although Glintborg et al determined that patients' self-reports are reliable when estimating recent use of cardiovascular and antidiabetic medication.507 Patient collaboration can also be used to recognise opportunities to improve medicines management services through feedback questionnaires at discharge.89 Involving the patient can reveal a significant amount of extra information not provided by other sources.78,121 Cochrane et al showed that a quarter of patients were taking regular OTC medication; information which cannot be gained from additional sources.379 These various studies demonstrate patients to be an important indicator of precisely what they are taking/using at admission.

In some cases, patients rely on relatives or carers to supervise medication379 and often would not notice themselves if anything was changed or omitted.515 Yet, where patients self-administer their own medications at home, their ability to recall their medications can be superior to that of other patients.515 It is

87 therefore, widely acknowledged that a range of sources must be used to verify the list of medication a patient is taking including the patient. 49,50,76,145,166,233-235

Some patients may not be able to recall their medication history without relevant prompts such as the use of a list or the medication in front of them.163,516 although Witting et al found that 70% of patients being admitted to A&E could accurately complete a self-administered medication form.517 The use of a pre- clinic questionnaire combined with a clinical interview has shown (in the UK) to improve the accuracy of a patient’s current medication list compared with either method alone as a source of information.294 This has also been demonstrated in Canada518 and the US.331

Nearly 30 years ago, both Mageean519 and Harding42 found that the GP often received important details from the patient, before any formal hospital discharge information was received. Although the timeliness has been improved by the use of the discharge advice summary,446 it is not clear from the literature whether this has improved. Cromarty et al demonstrates that the provision of any discharge advice summary to the patient at discharge is valued by the patient and also decreases the incidence of medicines related problems post discharge.392 Despite the evidence to suggest the patient’s usefulness as an information source, there is little evidence to show how widely they are involved in medicines reconciliation and what role they should play.

2.5.3 Potential limitations of collaborating with the patient to reconcile medication

There may be a number of challenges and limitations when using the patient during the medicines reconciliation process, including an appropriate assessment of their competence to provide accurate information. Sometimes the patient may be away from the bedside, such as having investigations or procedures, making them inaccessible, even so, every attempt must be made to include the patient in their care. There is little UK evidence which ascertains the proportion of patients who do not have the ability to contribute to an accurate medication history or the proportion who actually do contribute.

NICE & NPSA6 recognise the difficulties obtaining accurate information from patients about their medication, which may be caused by their acute condition, 88 sensory/cognitive impairment, lack of support or language barriers. It is supported by previous literature which highlights the challenges of patients being drowsy, asleep, confused or unable to recall names of medication.76 Patients may not be able to communicate or may feel so unwell that medication is not of interest.82 Patients may also be more vulnerable at different stages for example immediately post-discharge.520 Collins et al found that even when the patient had verbalised their list of medication to the pharmacist, it was often revealed through a further source that there were additional medications the patient was taking.76 Both Meyer et al163 and Lingeratnam et al332 showed approximately half of patients providing a medication lists from memory were accurate, reiterating the importance of relying on more than one source.80 Certain patient populations e.g. the homeless provide other challenges where due to their unstable living environment they are unlikely to have an accurate medication list.521

Low understanding by patients,145 along with lack of knowledge about what they are taking522 and why they are taking it523 can all lead to medication error. A review by Ostini and Kairuz suggests a possible U-shaped relationship between non-adherence and health literacy where people with low health literacy are more often non-adherent, often unintentionally; people with moderate health literacy are most adherent; and people with high health literacy can be non- adherent due to intentional non-adherence.524 Cognitive ability and health literacy can contribute to the person’s ability to recall information during medication history taking and to follow prescribed regimens.496-499 This is supported by the work of Marvanova et al who determined that lower health literacy and lower cognitive function and higher number of medications all independently contributed to lesser understanding of the intended medication regimen.500 Patients must be able to translate knowledge into safe practice. Hain et al525 showed patients’ lack of self-awareness of the problems they were having with medication. During telephone interviews 87% responded that they had no difficulties with their medication regimens, however, on home visits 52% had at least one issue with their medication. This was complemented by Cain et al526 who determined that patients reported having all the information they needed, however, this did not translate into appropriate medication

89 management. Patients often expressed uncertainty about their medication regimens such as when and how to take/use medication.

Whilst many patients show an active interest in their care, some do not want to be involved with decisions about their medication,332,527 making it important to tailor information to individual wants and needs.528 With all this in mind, there is debate whether it is reasonable to expect patients to identify error and to query medication changes or discontinuations.78 Although some patients may take an active role in their care, questioning changes to their treatment, others have a more passive role, expecting changes and assuming they are deliberate.527 This is supported by Lingaratnam et al who identified three types of patient group: those who were confident with their medicines management and perceived themselves to cope well; those who desired greater involvement, irrespective of their level of confidence; and patients who were reluctant to participate.

Much of the evidence relates to the secondary care setting and whilst most of the principles are likely transferable to the primary care setting, systematic study is necessary for any corroboration.

2.5.4 Role of patient representatives/caregivers in the provision of information

Care givers have an important role and can provide an independent source of information which may improve documentation of medication histories. Caregivers such as relatives, friends and paid carers often take responsibility for medication supply and administration. However, in practice carers are often not involved in counselling, with this being directed toward the patient rather than establishing who will be managing the medication. Pippins et al145 found that relying on family members or caregivers as a source was a risk factor for PADEs when used by the medical teams but were a source of accurate medication information when used by the study pharmacists. There was no explanation provided for this.

In the case of children, it is expected that the parent will play a central role in the provision of information about their child’s medication and they have been shown to be effective in ensuring seamless pharmaceutical care through the 90 delivery of the discharge summary to the community pharmacy and GP.88 This has been supported by Porter et al529 who demonstrated that parents could recall the name of the medication their child was taking approximately 80% of the time. Though, recall of the route, form and dose was less accurate. This was, nevertheless, greater than a subsequent study which demonstrated that only 50% of families were able to provide complete medication information when asked. Kimland et al530 demonstrated the importance of asking children about their medication showing 40% used prescribed medication, 65% OTC medication and 8% CAMs.

2.5.5 Collaborative interventions which support accurate medicines reconciliation

Several UK and international studies have concentrated on patient focused interventions that demonstrate the benefits of active patient involvement in medicines reconciliation. These include patients bringing their medication into hospital, self-administration of medication, patient completed questionnaires and patient counselling.116,334,405,416 These will be discussed below.

2.5.5.1 Bringing medication into hospital

Price et al suggested in 1986, that patients should play an active role by bringing their medicines into hospital at admission.247 More recently many NHS Hospitals are encouraging this further in line with NICE & NPSA recommendations. Methods such as the ‘green bag scheme’ are used, where green plastic bags are distributed at pre-admission clinics and used by patients and paramedics to bring medication into hospital.361 Although such schemes have not been evaluated in the UK, they have been shown to work in Australia333,334 and have recently been backed internationally by the WHO.1 The extent to which patients bring their medication into hospital is not widely cited within the literature; Daji and Urban531 found that it only occurred in 50% of patients, most of these being elective patients. This was mainly due to a lack of awareness of the need to bring them in by the patient. The use of PODs can produce demonstrable cost savings to hospital drug budgets, improve the accuracy of medication history taking78,400 and improve the patient’s ability to accurately describe the medication they are taking.247,515 Despite these advantages, only 50% of doctors stated that they always use PODs during 91 medication history taking.78 There can be reluctance from patients to bring their own drugs into hospital for fear they may be lost and not returned at discharge.75 Failure to bring medicines into hospital can lead to the potential for discharge medication and home medication to be taken concurrently post discharge. There is further work needed on the education of both patients and HCPs on the value of bringing medication into hospital.531

2.5.5.2 Self-administration

Self-administration of medication by the patient during their hospital stay has been shown by Shaw et al106 to increase the patient’s knowledge and maintain it three months post-discharge. This highlights the potential success of involving the patient in self-medication programmes prior to discharge.

2.5.5.3 Communication and patient counselling

International studies show that half of patients will have a medication change whilst in hospital with the majority of older patients having at least one medication change whilst in hospital.532,533 Lack of communication about these changes and poor patient understanding can lead to error during inpatient stay and post discharge.372,534,535 One study showed that a quarter of patients were never told by their physician that they were starting new medication, despite the feeling by physicians that they communicated with the patient.536 Another, showed patients had no understanding of two-thirds of medication changes.336 Further work looking at the frequency of communication about medication changes and relevant interventions to improve communication are needed. It is also known that where patients or their carers are given specific instructions they are more likely to follow them,88 however, the extent to which information is currently provided varies.

It has been established for over 30 years that patient medication counselling alone or with written information improves patient understanding, reduces the number of errors made by patients and leads to improved compliance.368,537 Counselled patients have been shown to have increased knowledge and to display fewer errors than uncounselled patients in recalling information about their medication82,383,395,424,537 which is imperative if patients are to be used as a source of information within the medicines reconciliation process. It has also

92 been shown in the Netherlands that it can assist with identifying discrepancies on discharge as patients are able to challenge perceived errors.405 This is reinforced by the WHO, who suggest patients be educated about safe medication use as part of medicines reconciliation and provided with access to reliable, relevant, and understandable information about their medications.1 Borgsteede et al528 go further to suggest that medicine information is tailored to the individual needs of the patient depending on their health beliefs and prior knowledge. Recently, the RPS445 has issued guidance on when and how to counsel patients about their medicines in the UK, although their evidence base is unclear.

Ensuring good communication with patients, however, can be time consuming424 and effective medication counselling has been shown to take up to 30 minutes.100 MacDonald et al537 suggested a member of staff, preferably a pharmacist, should spend approximately 15 minutes with elderly patients prior to discharge to ensure they fully understand their medication regimen. This was supported by Al-Rashed100 who found that pharmacist counselling linked to a medicines reminder chart and information discharge summary led to better knowledge and compliance compared to nurses counselling. A survey in 2000,366 found that in the majority of NHS Hospitals a pharmacist counselled fewer than 10% of patients about their discharge medication and in 82.2% of NHS Hospitals fewer than 20% of patients received written discharge information. Sexton et al went on to state that pressures on staffing and resources can often make services such as medication counselling at discharge unlikely.366 Almost 10 years on, there is little to demonstrate that this has changed.

There is evidence from Europe to suggest that community pharmacists can usefully counsel patients at discharge.416 Some hospitals in the UK are encouraging patients to visit their community pharmacy for a Medicines Use Review (MUR) post-discharge to inform them about their medicines, which is supported by the RPS437 although there is currently no UK evidence to support the effectiveness of this. Similar Home Medicine Reviews (HMRs) have been shown to work in Australia.523

2.5.5.4 Hand-held information

Sandler et al400 suggests patients should carry up to date written hand-held information around with them at all times to assist medication history taking. There is substantial evidence to suggest that the provision of hand-held records or information to patients improves their recall regarding details about their hospital stay, reason for admission, and medication, with patients often using the information as an aide-memoire.400 Both Irvine et al204 and Brackenborough 41 found that patients themselves did not think hand-held information would be useful which is contrary to findings by Sandler et al400 and Chae et al402 where it was found to have some use and improved actual and perceived patient knowledge respectively.

The APhA refers to this hand-held information as the patient medication list and defines it as ‘a record of current medications that an individual carries across the continuum of care to stimulate conversation between the individual and his or her healthcare providers regarding the patient’s current medications.’538 The provision of hand-held information allows patients to read it at home in their own time and not feel rushed. Many patients will also share this information with others including spouse, relatives and GP post discharge.400 Hand-held information is not intended to replace discharge information, rather to reinforce it and has been found to be particularly useful if combined with a regular medication review with a pharmacist.204

Hand-held records rely on information being kept up to date. This can be difficult and lists are often updated incorrectly or include omissions,204,332 putting some onus on the patient to prompt HCPs.41 Where hand-held records have been introduced, patients often did not carry their booklet or keep it up to date for a variety of reasons, including simply forgetting, disinterest and the belief that doctors already had access to this information.204

2.5.5.5 Self-completed questionnaires

Evidence suggests that although patient-self completed questionnaires at admission are not always complete and accurate, they can help to improve the accuracy of medication histories when combined with a clinical interview compared with either method alone.294,331,514,518 In France, Roulet et al tested a

94 questionnaire in which patients documented self-medicating behaviours. This identified double the number of self-medicating drugs compared with interviewing alone. It also had the potential to reveal iatrogenic conditions that were being self-treated, although this was not tested.330 Patient completed questionnaires need to be studied further to understand their role within the medicines reconciliation process. Marvanova et al500 suggests the use of a medication understanding questionnaire to establish the ability of a patient to understand their medication and give an indicator of their reliability. This would allow HCPs to exercise greater caution when using patients as a source of information.

2.5.5.6 Access to personal electronic health records (EHR)

Two studies from the US have demonstrated decrease in medication discrepancies through patient-access to their medication information prior to their doctor consultation,539,540 one via electronic access540 and the other a printed form.539 Schnipper et al found greater concordance between the documented and patient reported lists and fewer unexplained discrepancies when patients had access, in the study population.540 Webb et al determined that 80% of patients had some form of discrepancy on their EHR.539 In both studies, the prior access to the information encouraged patients to document their problems, concerns and queries, highlighting any discrepancies prior to review with the doctor. 539,540 Other patients, care givers and health providers have shown interest in this approach.541

2.5.6 Summary Part 3

Involving and engaging patients and/or their carers in the medicines reconciliation process should not be undervalued.222,529 They are the omnipresent link between all healthcare settings and HCPs. Patients, family members, and caregivers should be encouraged to participate in their care, bringing in their medication when they are admitted to hospital and keeping and maintaining an accurate list of all medications. This list should be updated and reviewed at each care encounter, and should include prescription and non- prescription medications, herbal and nutritional supplements, immunisation history, and any allergic or adverse medication reactions.1,64,402 Patients can

95 only do this if they are engaged and educated by HCPs so they understand their medication.522 HCPs must utilise patients as a source of medication information whilst recognising the patient’s individual limitations and taking into account their health beliefs. Whichever medicines reconciliation processes are established within an organisation, it does not reduce or eliminate the need to speak to the patient and verify the medicines they are actually taking or using. Further studies are needed to more clearly define the roles and responsibilities of patients within the medicines reconciliation process and to understand their usefulness and potential limitations.

2.6 Overall Summary Chapter 2

The scale of error associated with inadequate medicines reconciliation is well defined within the literature, and, although there is limited data on discrepancies at transition points from the UK it is recognised that there is a significant problem. It is also acknowledged that the overall process of medicines reconciliation consists of four key stages. Yet, the finer detail of the activities and systems which make up each stage or the roles of HCPs in the process are not clearly described. Without a clear understanding of practice and the stages within the medicines reconciliation process, as well as the surrounding context, it is difficult to assess where and how errors or other problems arise. This limits the ability to design reconciliation interventions, which ideally would be based on knowledge of the overall process so that any improvement in the quality and safety of reconciliation at one stage in the process (e.g. admission) contributes to reconciliation at the next and so forth. The patient has a role to play in the medicines reconciliation process, however, the extent to which they are currently involved and the precise nature of their input is not described in detail. The studies within this thesis explore the current frequency and type of medicines reconciliation discrepancies within the UK (Chapter 4) and look at the entire process of medicines reconciliation within secondary (Chapter 5) and primary care (Chapter 6) , including the role of the HCPs and their interaction with the patient.

CHAPTER 3 METHODOLOGY

3.1 Introduction

The overall study comprised three distinct stages, sequential in design (see figure 3.1). The first, a literature review to provide an evidenced-based picture of current practice and highlight any gaps in evidence; the second, a quantitative, multi-centre, descriptive, prospective study to determine the type and severity of discrepancies at admission to hospital; lastly the qualitative exploration of National Health Service (NHS) staff undertaking tasks associated with the transfer of information about medication. The third stage used both a framework of evidence-informed practice (deductive, evaluative approach) and complementary observation (inductive approach) to examine the process of medicines reconciliation and the interaction of individuals within the process to better understand the context in which medicines reconciliation errors and discrepancies occur. It elaborates and expands on the results found in the first two stages to determine broad ideas and concepts and explain the data gathered to establish a clearer understanding of the contributory factors which lead to error and error reduction within medicines reconciliation. This mixed- method approach542 was used to quantify, explore and explain medicines reconciliation-related errors and/or discrepancies which occurs at transition points in care. This chapter will justify the use of mixed-methods within the study, clarify the reasons for using each method, outline the nature of the data generated and signify how this relates to methodological theory. More emphasis is placed on the qualitative aspects of data collection within this chapter due to the differing theoretical perspectives compared with quantitative approaches (see section 3.2.1) and the associated application of rigour and reflexivity needed (see section 3.5). The weighting is also proportionate to the extent of qualitative data gathered within this study compared with quantitative. The specific methods employed and limitations of each method will be discussed further within Chapters 4, 5 and 6.

Figure 3.1 – Sequential study design

Stage 1 - Literature review (Chapter 2)

Stage 2 - Multi-centre, descriptive prospective study of type and severity of errors (quantitative approach - Chapter 4)

Stage 3 - Multi-centre, observational study of medicines reconciliation and related communications at care transitions (Qualitative approach - Chapters 5 & 6)

Stage 4 - Application to theoretical framework and practice recomendations (Chapters 5 &7)

3.2 Rationale for Methodology

3.2.1 Theoretical perspectives

Various theories exist in the methodological literature which describe the philosophical assumptions that influence research design. These are based on the researcher’s ‘world view’ (epistemology) and how they view existence within their world (ontology). They include objectivism, social constructivism and subjectivism. Objectivism assumes that the world exists independent of human belief and consciousness, therefore is deemed measureable. In contrast, social constructivism assumes the world is socially constructed in order for groups to make sense of the world, therefore multiple realities exist. Subjectivism assumes an individually constructed world where reality is experienced by each person. Crotty describes an inter-relationship between the researcher’s world view and the methods they choose to use.543 Theoretical perspectives exist which are congruent with these differing epistemological views. These theoretical perspectives lead to a variety of different methodologies and ultimately, methods (see figure 3.2).

Figure 3.2 Relationship between epistemology, theoretical perspectives, methodology and research methods. Gray544 adapted from Crotty543

Epistemology Theoretical Methodology Methods Perspective

• Objectivism • Positivism • Experimental • Sampling • Constructivism • Interpretivism research • Statistical analysis • Subjectivism • Critical Inquiry • Survey research • questionnaire • Feminism • Ethnography • Observation • Post-modernism • Phenomenological • Interview research • Focus Groups • Grounded theory • Document Analysis • Heuristic inquiry • Content analysis • Action research • Discourse Analysis

Objectivism gives rise to post-positivism, testing ideas and hypotheses through gathering data, evidence and rational consideration. Those who are true to post-positivism hold the philosophy that cause directly influences outcome, and truth and reality can be gained. It is an objective approach and produces quantitative data which can be measured, deeming an egalitarian researcher- subject relationship. Conversely, social constructivism and subjectivism assumes that humans engage with the world to produce meaning and prioritise understanding based on their experience. This viewpoint lends itself to qualitative research which seeks to understand the complexity of a situation in the context of the participants. With this, truth and meaning are constructed and created by an individual’s interaction with the world, rather than meanings based on the objective measurement of reality valued in post-positivism.

Although the theoretical debate about the strengths and qualities of the different approaches to research is less polarised, no specific method of data collection is perfect nor are any disregarded as inappropriate for obtaining evidence through applied research. Barbour agrees with this notion, reiterating that to ensure good quality evidence-based medicine the methods chosen must have the ability to address the research question’s relevance to practice.545

Pragmatism is not committed to one specific philosophy; instead it focuses on solutions to problems, allowing segments from both quantitative and qualitative methodology to be considered when designing research. It allows methods and techniques to be chosen which best meet the needs, purpose and

99 understanding of the research, provided that these can be justified and a sound rationale made. It recognises that each method has its own set of strengths and weaknesses dependant on the aims and practical constraints of the phenomena under study. It can be achieved either through the use of individual research paradigms or alternatively through the employment of multiple research strategies (mixed methods).546 This research study takes a pragmatic worldview.542

3.2.2 Justifying mixed-methods

Using mixed methods has been defined as:

‘the collection or analysis of both quantitative and qualitative data in a single study in which the data are conducted concurrently or sequentially and involve the integration of data at one or more stages in the process’ 547

As previously described, a divide has been seen between the use of qualitative and quantitative research with many researchers being specifically loyal to one or the other. They argue that the methods are mutually exclusive as they arise from different ontological and epistemological positions (see section 3.2.1). In contrast, Newman and Benz describe quantitative and qualitative approaches as different ends on a continuum rather than polar opposites.548 In the middle of this continuum resides mixed-methods, incorporating elements of both qualitative and quantitative research.542 Mixed-methods 542,549,550 is gradually being recognised as the third research paradigm551 and is becoming increasingly common in health research.552

Combining methods produces different kinds of data on the same topic which allows events to be seen from differing perspectives and can provide alternative or confirmatory explanations, enhancing the validity of the data. It provides a richer more complete and comprehensive picture than that achieved by one method alone553 and allows the biases within one method to be counteracted by the strengths of another. With this aim, it is important that the methods chosen complement and support each other to add value to research.

A well-documented advantage of mixed methods is the opportunity for triangulation. Two or more independent processes of measurement reduce the 100 uncertainty of data interpretation, thus any variance is likely to be that of the trait under study and not due to the method.554 Methodological triangulation has been defined by Denzin [555] as ‘the combination of methodologies in the study of the same phenomena’. It reduces the sources of error by gathering data from multiple sources or using different approaches. For these reasons mixed methods were employed in this study. The different methods, however, remain autonomous (see section 3.5).

Sieber proposes that fieldwork can help to interpret relationships and clarify unusual findings found using quantitative methods, shedding light on phenomena.556 This is reinforced by Jick who adds that it can also capture a more complete holistic and contextual portrayal of the phenomena under study and may add to the transferability of the qualitative findings.557 Mixed-methods can therefore not only examine multiple perspectives, but also enrich understanding. This is summarised by Greene et al who demonstrates the five main reasons to combine quantitative and qualitative research.558 These include:

1) Triangulation – to seek corroboration 2) Complementarity – seek elaboration, enhance, illustrate and clarify the results of one method with another 3) Development - the results from one method are used to develop/inform another method 4) Initiation – seeks contradictions/new perspectives through the use of another method 5) Expansion – extend the breadth of inquiry

This study integrates quantitative and qualitative methods to generate an improved understanding of current practice concerning medicines reconciliation, enhancing the validity, credibility and trustworthiness of the findings. The initial data collection in Chapter 4 outlines the scale of the current problems of medicines reconciliation, providing some context and helps to focus the rest of the study on the predominant issues. The subsequent stage (Chapters 5 & 6) allows expansion so as to gain a wider knowledge surrounding medicines reconciliation practice, analysing and explaining the findings in the initial study to provide a more detailed picture. Creswell defines this as ‘sequential

101 explanatory strategy’ where qualitative data is used to build and expand on the quantitative findings.542 Overall, the empirical work here draws upon a qualitative rather than quantitative methodology so as to explore and appraise the behaviours, actions, systems and practises at different stages within the medicines reconciliation process. As demonstrated in the literature review, much is known about the scale and severity of the errors made at transitions in care, emphasising that medicines reconciliation sometimes fails or does not take place at all. The majority of evidence published to date is, therefore, quantitative with several interventions which demonstrate improvement in error rates and quality of medicines reconciliation. There are few published qualitative studies, nor are there any studies looking at the contributory factors which may lead to medicines reconciliation error and error at transitions in care.

3.2.3 Exploring quantitative and qualitative approaches

3.2.3.1 Quantitative approaches

Quantitative data by definition is numeric and if the measurement process is accurate, data are confidently used to objectively report reality. Although there is some inherent risk of bias with quantitative data (e.g. attrition bias; selection bias) it is thought not as inevitable as with qualitative data.559 Quantitative data can be used to define the characteristics of a population (e.g. epidemiology or prevalence) or relationships between variables (cause and effect). The outcomes can be statistically generalisable based on whether the study has been powered correctly and whether the sampling strategy is robust. The quantitative data collected during the second stage of this study (described in Chapter 4) provides information on the number type and perceived severity of discrepancies experienced by the patient. This establishes the scale of the problem and determines the accuracy of the medication history taken at admission by the doctor and also the sources used by pharmacists to elicit information regarding medication being taken by the patient prior to admission.

However, any emerging questions as to why these problems are occurring may only be answered through qualitative research. Quantitative research involves little or no contact with people in the field setting.544 This is clearly apparent in Chapter 4. Only the charts of the patients rather than the interaction, decisions

102 and processes which lead to the production and verification of the IMC and discharge were studied at this stage. The complexity and inter-relationship between the many different aspects of medicines reconciliation presents a particular challenge to the researcher; hence the third stage (a qualitative approach) was critical to understanding the detail, context and contributory factors which may impact and influence the medicines reconciliation process.

3.2.3.2 Qualitative approaches

Qualitative research, in contrast to quantitative, is highly contextual and conducted through intense contact with the field, in a ‘natural’ real life setting.544 It aims to identify, describe and analyse which allows the researcher to gain a holistic or integrated overview of the phenomenon being studied in order to understand the ways in which people act and account for their actions.560 Gantley et al561 describe qualitative data as ‘the what?’, ‘the why?’ and ‘the how?’ rather than ‘the how many?’ or ‘how frequent?’ found in quantitative data. This is explained further by Thorne, who illustrates that qualitative methodology is often used to understand phenomena through deconstruction, i.e. the distinction between explaining how something operates (explanation) and why it operates in the manner it does (interpretation).562 The qualitative methods employed search for patterns and inter-relationships rather than the specified set of variables in the quantitative stage. Qualitative data can also determine whether any statistical significance determined in the quantitative stage is due to random variation or systematic differences. This is important here in triangulating the quantitative findings in Chapter 4 with the qualitative observations in Chapter 5.

Qualitative research is more suited to studying process i.e. how outcomes are achieved and the mechanisms involved. It can help to understand why interventions may be successful or fail and can help to understand issues with communication and interpersonal co-operation. The qualitative approach is, thus, better suited to the exploration of the medicines reconciliation process and its associated systems. Qualitative methods can also explore outliers (often called deviant cases) to determine why they diverged in the quantitative sample.542 It therefore captures those results which, although are not statistically significant, may be clinically significant and can inform

103

Table 3.1 The application of quantitative and qualitative methods within this study and their relevance: adapted from Bryman563 and Gray544

Quantitative Qualitative Relevance to this study Methods Methods Epistemological Objectivism Constructivism Combining both perspectives produces a pragmatic approach which contributes to position understanding the current picture of medicines reconciliation Relationship Distant / Close/ insider The close proximity of the researcher to the field during the observation allows between outsider increased detail to be gathered to explain the findings of stage 2. researcher and subject Research focus Facts Meanings The qualitative findings (stage 3) apply meaning to the facts determined in the quantitative stage (stage 2). Relationship Deduction/ Induction / The deductive approach confirms the extent to which medicines reconciliation errors between theory/ confirmation emergent arise within the sampling frame; the inductive approach determines the reasons why concepts and this occurs. research Scope of Nomothetic Ideographic A large sample is investigated to determine the types and severity of discrepancies, findings then individuals are observed, in-depth to understand where contributory factors associated with medicines reconciliation may occur. Nature of the Quantitative Qualitative Stage 2 is reported using numbers, frequencies and statistical tests. Stage 3 is data based on based upon reported using themes, description and thematic analysis. numbers text

104 for practice.564 Barbour concludes that ‘qualitative research methods can provide useful data on a range of topics and research questions which are not amenable to study using conventional quantitative methods.’’564 (The quantitative and qualitative methods used within this study are shown in table 3.1)

3.3 Data Collection Methods

Empirical data were gathered during the 2nd and 3rd stages (see figure 3.1), each set of data informing the next stage. All data were collected prospectively. Quantitative data were collected in the second stage; the method and results are described in Chapter 4. Qualitative data were collected using observation in the 3rd stage; the method and findings are discussed in Chapters 5 and 6.

3.3.1 Survey methods

Survey is a general term for a design which aims to collect the same set of data for every case within a study. Survey methods are used when the researcher wants to obtain factual information and are the design of choice for descriptive quantitative research questions. Observational surveys are used to make detailed, factual records of specific behaviours, events and settings. They usually employ a systematic tool. They focus on what people actually do rather than say and can be used to look at particular events e.g. when they occur and how often. They are usually designed to record things objectively. Most people think of questionnaires when discussing survey methods, however, where previous studies have recorded the prevalence, type and severity of medication discrepancies, the authors have described their study method as a survey.182,193,196 Survey methods are rooted in positivist epistemology and are therefore usually associated with quantitative data. In Chapter 4 a prospective, descriptive observational survey is used to collect data on the frequency, type and severity of discrepancies plus the frequency and type of sources used.

3.3.2 Observation

Observation is the systematic viewing of people’s actions followed by the recording, analysis and interpretation of their behaviours.544 It is primarily descriptive of settings, people and events, and the meanings that the 105 participants assign to them. In this case, the healthcare setting (namely the ward environment) and the actions and interactions of staff within the healthcare settings; all in the context of medicines reconciliation. The choice to use observation stems from the view that rich multidimensional data is needed to examine the medicines reconciliation process and contributory factors associated with medicines reconciliation error. Observation is often used to understand cultures and behaviours and is useful when individuals would be unable to provide objective information about the phenomena being studied.559 The researcher can determine what actually occurs in practice rather than relying on second-hand information or participant’s perception of what occurs; the type of data that would be elicited using, for example, focus groups or interviews. Accurate retrospective accounts of interactions in particular settings (e.g. those found in interviews and focus groups) can be impossible to achieve because the situational dynamics are often not reported by those who participate in them. They may not hold an objective view and may only have partial knowledge or understanding, based on the role they play in the setting; a non-holistic picture. It is felt that observation occurs in a ‘natural’ context rather than being reconstructed, although some argue that it is reconstructed by the researcher during data analysis which may cause bias.565

Within this study, observation was chosen as the researcher felt that interactions, actions and behaviours and the way people interpret these were central to explaining the findings in stage two.566 Also observation in the context of the setting was important to determine any relevant human factors associated with the task, the individual and the organisation. Pragmatically this data cannot be acquired using alternative methods.

The observational methods used have some similarities with ethnography as the researcher spent much time in the field. Ethnography studies cultures and groups using a holistic approach which looks at processes, relations, connections and interdependency amongst component parts.567 In this study, however, the researcher’s main aim was not to examine the characteristics and cultures of the HCPs and patients within the hospital setting, rather to study the systems and processes used to reconcile medication together with the HCP’s interaction with the process. Ethnography looks to provide a more

106 multidimensional picture, with richer description and attempts to see things from the observed party’s point of view in more depth than studied here.

The study also draws on some phenomenological principles, where the patient’s interpretation of events gives explanation to their experience. Phenomenology accepts that things can be seen in different ways by different people at different times, with alternative explanations being equally valid. Although this is not the main aim of the research, drawing on these principles goes someway to explaining the phenomena seen. The researcher needs to understand their thinking to interpret and understand their actions. Miles and Huberman560 suggest that research must be specific and focused, and that the looser the initial design the less selective the data will be during the observation. With this in mind the observation focused on medicines related tasks which may impact on medicines reconciliation and the medicines reconciliation process itself to ensure that the most appropriate and relevant data were collected.

Most observational data collection is overt, where the participants are aware that the observation is taking place. Most agree that it is unethical to conduct research in a covert or deceitful manner. Although it is accepted that researchers may choose to reveal/inform participants as they see fit, as it is not always possible in a setting to inform everyone of the researcher’s presence or intentions. All observation within this study was overt within the hospital ward setting, however, there were some staff on the ward who may not have been aware of the researcher’s presence e.g. housekeeping and administrative staff.

3.3.2.1 Deduction and induction

This study used both an inductive and deductive approach. This is supported by Gray who suggests that inductive and deductive approaches are not mutually exclusive.544 Deductive approaches take a prior theory, often with pre- defined measures and manipulate an independent variable e.g. increasing the number of sources used to illicit a medication history to test the impact of a potentially dependent variable e.g. the number of discrepancies found. Whereas in the inductive approach, theory arises from the research; data is collected and analysed to see whether patterns emerge which suggest relationships between phenomena e.g. the observation of medication history taking. This study incorporated a quantitative, deductive element that focused 107 on the frequency and type of medication discrepancies (Chapter 4) and a framework against which to measure medicines reconciliation during the observational stage (Chapter 5). An unstructured inductive element was also employed in Chapter 5 which embeds the medicines reconciliation process in a social context of how people interact and behave.

Deductive approach

Deductive reasoning starts with general theory on the topic being studied and utilises this to form a hypothesis that can be tested to determine whether it fits with the original theory (see figure 3.3). Within this study, a deductive approach is used to develop the data collection tool used to collect data on the type and severity of discrepancies within Chapter 4 and the evidence-informed framework used during the observations in Chapter 5 and 6. In both cases they determine how current medicines reconciliation practice fits with the relevant evidence base. The structured approach used allowed the systematic collection of data, and increased the consistency of approach within organisations thus increasing reliability.568 Structured observation is more economical as it allows a high focus and systematic checking of elements of behaviour which can be replicated by the same researcher or others at a later date. It also allows data to be collected at the time it occurs, however, this relies on the observer being present when it does occur.544 Gray suggests seven stages in the deductive process highlighted in table 3.2.544

Figure 3.3 Deductive reasoning569

Theory Hypothesis Observation Confirmation

108

Table 3.2 Seven stages in the deduction process applied to this study (based on the work of Gray544)

Stage in deduction process How it was deduced within this study. 1. Organisational mission To determine an evidence based framework which looks at the factors that should be incorporated into the medicines reconciliation process 2. Theory Gathered and synthesised during the literature review (stage 1) 3. Hypothesis Current practice does not mirror recommendations outlined in recent literature 4. Operationalise Devise data collection tool (stage 2) and observation framework (stage 3) to assist data collection 5. Testing by corroboration or Data collected and tested against the attempted falsification theoretical frameworks

6. Examine outcomes Data is analysed to reject or confirm hypothesis 7. Modify theory if necessary Determine reasoning through inductive approach

Inductive approach

Inductive reasoning works in the opposite way to deduction, moving from specific observations to broader generalisations and theories (see figure 3.4). Inductive reasoning begins with specific observations and measures to detect patterns and regularities, which can then be utilised to construct generalisations, relationships and theories. It does not set out to corroborate or falsify a theory, instead it attempts to establish patterns, consistencies and meanings. Although much is known about the frequency and type of medicines reconciliation discrepancies occurring in practice, there is little evidence as to why they occur. An inductive exploratory component is used within this study to 109 try and investigate contributory factors which can lead to medicines reconciliation error and determine how these fit in with the human factors approach. During the observation, initial description was used to set the scene and gain an overview, followed by more focused observation to concentrate on the processes and problems associated with medicines reconciliation, with tasks specific to medicines reconciliation being carefully observed. Inductive approaches can lead to a detailed, descriptive or narrative account rather than a critical or analytical insight. The use of case studies or stories can be useful to illustrate specific points, however, overall themes must be sought to enable comparison with theoretical frameworks.

Figure 3.4 Inductive reasoning569

Tentative Observation Pattern Theory Hypothesis

3.3.2.2 Participant and non-participant observation

Often, a very clear divide is seen between participant and non-participant observation. In non-participant observation the researcher is not involved in the activities of the group. They remain a passive observer, watching and listening to activities, then drawing conclusions. In contrast, participant observation enables the researcher to participate in the same manner as those being observed, with or without them knowing.570 The researcher must maintain a position where they are considered part of the group without becoming immersed.571

Both Mason566 and Gillham,572 however, argue that no observation is entirely non-participant and the researcher must participate in some form if the subjects are aware of the presence of the researcher. Similarly, Atkinson and Hammersley state

‘in a sense all social research is a form of participant observation as we cannot study the world without being part of it’ (p249).567

110

Being entirely non-participatory can also lead to the researcher’s actions and behaviours being interpreted as unhelpful or cold. The observation described within Chapter 5 is non-participatory. However, the potential conflict of interest between the investigators role as a researcher and their role as a HCP means there were times when the researcher had to move from a level of moderate participation (acknowledging the patient and explaining the study) to active participation, depending on the situation. Active participation included interventions in collaboration with the HCP where the researcher noticed something which could cause potential harm to the patient. Depending on the potential patient harm, the researcher intervened either at the point of noticing the potential harm or following completion of observation of the particular activity being undertaken.

3.3.3 Possible alternative methods

Rather than observation other qualitative methods could have been considered to explore and explain the phenomena being studied. Questionnaires would have provided greater anonymity, however, would have presented participants perceptions rather than reality with the potential for low response rates, self- selecting bias, lack of opportunity to clarify issues and non-spontaneous responses. Although observation sometimes leads to observation biases the busy nature of the NHS does not allow sufficient time for clinicians to reflect too deeply on their behaviour in the care setting. Alternatively, interviews would have provided the chance to explore more in depth and allows explanation and probing to gain more complete answers, although these reveal participant perception rather than actual practice.

3.4 Recording Data

Various methods can be employed to record data during observation. These include field notes, video and tape recording. Field notes are ‘writings produced in or in close proximity to the field.’573 Proximity means they are written almost contemporaneously with the events and experiences that they describe and recount’.573 Emerson points out that there are different ways to produce field notes which reflect the theoretical and methodological orientation.574 For example, they may be recorded as raw data which is gradually built up or

111 alternatively they might be used to record hunches, everyday perceptions or interpretations. Bailey suggests that field notes develop out of an analytical process; the researcher mentally captures as much detail as possible which assists recall and interpretation of jotted notes.575 The way in which field notes are recorded determines how they are read later during analysis i.e. whether they are literal, interpretative or reflexive. Mason believes it is important to make explicit the researcher’s interest (see also section 3.5 Reflexivity), framework and how these will be recorded.566 With field notes there is more chance that the researcher may be selective and choose what they deem important to document. This is where the use of recording equipment would be beneficial as the researcher can return to the tapes or transcripts at any time to re-define categories. Some participants may feel uncomfortable being recorded; which makes any observed interaction less than a true reflection of the situation.576 The choice of data recording within the hospital ward setting is further limited due to ethical considerations. It would be unethical to video or tape record in the ward setting without explicit consent and relevant ethical approval. This is difficult to obtain due to the many people who pass through a hospital ward, hence in this study contemporaneous field notes were used throughout.

3.5 Ensuring Rigour - Concepts of Reliability and Validity

Reliability is the capability of the research instrument to produce consistent results or measurements which are dependable. The researcher could, however, use two different instruments to produce the same outcomes; or the same instrument could be administered by two different researchers.544 When the instrument is a human and the setting is characterised by regular human interactions, as is often the case in social research, the results are of course unlikely to be replicated. The researcher being the main instrument of data collection is more responsive to the situation and they are able to adapt to any changing conditions. In qualitative research, the level of reliability is often referred to as ‘trustworthiness’ and is assessed through the detail with which the data collection methods are described and the reflexivity which the researcher 559 has applied so as to counteract any a priori assumptions. The researchers feelings were documented simultaneously to the data collections and were then reflected on to ensure observations remained as objective as possible

112

Researchers conducting qualitative descriptive studies should seek to accurately account events that most people observing the same event would agree is accurate ‘descriptive validity’ and accurately assign meaning that 577 others would agree was accurate ‘interpretative validity’. Validity is defined as the degree to which the researcher has measured what (s)he has set out to measure.578 It is often further subdivided into ‘internal’ and ‘external’ validity; internal validity being the confidence that the conclusions drawn are reliable and external validity being the extent to which the results of a study can be generalised to other situations or to other people. Generalisability is more commonly associated with quantitative methods. Although qualitative research findings are not statistically generalisable, they can be theoretically generalisable which refers to the capacity to apply the findings to other situations.564 Lincoln and Guba suggest the use of the word transferability rather than generalisability in the context of qualitative data.579 Within this study generalisability is referred to when discussing quantitative data and transferability, qualitative data.

Various threats to reliability and validity exist depending on the type of study and methods used. Johnson and Harris suggest that universally agreed standards for achieving reliability and validity within qualitative research have not been established due to the variable nature of qualitative research and individual study design.580 Some researchers argue that reliability is only of concern in quantitative research and that if social reality is constantly changing then reliability is not important. However, taking into account that a single social reality is a controversial assumption it would be neglectful not to triangulate or cross-reference findings with other studies in the same field.559

Using the framework developed and commonly cited by Campbell and Stanley (1963) the main threats to validity within this study are sample bias, Hawthorne effect, halo effect and experimenter or observer bias.581 The latter three consider the conscious and unconscious influences of both the participant and observer (cognitive biases). The Hawthorne effect takes into consideration that participants being observed may alter their behaviour in an attempt to appear more efficient. This means that any findings are likely to be real and a best case scenario, but may be an underestimate. The potential Hawthorne effect could be minimised through covert observation, though this is not usually deemed 113 ethically acceptable. Previous observers of health-related situations have noted that their impact is often minor due to the complex, demanding work needed to be undertaken in a timely manner. Strong observed that his own impact on the setting was likely to be minimised by the presence of multiple other audiences including nurses, other doctors, medical students and so on.582 The researcher was only one small part of the audience in the setting under study. Recently the interpretation of the original work which uncovered the Hawthorne effect is also being questioned due to the quality of the methods used in the original study and the conclusions drawn from the data.583 The Halo effect is a cognitive bias where the judgments of a person’s character is influenced by the overall impression of him or her by the researcher and the values of the researcher.584

Observer bias is sometimes confused with the Hawthorne effect. It acknowledges the potential contamination effect of the researcher’s preconceptions. Silverman (1993) highlights that the researcher will always enter the field with some perspective or broad focus and it is important to be as explicit as possible about this. The researcher must be aware of their own cultural assumptions and any preconceived ideas especially given the researcher’s additional, ongoing role as a pharmacist. This can be countered by being reflexive. This will be discussed further in Chapter 5.

Maxwell (1992) also highlights further areas which may affect validity in terms of qualitative data.577 These include description, interpretation and theory. ‘Descriptive validity’ refers to the accuracy of the data and how well it reflects what the participant has said or done.577 This means any transcription must be an accurate account or portray the events in an accurate manner. Descriptive validity is the base on which all the other types of validity are built. Without an accurate account of the initial data, all else is thought to be irrelevant.585 ‘Interpretive validity’ describes how well the researcher reports the participants’ meaning of events, objects and/or behaviours.577 The interpretations must not be based on the researcher’s perspective, but that of the participant, which presents very particular challenges in observational studies such as the stage 3 study here. This was addressed where possible by asking questions of the participants to clarify observations. Lastly theoretical validity ‘goes beyond concrete description and interpretation’(p291) and addresses the theoretical underpinnings that the researcher uses or develops during the study.577 114

Threats to reliability and validity can be minimised through the employment of different strategies such as mixed-methods where the findings can be triangulated and corroborated as previously discussed (see section 3.5), providing credibility (internal validity). Denzin [555], however, also highlights the importance of within-methods triangulation to enhance claims on the validity of the conclusions drawn (see table 3.3).

Table 3.3 Types of triangulation described by Denzin (1978) and their application to this study555

Type of Further division of types of How triangulation is achieved triangulation 555 triangulation586 within the study Data Time triangulation – collected Data were collected over an 8 triangulation on the same phenomena over a week period at each different period of time site

Space triangulation – from Four sites were included within multiple sites the study Person triangulation – Individual HCPs and patients collected at three levels e.g. were observed. The data were individual, group, department also analysed by professional group, by hospital and specialty Individual HCP group, specialty

Investigator Using more than one observer Co-raters were used to minimise triangulation to reduce observer bias observer bias Multiple A combination of multiple Medicines reconciliation triangulation methods data types, observers literature was used to develop and theories are combined an observational framework. within the same investigation This was used with other relevant theories e.g. human factors to analyse the data Methodological Between method where a Both quantitative and qualitative triangulation variety of different methods are approaches were used used i.e. quantitative and qualitative

115

3.5.1 Reflexivity

Qualitative research is inevitably influenced by the researchers values.587 Reflexivity looks at the effect that the researcher’s background attitudes, values and beliefs have on the research process. It recognises that the researcher is not a neutral observer and may be part of the setting or context that they are seeking to understand. Miles and Huberman suggest that any researcher enters the field with some orienting ideas, motivations and tools.560 These are subsequently implicated in the data gathering and construction of knowledge. Self-reflection is therefore needed to determine any bias that may occur, and inadvertently influence the research process. Other influences, such as gender and appearance, must be also considered in observational research. Warren suggests female fieldworkers have greater communicative skills and are less threatening than their male counterparts, potentially giving them greater access to the field.588 This may have potentially improved access within this study. The researcher documented any potential feelings which they felt may bias the observation or analysis within the field notes.

3.5.2 Inter-rater reliability and minimising observer bias

During observation humans are used as a part of data collection. This can lead to questions about the consistency between recordings and the potential misinterpretation of phenomena being observed. The potential for different researchers to view phenomena differently when observing the same event can be reduced through more structured observation, and investigator triangulation to ensure inter-rater reliability. Within this study inter-rater reliability was addressed through a process of co-rating. This is described in further detail in Chapter 5.

Inter-rater reliability when coding data (also known as multiple-coding) aims to counteract potential subjectivity during data analysis.589 The nature of any disagreement is important together with any subsequent discussion that leads to the refining of codes. Also inter-rater reliability is also tested within Chapter 4 to determine the extent of agreement between the pharmacists categorising the potential severity of the medication discrepancies. In both cases a statistical approach was taken; the agreement between co-raters is measured using

116 weighted Kappa coefficient in Chapter 4 and Kappa coefficient in Chapter 5. These are described further within each chapter.

3.6 Sampling and Sample Size

Brannen550 explains that statistical sampling (associated with quantitative data) and theoretical sampling (associated with qualitative data) strategies can be used in combination to determine which cases are representative. Several considerations regarding quantitative and qualitative sampling are discussed below.

3.6.1 Quantitative sampling

In quantitative research, two main factors influence the conclusions drawn from a sample. The first is the sample size and the second the extent of variation in the sampling population. The general rule is the larger the sample size the more accurate the findings, however, the greater the variation in the study population the greater the uncertainty – in statistical terms the greater the standard deviation the greater the standard error.590 A power calculation must be undertaken to calculate the minimum sample size required so that it is likely to detect an effect of a given size. Quantitative sampling and power calculations are used in Chapter 4 to determine an appropriate sample size through which the number and type of discrepancies can be quantified.

3.6.2 Qualitative sampling

Various sampling methods can be used to elicit qualitative data; the approach used is influenced by the research methodology. Non-probability sampling techniques are more often used in qualitative research. This follows the theory that qualitative research is one of discovery rather than hypothesis testing. However, the argument for generalisation or transferability is stronger when findings are observed with different participants in different settings. Non- probability sampling methods include quota, purposive, theoretical, snowball and convenience sampling.

Purposive sampling (theoretical sampling) was used within this study as it offers some degree of control and allows the judgement of the researcher to

117 determine who can provide the best information and value to achieve the objectives of the study.591 It also allows information-rich cases to be identified which can be studied in depth.592 Instead of selecting a cross section of the population using probability-sampling methods such as random systematic and stratified sampling, purposive sampling aims to produce a variety of cases; mainstream and potentially extreme cases. This sampling method is not concerned with representativeness but with representing the diversity within the concept being studied.559 It deliberately seeks to include outliers or deviant cases.591 It is useful to describe a phenomenon or develop something of which little is known, in this case medicines reconciliation. Thus, the hospitals included in the study varied in size, patient population profile, teaching hospital status and medicines management systems.

Whilst it was known which hospitals and wards were going to be investigated at the outset, it was unknown how many HCPs and patients would be observed throughout. The sample was ‘emergent and sequential’ as decisions were made as the research progressed as to which parts of the process and which HCPs required further study and where attention should be focused during observation. Participants were continually selected until the research reached theoretical saturation when the data started to confirm the analysis rather than provide anything new.559 There were a large number of observed patient encounters. This, however, was necessary to identify cases which provided the richer data. The researcher, as well as looking at common occurrences and similarities in process, needs to look for the outliers and include ‘special instances’ which are more extreme. Within medicines reconciliation the aim was to gain ‘maximum variation’ and a broad spectrum of problems which arise as well as the more mainstream day to day difficulties which occur.

3.6.3 Patient specialties observed during the study

Three specialties were chosen for the overall study; elderly care, renal medicine and general medicine. Elderly patients were chosen as they are often prescribed four or more medications,220 making it more difficult to capture a complete and accurate list of medication.145,177,197 Elderly patients may also exhibit a number of attributes including problems with recall, communication and comprehension which can make medicines reconciliation challenging. The 118 elderly, those with memory problems and those with increased number of medicines have been shown more likely to experience medicine related problems.593 General medicine was chosen to draw comparisons with a younger population, who are generally prescribed fewer medications and whose memory and comprehension usually, on average, greater functionality. In contrast, renal specialties were included, as through the nature of their condition, these patients often have complex medication requirements and are frequently admitted to hospital requiring continuous communication between primary and secondary care.290 Patients with end stage renal disease frequently experience high numbers of medication related problems specifically related to gaps in medication information at transitions in care.226 A significant proportion of DRPs have been shown in haemodialysis patients and medicines reconciliation has been shown effective in identifying and rectifying errors in this population.319,594

3.7 Access and Recruitment

‘Gate keepers’ are key people who can grant permission and can be negotiated with for access to places and facilitate contact, for example, within the hospital between the researcher and study group.595 The NHS is a public service but in a private setting (i.e. no public access) with various gatekeepers. Formal gatekeepers include managers e.g. matron or sister on a ward, whilst informal gatekeepers do not have institutional powers but may exert influence over the setting e.g. consultant physician of a medical team do not manage the ward, however exert influence over the management. Patton et al suggests a reciprocity model of gaining entry where both parties see mutual advantage in the research.592 In this case, many HCPs approached were keen to improve their medicines reconciliation practices. Observational methods can create particular challenges for host organisations. These include intrusion of privacy and disruption, additionally a study focussed on patient safety such as this, may be seen as a form of scrutiny, however, the HCPs’ willingness to improve patient safety predominated. Access during this research project was an ongoing process. The research occurred over a long period of time and the researcher required periodic access to various settings, people and professional groups. Access was gained by approaching relevant ‘Gate keepers’ for example the consultant, ward manager or practice manager. Specific 119 approaches for each part of the study are discussed further in the methods sections of Chapters 4, 5 and 6.

3.7.1 Relationships in the research setting

Whether or not those who routinely inhabit the research setting have a knowledge of social research, they are likely to be concerned with the researcher’s personality and behaviour. Berg suggests a number of strategies for successful observation which includes familiarisation with the environment and the process of observation but also the development of rapport with the participants.596

Developing relationships can be difficult as the researcher may be seen as over-friendly, intrusive, suspicious or threatening. Inevitably, the researcher will relate better to some participants than others, which can lead to a key informant being adopted.565 The various relationships must not adversely affect the researcher’s credibility or the focus of their attention. Termination of relationships once the researcher departs from the setting must be considered and whether any enduring contact is to be maintained, which could affect the subsequent data analysis.

3.7.2 Informed consent

It is considered unethical to collect information without the participant’s knowledge. Researchers must have consent from all participants prior to data collection. Informed consent demands that subjects are made adequately aware of the type of information being gathered, the reason for collection, the purpose, how they will be expected to participate, and how the research may affect them. Consent was collected from all participants involved in the observational studies. Specific information regarding consent for each of these studies is discussed in Chapters 5 and 6.

3.8 Data Analysis

As a mixed-methods paradigm is used within the study, both quantitative and qualitative data are generated. It cannot be assumed that qualitative data is exclusively generated through qualitative research.562 For example, open

120 ended questions included in questionnaires can produce qualitative components. It is the inductive reasoning used to interpret the significance and importance of the data which largely makes a study qualitative.562 Data is differentiated according to the assumptions and principles regarding truth and reality discussed earlier in the chapter. General principles regarding quantitative and qualitative data analysis are discussed below. Data collected during each stage has been analysed individually in Chapters 4, 5 and 6 and conclusions drawn separately. The findings from both studies are then brought together and discussed collectively in Chapter 7.

3.8.1 Quantitative data analysis

Quantitative data is used to describe data that can be measured and expressed numerically. The data focuses on numbers and frequencies rather than descriptions. It can be analysed statistically to determine relationships between variables. This is demonstrated in Chapter 4.

3.8.2 Qualitative data analysis

Qualitative analysis looks at relationships and links within data, the researcher attempts to understand why these linkages occur and then considers them in the context of other knowledge and literature.597 Various qualitative data analysis methods exist which either stem from the theoretical position e.g. interpretive phenomenological analysis (IPA)598 and grounded theory585 or methods which are independent of theory and can be applied across all approaches e.g. thematic analysis. Dey describes qualitative analysis as a circular process which involves continual describing, classifying and connecting the data (see figure 3.5).599 In a more detailed paper, Morse describes four intrinsic cognitive processes which occur irrespective of the qualitative data analysis method used: comprehending, synthesising, theorising and re- contextualising.597

121

Figure 3.5 Dey’s model of qualitative analysis 599

Describing

Qualitative analysis

Connecting Classifying

This study uses thematic analysis which Braun and Clarke600 argue is widely used but rarely acknowledged. It is a method for identifying, analysing and reporting themes within data. A theme captures something important about the data in relation to the research question. They can be identified at an explicit (content) or latent (interpretive) level, producing either descriptive or analytical codes.

Chunks of raw data are open-coded and as the codes take shape the researcher looks for relationships between the codes. This involves reading and systematically labelling data including key, essential, striking, odd and interesting items.601 Links and associations allow certain codes to be listed under broader headings where some codes seem more significant than others (axial codes). As codes emerge a hierarchical coding scheme with higher level themes containing a series of sub-themes are developed, eventually arriving at core codes to determine concepts that help to explain what is happening in practice. This determines how the categories relate to each other and explains why things happen as they do. Quantification of the qualitative data allows comparison with the quantitative data and also adds weight to the significance of the themes. This is also part of triangulation of the data. The coding process is summarised using Braun and Clarke’s phases of thematic analysis in table 3.4.600

122

Table 3.4 Six phases of thematic analysis 600

Phase Application to study 1) Familiarisation with the Repeated active reading of the data data searching for meanings and patterns

2) Generating initial codes Work systematically through the entire data set to produce initial codes from the data 3) Searching for initial Codes are sorted into potential themes and themes considered how they may combine to provide overarching themes and subthemes 4) Reviewing themes Themes are reviewed and refined. May become evident that some themes are not themes and may collapse into each other 5) Defining and naming Define and further refine themes – identify themes the essence of what each theme is about and determine what aspect of the data each theme captures

Identify the story that each theme tells and consider how it fits into the broader overall story in relation to the research question

Identify whether they contain sub-theme(s) – provides structure to particularly complex themes 6) Producing the report Concise and clear with evidence to support the themes and related to research question

Theoretical saturation is achieved through constant comparison of incidents until no new themes or dimensions are emerging and was aimed for within the study.552,602 This could not be predicted accurately at the start of the research but was reached when data collection ceased to contribute anything new to the codes categories or concepts developed. Specific details on the methods used for qualitative data analysis are discussed further in Chapters 5 and 6.

123

3.9 Ethical Considerations

All professions are guided by a code of ethics, especially within health. All have common principles such as avoiding harm, ensuring confidentiality and using information properly. These principles apply to the research process, from the collection of data, through analysis, to the dissemination of findings. This is enshrined in legislation to protect both participants and researchers. It includes respecting participants’ rights and dignity, avoiding harm and behaving with honesty and integrity. Participants must not experience any detrimental effects from the research process; and the long-term impact of the research on the participant should be considered including physical harm, psychological harm (discomfort and anxiety), and a sense of being intruded.575 Researchers must avoid deception and misrepresentation; providing factual and truthful information about the nature of the investigation and the role of the participants in the research before any involvement is considered.

Mason suggests that researchers need to develop a self-conscience and not simply, or indeed blindly, follow a strict ethical code.566 Within the NHS the ethical approval process is designed to protect the patients, staff and organisations within it.

Ethical approval for this study was considered separately at each stage of the research. The data collection described in Chapter 4 was considered within each organisation as a service improvement activity as it measured the rate of discrepancies and number of sources used against standards within the literature and facilitated a benchmarking opportunity for those involved. Thus formal ethis approval was not required for this study.

Ethical approval was sought and received from Bradford, Local Research Ethics Committee (LREC) on 21st March 2011 for the observation and documentary analysis phases of the study (Chapters 5 and 6) (see Appendix A). Additional university ethical approval is not required where NHS LREC approval has been obtained, however, the study was logged with the University of Bradford Ethics Department as per institutional requirements. Subsequently, applications were also sent to relevant NHS R&D departments to gain site specific approval for

124 each of the hospitals sites and the general medical practices. This was received from:

. NHS Bradford and Airedale – 5th May 2011 (see Appendix B) . Calderdale and Huddersfield NHS Foundation Trust – 16th June 2011 (see Appendix C) . Bradford Teaching Hospitals NHS Foundation Trust - 23rd June 2011(see Appendix D) . Airedale NHS Foundation Trust - 10th January 2012 (see Appendix E) . The Leeds Teaching Hospitals NHS Trust – 22nd August 2013 (see Appendix F)

However, during data collection, the rules regarding the ethical requirements for involving HCPs in research changed nationally and the study no longer fell within the remit of the LREC. A letter was received 21st August 2012 excluding their involvement (see Appendix G). See table 3.5 for summary of approval.

125

Table 3.5 Summary of ethics approval for study

Stage NHS LREC Bradford Airedale NHS The Leeds Calderdale NHS Bradford and Airedale Teaching Foundation Teaching and Hospitals NHS Trust Hospitals NHS Huddersfield Foundation Trust NHS Trust Foundation Chapter 4 Not required as considered Service Improvement

Chapter 5 21st March 2011 23rd June 2011 10th January 22nd August 16th June 2011 subsequently 2012 2013 informed didn’t need it on 21st Chapter 6 August 2012 5th May 2011

126

CHAPTER 4 STUDY 1: OBSERVATIONAL SURVEY OF MEDICINES RECONCILIATION FOLLOWING ADMISSION TO SECONDARY CARE

This work has been published in Research in Social and Administrative Pharmacy (see Appendix H):

Urban R, Armitage G, Morgan J, Marshall K, Blenkinsopp A, Scally A. Custom and Practice: a multi-centre study of medicines reconciliation following admission in four acute hospitals in the UK. RSAP. 2014. 10 (2) 355 – 368.

4.1 Introduction

Many studies have highlighted the high number of errors in medication history taking at admission which led to subsequent unintended discrepancies on IMCs.48-50,145,189,603,604 The most common error has consistently been the omission of clinically relevant medication from in-patient charts with an incidence of 39 to 72%.74-79,247,442 Despite the number of interventions which have shown to decrease error, 48,53,64,74,185,300,329,477,603 there is little to suggest, however, that practice has significantly changed over the last 15 years within the UK, creating the need for further improvement to try to reduce the number of medication errors which occur at admission.

Current UK guidance on medicines reconciliation is based on error rates from one Canadian and one US study, making its relevance and applicability to the UK uncertain.48,50,189 This chapter determines the current picture of unintended medicines reconciliation discrepancies within the hospitals studied. It compares these to existing data within national and international literature to contextualise the scale of problems with medicines reconciliation and highlight the relevance of national guidance. It quantifies the number and type of sources used to verify medication histories and measures the type and potential severity of UIDs identified during the process. Recognising that medicines reconciliation may be open to various interpretations, this chapter addresses the first two stages (according to the NPC’s 3Cs approach); collection (collection of the medication history by doctors at admission) and checking (subsequent verification of the medication history by pharmacy staff).80

127

4.2 Aims and Objectives

4.2.1 Aim

To determine current medicines reconciliation practice within one region of the UK and compare it to published best practices.

4.2.2 Objectives

 Determine the time taken for medicines reconciliation to occur following admission  Determine the number and type of sources used by pharmacy staff to verify the medication prescribed at admission  Quantify the number and type of UIDs discovered by pharmacy staff during verification of the medication history (post-admission)  Determine the potential severity of unintended discrepancies identified during verification of the medication history (post-admission)

4.3 Method

This is a prospective multi-centre observational survey conducted by pharmacists during the verification of medication histories at admission.

4.3.1 Participants

Organisations were purposively selected based on their status e.g. teaching hospital, patient population and their approach to medicines reconciliation e.g. pharmacy technician involvement and use of electronic discharge, identified through discussions with various hospitals. Consequently, a cross-section of four NHS acute hospitals in one region of the UK, including one of the largest university teaching hospitals in Europe, two district generals and a further teaching hospital, were recruited to increase the representativeness and diversity of data.

Ethical approval was not sought following agreement with the hospitals that this was a service improvement initiative.

128

4.3.2 Data collection

4.3.2.1 Development and piloting of data collection proforma

A data collection proforma was compiled using existing literature and validated audits.6,77 It included a taxonomy of medicines reconciliation discrepancies developed by the universities of Sydney, Leeds and Wisconsin (David Alldred. Senior Lecturer in Pharmacy Practice, University of Bradford. Personal communication 2011), which focuses on the categorisation of medication discrepancies, rather than the categorisation of medication error (see Appendix I). The indicators were chosen based on those which were most commonly cited in the literature and in discussion with experts in the field (Linda Dodds, Associate Specialist Lecturer, Medway School of Pharmacy, Personal communication, 21st October 2010 and David Alldred, Senior Lecturer in Pharmacy Practice, University of Bradford, Personal communication, 8th February 2011). Guidance notes were also developed to assist completion of the data collection proforma and to ensure consistency. The data collection tool and guidance notes were reviewed within the supervisory team plus also by two hospital pharmacists. Minor amendments to wording were made to increase clarification. Also, the data collection proforma was changed from landscape to portrait orientation to ease completion. The data collection tool was also piloted in 5% of sample (10 patients) by the researcher and two further hospital pharmacists, then data reviewed to check face validity.

4.3.2.2 Collecting the data

Prospective data were collected from the four hospitals (Hospitals A, B, C and D) from elderly care, medical and renal wards, over a total of 3 days per hospital, during the weeks commencing 5th and 12th September 2011. This data collection period was selected as it was deemed sufficient time to reach the calculated sample size needed to power the study. Hospital C was a single hospital split over two sites, data were collected from both. A period of two weeks was chosen to allow for natural fluctuations in working patterns and patients admitted to the ward.

Data were collected by 23 pharmacy staff undertaking routine medicines reconciliation post-admission. For each patient, the pharmacy staff initially

129 recorded the length of time between admission and when medicines reconciliation was conducted. (It was classified as being within 24 hours if the medication history was collected the same day or the next day). Next, the start and finish times, plus time taken were recorded. If the staff member had to break off to wait for information from general medical practices or for information to be confirmed by other HCPs, this time was deducted to give an adjusted time. The number of sources used to ascertain current medication and the number, type and severity of unintended discrepancies between the sources used and the medication history documented were also recorded (see Appendix I). An unintended discrepancy was recorded where no intentional reason for a difference between the medicines the patient was taking prior to admisson and the medicines prescribed on admission was identified. Guidance notes were provided with the proforma to ensure consistency of recording, together with a completed example proforma (see Appendix I). In addition, the study was explained verbally by RU (the principal investigator), either in person or by the telephone, to each staff member prior to any data collection.

4.3.3 Classifying severity

As discussed in the literature review, a number of different scales are used within the literature to rate the severity of error and or potential error. Some use severity i.e. lethal, serious, significant, minor65,210,218,219 Some look at discomfort and monitoring e.g. Climente-Martí et al189 and Arora et al.217 Others use the likelihood to cause harm e.g. unlikely, possible, probable.183 Each scale was reviewed for appropriateness and relevance. Ultimately a scale was chosen which assessed discomfort and monitoring e.g. Climente-Martí et al189 and Arora et al.217 as this was deemed the easiest to predict the potential severity. The scale chosen for this study reflects the potential impact rather than drawing on those scales which have routinely been used to establish the level of harm or the likelihood of recurrence (the traditional risk matrix) in an ADE. The severity scale used within the data collection proforma (Appendix I) was adapted from the work of Climente-Marti by adding in the prefix ‘If left unchanged, the discrepancy is likely to…’ to recognise that not all discrepancies result in harm and that some may be identified before they reach the patient.

130

The severity scale was pre-tested using a convenience peer group comprising four community pharmacists and four hospital pharmacists who were independent of the pharmacy staff involved in data collection. During the pre- test each pharmacist was asked to categorise 30 discrepancies which may arise during medicines reconciliation, according to their potential severity. This determined where they should fall within the scale and highlighted any difficulties using the scale. Participants were instructed to conduct the task individually, without discussion with other colleagues, so as not to influence their answers. The categorised discrepancies were used as examples in the guidance notes. A consensus or majority was reached in 28 of 30 categorisations, the remaining two were excluded. Following the peer categorisation, the wording within the guidance was amended from ‘would cause injury’ to ‘could cause injury’ on the severity scale to take into account the potential rather than actual risk.

During the data collection period, staff categorised the potential severity of each discrepancy. Categorisation was based on the assumption that had medicines reconciliation not taken place, and the discrepancy not detected then there would be impact on the patient (see Appendix I). At the end of data collection, all data sheets were collected and passed to RU for analysis.

The classification of severity was validated by taking 10% of each organisation’s data and verifying with the peer group of 8 pharmacists who had originally categorised the discrepancies to determine a consensus or majority. Each pharmacist rated the data from all four sites. Where the severity had not been recorded by a pharmacist during data collection the peer group reviewed the description of the discrepancy and classified the severity. The level of agreement between the peer group and the pharmacists collecting the data was calculated using Weighted Kappa.

131

4.3.4 Sample size estimation and data analysis

Using a sample size of 250 patients, each taking an average of 8 items, approximately 2000 prescribed items needed to be observed to detect relevant discrepancy rates between the four sites. Using previous studies the true discrepancy rate was estimated to be approximately 20% i.e. 400 discrepancies (95% CI ± 2%) providing an exact binomial confidence interval of (0.18 - 0.22).49,50,77

Data distribution was examined using box plots then data were tested for normality using Kolmogorov-Smirnov. Differences between hospitals were analysed using non-parametric tests (Kruskal-Wallis analysis of variance and Pearson’s chi-squared test.) Two outlier values were excluded from the ‘time taken’ data which were disproportionate to the other results (i.e. >500minutes). All tests were conducted using Stata® version 9, using a significance level of P<0.05.

4.4 Findings

A total of 250 medication charts were reviewed from the four hospitals (54 Hospital A, 61 Hospital B, 69 Hospital C, 66 Hospital D). In some cases the pharmacy staff did not complete the full data set. As a result the denominator varied for the number and type of sources used, the number and type of discrepancies and the categorisation of severity of the discrepancies. The denominator used is defined at the beginning of each relevant section. (See table 4.1 for hospital demographics).

4.4.1 Length of time before medicines reconciliation is undertaken post- admission

The majority of medication histories assessed were verified by a member of pharmacy staff within the first 24 hours post-admission with similar figures in all hospitals except Hospital A which was significantly lower (Fishers exact test, P = 0.005) (see figure 4.1).

. 132

Table 4.1 Hospital demographics

Hospital Hospital Number Medical Renal Policies in Standardised Training for Dedicated Dedicated Type of Beds Admissions Wards Place for Documentation Medicines Admission Admissions Medicines for Medicines Reconciliation Pharmacist Technician Reconciliation Reconciliation A Teaching 678 Separate elderly Yes Yes No No Yes on No Hospital (> 75 years) and elderly ward general medical not on general medicine B District 296 Combined No Yes Yes Yes Yes Yes General elderly and general medicine C District 735 Combined No Yes Yes Yes Yes Yes General elderly and general medicine D Teaching 1831 Separate elderly Yes Yes Yes Yes Yes Yes Hospital (> 80 years) and general medical

133

Figure 4.1 Percentage patients with medication history verified by a pharmacist

100

70 24 hours 60 48 hours 50 72 hours 40 96 hours 30 120 hours Percentage Patients (%) 20

0 All Hospitals Hospital A Hospital B Hospital C Hospital D

4.4.2 Time taken to undertake medicines reconciliation

The mean total time recorded for the pharmacy staff to complete the verification of the medication history from start to finish was 35.4 minutes. The mean adjusted time, after time taken waiting for information from other HCPs (e.g. the GP) was deducted was 14.8 minutes, a mean time difference of 21.0 minutes. The total time taken minus the adjusted time represented a total of 83 hours and 6 minutes. Seventy-five percent of medicines reconciliations took less than an hour to complete in Hospitals A, B and C from the time they were started to the time they were complete (adjusted time) (see table 4.2). The total time taken by Hospital D was significantly different to the other hospitals (excluding the outliers) (P=0.0001, Kruskal-Wallis) taking up to 2 hours in 75% of cases. The adjusted time between the hospitals was also significant (P=0.0117, Kruskal Wallis), however, this result may be spurious as a large proportion were reported at 10 minutes by Hospital A suggesting the time was not measured as precisely in this hospital. The distribution of data for adjusted time was very similar, demonstrating a degree of consistency (see table 4.3).

134

Table 4.2 Descriptive statistics for total time taken to conduct medicines reconciliation (minutes)

Hospitals A B C D Minimum 5 5 5 5 Lower Quartile (Q1) 5 5 5 10 Median 10 20 20 30 Upper Quartile (Q3) 60 60 60 120 Maximum 120 100 1440 360

Table 4.3 Descriptive statistics for actual time taken to conduct medicines reconciliation (minutes)

Hospitals A B C D Minimum 2 2 5 5 Lower Quartile (Q1) 5 5 5 5 Median 10 15 15 15 Upper Quartile (Q3) 25 30 30 30 Maximum 45 45 35 60

4.4.3 Sources used to conduct medicines reconciliation

The number and type of sources used were recorded for 247 patients. The number of sources used to reconcile medicines varied (see figure 4.2). Only one source of information was used in 36.8% (91/247) of cases, however, overall, Hospital D used significantly more sources than the other hospitals (P<0.005). The patients’ own medication was used in less than half of medicines reconciliations (110/247, 44.5%) (see figure 4.3), with no significant difference between the participating hospitals (P=0.130, Pearson’s Chi- squared). Discussions with the patient occurred in less than half of all medicines reconciliations (113/247, 45.7%) and there was a significant 135 difference between hospitals with Hospital D utilising the patient more often than other hospitals (P=0.001, Pearson’s Chi-squared).

Figure 4.2 Number of sources used to reconcile medication within each hospital

100 90 80 70 60 1 50 2 40 3 30 4

Percentage of Patients (%) 20 10 0 All Hospitals Hospital A Hospital B Hospital C Hospital D

Information from the general medical practice was used in 57.5% (142/247) patients with Hospitals C and D using the general medical practice significantly more than the other two hospitals (P<0.005, Pearson’s Chi-squared) (see figure 4.3). This included information received over the telephone, via fax or in a letter. Hospital D also used significantly more ‘Other sources’ than hospitals A - C. (P=0.048, Pearson’s Chi-squared).

Other sources included hospital pharmacy records, mental health hospital records and the community pharmacy records (see table 4.4)

136

Figure 4.3 Type of source used to reconcile medication in each hospital

100 Patient's Own Drugs (PODS) 90

80 GP Practice 70

60 Medicines Administration Record (MAR) Chart 50 Discharge Advice Summary 40

30 Patient

20 Patient Representative

Percentage of Medicine Reconciliations (%) 10

0 Other Source All Hospital Hospital Hospital Hospital Hospitals A B C D

Table 4.4 Other sources used to reconcile medication across all four hospitals

Source Number of times used Patient’s own list 7 Hospital pharmacy 5 Patient notes 4 Community pharmacy 3 Pharmacy ward technician 2 Mental Health Trust 1 Private hospital chart 1 Liverpool Care Pathway (LCP) 1 Warfarin clinic 1 Previous inpatient medication 1 chart (IMC) Renal unit records 1

137

4.4.4 Number and type of discrepancies identified

The frequency and type of discrepancy were recorded for 245 patients. The mean number of prescribed medicines taken per patient was 8.4. A total of 420 discrepancies were recorded. Of these, seven were omitted as they were documentation errors rather than unintended discrepancies e.g. missed strength of inhaler and 8 were reclassified as they fitted within the other 13 categories, leaving a total of 413 discrepancies from 2013 items. This yields a discrepancy rate of 19.6% (95% CI [0.180, 0.214]). The mean number of discrepancies per patient was 1.69. This varied between hospitals from 0.95 to 2.31 discrepancies per patient (see figure 4.4). The percentage of patients experiencing at least one discrepancy on their IMC was 37.6% (92/245). This varied between hospitals with nearly one in two patients experiencing at least one discrepancy in Hospital A (25/54, 46.3%) and Hospital B (28/61, 45.9%) (see figure 4.5). The difference, however, was not statistically significant (P =0.061, Pearson’s Chi-squared). Of the 413 discrepancies, 351 were categorised into discrepancy type with omission being the most frequent (237/413, 57.4%) (see table 4.5). The number of omissions per patient reviewed varied from zero to thirteen. The pharmacists recording the data provided detail of which medication was omitted in 169 cases. Inhalers (24/ 169, 14.2%) were the most commonly missed individual drug class, with pain relief (14/169, 8.3%) then nutritional and blood supplements e.g. ferrous sulphate and calcium D3 (10/ 169, 5.9%), proton pump inhibitors (10/ 169, 5.9%) and antidepressant next (10/169, 5.9%) (see figure 4.5). The more medicines a patient was taking the more likely they were to experience an unintended discrepancy (P<0.001, linear regression).

138

Figure 4.4 Number of discrepancies experienced per patient within each hospital

2.5 2.31

2.0 1.76 1.69 1.7

1.5

0.95 1.0

0.5 Numberof Discrepancies perPatient

0.0 All Hospitals Hospital A Hospital B Hospital C Hospital D

Figure 4.5 Percentage of patients experiencing at least one discrepancy on their inpatient medication chart (IMC)

50 46.3 45.9 45

40 37.6

35 33.3

30 26.6 25

15 Percentage of Patients (%) 10

0 All Hospitals Hospital A Hospital B Hospital C Hospital D

139

Table 4.5 Number of discrepancies identified and categorised by type

All Hospitals Hospital A Hospital B Hospital C Hospital D n % n % n % n % n % Drug Omission 237 57.4 55 57.9 34 58.6 59 52.7 89 60.1 Dose Decrease 20 4.8 3 3.2 6 10.3 3 2.7 8 5.4 Difference in timing 18 4.4 4 4.2 1 1.7% 5 4.5 8 5.4 Drug Addition 17 4.1 4 4.2 6 10.3 5 4.5 2 1.4 Formulation/ route of administration 15 3.6 2 2.1 3 5.2 1 0.9 9 6.1 Dose Increase 12 2.9 5 5.3 1 1.7 3 2.7 3 2.0 Frequency Increase 7 1.7 2 2.1 1 1.7 1 0.9 3 2.0 Other 6 1.5 0 0.0 1 1.7 3 2.7 2 1.4 Drug Duplication 4 1.0 1 1.1 0 0.0 1 0.9 2 1.4 Therapeutic class substitution 4 1.0 2 2.1 2 3.4 0 0.0 0 0.0 Strength Decrease 4 1.0 3 3.2 0 0.0 0 0.0 1 0.7 Frequency Decrease 4 1.0 1 1.1 1 1.7 1 0.9 1 0.7 Strength Increase 2 0.5 1 1.1 0 0.0 0 0.0 1 0.7 Change in Brand 1 0.2 0 0.0 1 1.7 0 0.0 0 0.0 Uncategorised 62 15.0 12 12.6 1 1.7 30 26.8 19 12.8 Total Categorised 351 85.0 83 87.4 57 98.3 82 73.2 129 87.2 Total discrepancies 413 100 95 100 58 100 112 100 148 100

There was borderline significance demonstrating the more sources used, the higher number of discrepancies discovered (P=0.0727, Kruskal-Wallis). There was no significant difference between the proportions of different types of

140 discrepancies within the hospitals except the difference in administration times where Hospital D was significantly higher than other hospitals (P=0.011, Pearson’s Chi-squared).

4.4.5 Severity of discrepancies identified

Three hundred and forty-four discrepancies were categorised according to their potential severity if left unchanged. The remainder (69) could not be categorised due to incomplete datasets. In Hospitals A, B and C the severity of discrepancies was most frequently categorised as potentially causing reversible injury which requires increased monitoring or intervention to preclude harm; however, the potential severity of discrepancies in Hospital D was most frequently categorised as not causing injury which may have skewed the overall results (see figure 4.6).

Figure 4.6 Categorisation of severity of discrepancies

50 1* 40 2*

30 3* 4*

Percentage (%) 20 5* 6* 10

0 All Hospitals Hospital A Hospital B Hospital C Hospital D

1* If left unchanged, the discrepancy would not cause injury 2* If left unchanged, the discrepancy could cause reversible injury with no changes to care 3* If left unchanged, the discrepancy could cause reversible injury which requires increased monitoring or intervention to preclude harm 4* If left unchanged, the discrepancy could cause reversible injury which would require additional care e.g. increased hospital stay, medication 5* If left unchanged, the discrepancy could cause irreversible or disabling injury 6* If left unchanged, the discrepancy could cause death

141

Figure 4.7 Omissions categorised by drug group

Percentage (%) 6

0 HRT PPIs Insulin Statins NSAIDs Antacid Quinine Inhalers Nitrates Diuretics Laxatives Dementia Eye drops Emollients Pain Relief Antibiotics Antiemetic Antithyroid Rubifacient Betablockers Glucosamine Antipsychotic Antiarrythmic Anticoagulant Antihistamine ACE Inhibitors Corticosteroid Antidiarrhoeal Mood Stabiliser Antidepressants Benzodiazepines Bisphosphonates Contraceptive pill Oral Hypoglycaemic Parkinson's Medication Antiplatelet medication Anti-obesity medication Calcium channel blocker Nutrition and Blood Supplements

142

Overall, there were four discrepancies that could have potentially caused irreversible injury or death. These included omission of insulin, anticoagulant therapy and prescribing penicillin in a penicillin allergic patient. Inter-rater reliability between the pharmacists recording the data and the professional panel was tested using Weighted Kappa (K=0.4048, std. err. 0.0483), showing moderate agreement.

4.5 Discussion

The results found varying levels of compliance with current UK guidance on medicines reconciliation. All participating centres strongly relied on GP-based sources, with this type of source being the only one used to ascertain the medication history in a substantial proportion of cases. This is against recommendations that information should be cross checked and verified.80 Patients were not involved in the majority of reconciliation encounters despite recent research indicating that patient involvement can improve safety.121,502 A significant number of discrepancies was identified by the pharmacists, of which, omission was the most common. Importantly, a minority of these discrepancies had the potential to cause harm to patients and therefore potentially increase the utilisation of healthcare resources.

4.5.1 Time taken to complete medicines reconciliation

The WHO recommends that medication histories should be verified by a member of pharmacy staff within 24 hours of admission.1 This occurred for the majority of patients included in the current study. Medicines reconciliation by pharmacy staff was often interrupted with the time from start to finish being much longer than the time spent on the task itself. It is unclear whether this time was spent waiting for information, whether the member of staff had been interrupted to undertake another task or some other reason. The idea of sharing access to the patient’s medical record between primary and secondary care could improve the time taken to reconcile medication, although this is not without its own difficulties.235 It could also be argued that the clinical significance of the time taken to verify a medication history is greater than the statistical significance as this may ultimately lead to delayed or omitted medication.

143

4.5.2 Sources of information used in medicines reconciliation

Many studies have looked at how to elicit appropriate information from patients through medication histories and have identified the sources of information which can contribute to accurate reconciliation 49,50,145,166,233,603. The current study demonstrates variability in the number and type of sources used within each hospital to reconcile medication, with some hospitals routinely using fewer sources. Although there was no statistical correlation between the number of sources used and the number of discrepancies found, there may be clinical significance in ensuring there is triangulation between sources. This is supported by Collins et al and Andersen et al, both of whom demonstrated the benefits of referring to a second source to corroborate the first during medication history taking, plus guidance from the National Prescribing Centre (NPC) which indicates that information should be cross-checked and verified.76,80

All four centres demonstrated a strong reliance on GP-based sources. Indeed, this was the only source in a substantial proportion of cases. It is often assumed that the GP record is the most accurate source of information; however, they have been shown to contain incorrect information and omissions.76,78,244 Collins et al76 found that 5% of medicines the patient was taking were on neither the GP nor nursing home record, and Green et al244 identified that 26.6% of discrepancies were attributable to inadequate or incorrect primary care information. This reinforces the need to utilise more than one source of information. Further work to identify whether the specific GP source used e.g. the repeat prescribing slip, GP printout or direct access to the patient’s GP electronic health record had an impact on accuracy may be useful. Hospital D used significantly more sources than the other hospitals which may account for the increased length of time taken to reconcile medication within that hospital.

‘Other sources’ were rarely used to verify the list of medications a patient was taking. These included the community pharmacist, mental health hospitals and warfarin clinics. Use of community pharmacy records has been shown to add value in some patients and is worthy of further attention to determine the reasons for low usage.107

144

4.5.3 Patient involvement in medicines reconciliation

Although early research indicates that safety can be improved through patient involvement, this study found that patients were not involved in the majority of reconciliation encounters.121,502 Patients have been encouraged since 1986 to bring their medicines in on admission to hospital as it reduces hospital expenditure, can increase the accuracy of medication histories and can improve the patient’s ability to accurately describe the medication they are taking.247 Over the years the percentage of patients doing so has increased and there has been a recent drive in many NHS hospitals to encourage patients to bring their medication, using methods such as the ‘green bag scheme’.249,361 Although such schemes have not been evaluated in the UK, promotional interventions have been shown to encourage ambulance staff to bring patient’s medication into the hospital setting in Australia which directly impacted on the increased accuracy of prescribing at admission and have recently been endorsed for international use by the WHO.1,333,334 Within the current study, patients’ own drugs were used in less than half the medicines reconciliations, with no significant difference between the participating hospitals. It was not possible to determine whether this was all patients who had brought their own medication into hospital or whether some patients’ own medication was available but not used. A subsequent study has shown that approximately 50% of patients bring in their own medicines to hospital which suggests the former.531 There is sometimes reluctance from patients to bring their own drugs into hospital for fear they may be lost and not returned at discharge.75 Leaving current medicines at home can, however, lead to the potential for discharge medication and home medication to be taken concurrently post-discharge.

The WHO suggests that to be optimally effective, patients and their families must be involved in the medicines reconciliation process.1 There is no evidence to suggest that the patient is more useful or accurate than other sources, however, involving the patient can reveal information not provided elsewhere.78,503 Studies of patient compliance have shown rates of 30-50% non-adherence, and demonstrated that, non-adherence can lead to increased medication discrepancies.194,288,499,605 Cochrane et al showed that a quarter of patients were taking regular OTC medication and that often the patient decides

145 themselves whether to stop or continue prescribed medication; information which cannot be gained from additional sources.379 Patients are best placed to be aware of all medication prescribed by multiple caregivers including self- purchased medication and are likely to know exactly what they are currently taking. It is unclear whether the patients who were not asked about their medicines were unable to communicate, asleep, lacked capacity, challenges previously highlighted by Collins et al or whether there was some other reason.76 Further research is therefore needed in this area.

4.5.4 Medication discrepancies identified

Within this study, the hospital which discovered the largest number of unintended discrepancies per patient did use more sources than the other hospitals; however, the statistical correlation was borderline between the number of sources used and the number of discrepancies found. This lends some support to the suggestion that more than one source should be used, but, further study is needed. Overall, four in ten patients experienced at least one discrepancy with a mean of 1.69 discrepancies per patient, (range 0.95 to 2.31 across hospitals). This is higher than a previous studies in the UK, the US and Canada.6,49,50,77 The UK guidance is based on the latter two.6,49,50

The results show that the more medicines a patient is taking the more likely they are to experience an unintended discrepancy, confirming findings by Duggan et al and Irvine et al.204,432 The most common discrepancy found was the omission of medication, which is in line with previous UK and international studies.74-79,247,442 The present study found a higher number of omissions than other studies, however, omissions were approximately 10% lower than two previous UK studies.49,50,76-79 It is unclear whether the hospitals with higher rates of discrepancies have poorer prescribing on admission or whether pharmacy staff are better at identifying the discrepancies.

Some omitted medications were potentially fatal e.g. insulin and warfarin, supporting previous findings.76,442 A number of studies have looked at common omissions in medication histories, the most frequently reported being the contraceptive pill, inhaled corticosteroid strength, type of inhaler device, brand of medication (where clinically significant), CAMS and OTC medication.76,79,432

146

The WHO suggests information should be collected on both prescription and non-prescription (OTC) medications including vitamins, nutritional supplements, potentially interactive food items, herbal preparations, and recreational drugs together with the dose, frequency, route, and timing of last dose of medication.1 The frequency of different types of medication omitted may have been higher but it was difficult to extrapolate this due to the lack of recording of type of medication omitted.

4.5.5 Severity of discrepancies identified

Although the majority of discrepancies were not classed as severe, a minority, importantly, had the potential to cause significant injury to patients or require increased monitoring of the patient. This may have meant an increased hospital stay or subsequent readmission, enhancing hospital expenditure, as well as increasing the distress of patients and undermining their confidence in existing systems.59 The findings in this study are similar to findings of Climente- Marti et al and Karkov et al but more severe than those cited by Lessard et al who determined that the majority of medication discrepancies identified were unlikely to cause harm to the patient.189,194,218

4.5.6 Strengths and limitations of Study 1

This part of the study demonstrates a number of strengths, most importantly that it was multi-centre and involved a range of specialties. This ensured that different patient groups and organisations were represented. However, hospitals B and C did not distinguish between general medical and elderly patients as their hospital admission wards contained mixed patient populations. This created insufficiently defined data to enable robust comparisons to be drawn between specialties. Within the other two hospitals the patient numbers within each specialty were too small to draw any conclusions. Further work looking at comparisons between specialties would be beneficial.

These data may also be limited as they were collected voluntarily by employees of the hospitals involved, relying on their engagement with the data collection and subjectivity in categorising the severity of the discrepancies identified. The varying level of engagement led, on occasion, to incomplete data. Alternative

147 methods could have been employed, e.g. the retrospective review of the medication charts and notes, however, due to differences in systems the relevant data was not always routinely captured within the hospitals. This would make retrospective evaluation difficult. The sample size in this study was relatively small - with fewer centres in comparison to the UK study conducted by Dodds et al.77 However, it does explore the nature and type of discrepancies in greater depth. Furthermore, the study had a larger sample size and more centres than the work of both Gleason et al and Cornish et al – both cited within national guidance.6,49,50

Additionally, the data collected were predominantly on admissions wards which may have skewed the results as patients often do not stay on this ward longer than 24 hours. It does though, give an indication of different practice within the hospitals in reviewing patients immediately post-admission. Ideally, studies of this nature should gather data from a wider cross-section of the hospital and including downstream wards, backed up with observational data of current practice.

It was acknowledged that the severity of discrepancy may be affected by a variety of factors including the patient’s profile, the specific medicine, strength, indication and co-morbidities, however, this was not taken into account for ease of categorisation and comparison.

4.6 Conclusion

This is the first UK study to present a detailed analysis of the medicines reconciliation process across four centres. Higher rates of unintended medication discrepancies per patient were found than in other studies; the majority were omissions, consistent with other research. None of the four centres adhered to current UK guidance on medicines reconciliation. Despite the evidence suggesting that at least two sources should be used to confirm the medication history, this is not happening in practice, with high reliance on data from general practices. The patient and patient’s own medication were used in less than half of all medicines reconciliations with much time being spent waiting for information from general practice. There is an increased need to educate staff involved in medication history taking on common omissions from

148 medication histories, particularly inhalers. Further qualitative research needs to be conducted on the barriers to obtaining information, including the reasons for utilising only one source. This chapter has provided baseline data to compare practice within the study population with published literature. It explains what is happening in practice, but does not explain why. To ascertain the reasons for the findings within Chapter 4 and to look at why these errors are occurring further more comprehensive observational work is needed. This will be explored in Chapter 5.

149

CHAPTER 5 STUDY 2: OBSERVATION OF MEDICINES RECONCILIATION IN SECONDARY CARE

Work within this chapter has been presented at the Health Services Research and Pharmacy Practice (HSRPP) and Royal Pharmaceutical Society (RPS) conferences and published in the International Journal of Pharmacy Practice. It has also been used to inform the national RPS Medicines Optimisation guidance.

Urban RL, Armitage G, Morgan J, Marshall, K. Is electronic discharge the answer? IJPP 2012, 20 (Suppl. 2): 40-1

Urban RL, Armitage G, Morgan J, Marshall K, Blenkinsopp A. Communicating about medication: Who, what, where, when, how? IJPP. 2013, 21 (Suppl. 1): 20-1.

RPS (2013). Medicines Optimisation: Helping patients to make the most of medicines. Good practice guidance for HCPs in England.

5.1 Introduction

The previous chapter identified the scale of error at admission to hospital from inadequate medicines reconciliation during the first two stages of the National Prescribing Centre’s (NPC’s) 3 C’s approach; collection and checking.80 It described the type and severity of discrepancies, which occurred between the medication the patient was taking prior to admission and the medication history taken at clerking, and, the sources used to verify the medication history. The findings demonstrated a high number of discrepancies of varying type and severity at admission to hospital within the study population. They were in keeping with previous UK and international literature which demonstrates high numbers of discrepancies, especially omission of medication.77,109,125,145,166,171- 183,201 Although Chapter 4 determined the scale of error within the health economy being studied, it did not identify the reasons why these errors were occurring or what was contributing to medicines reconciliation related discrepancies. The observation reported in this chapter seeks to explain the findings within the previous study and literature review, using observational methods. It will also build on the previous study by looking at the third ‘C’ within 150 the NPC’s 3C approach; communicating.80 Previous studies have examined stages of the medicines reconciliation process in isolation, focusing on one specific transfer point. They have segregated primary and secondary care, rather than looking at the impact on safety across all interfaces, and transitions of care. This chapter will look at the entire medicines reconciliation process within secondary care to identify contributory factors and defences which may lead to or ameliorate error associated with medicines reconciliation. The subsequent chapter will look at primary care and the interface. HCPs were systematically observed whilst undertaking the tasks that impact on the accuracy of medicines reconciliation: nurse handover, medicines administration, ward rounds, pharmacy staff verification of the medication history, patient counselling, prescribing & transcribing and the writing of discharge summaries.

5.2 Aims and Objectives

The aim of this chapter is to describe current medicines reconciliation practice within four acute hospitals together with the role of Health Care Professionals (HCPs) and patients. The specific objectives were to:

 Identify the process and stages of reconciling medication within secondary care  Determine the roles of secondary care staff and patients with respect to medicines reconciliation  Identify contributory factors which may lead to error associated with medicines reconciliation  Identify any practices and solutions which enhance the accuracy of the process or defences which prevent error

5.3 Method

5.3.1 Overview

This study employed qualitative observation using both deductive and inductive approaches to determine current practice within medicines reconciliation. The methods described here are based on the qualitative methodological theory described in Chapter 3; a pragmatic observational approach.

151

5.3.2 Sample

The four hospitals previously studied in Chapter 4 were used as the study sites. HCPs including doctors, pharmacy staff and nurses, were observed within these hospital settings undertaking the different stages which contribute to medicines reconciliation. Each stage was observed at least once in each area of each hospital being studied. The study used maximum variety sampling, therefore different participants were observed doing the same task, or different tasks.

5.3.3 Access to hospitals and recruitment of healthcare professionals (HCPs)

The researcher met with key personnel within each hospital to discuss the study prior to making an NHS Research and Development (R&D) application. Staff were assured that the tasks being observed were in line with the researcher‘s usual role as a pharmacist, bound by the GPHC Standards of Conduct, Ethics and Performance, which also contains a code of confidentiality.606 If a potential or actual, error or discrepancy, were noticed they were relayed to the relevant HCP. Post R&D approval the researcher then met with relevant personnel in the area/department being studied, for example, the ward sister, consultant, other relevant managerial staff, to explain the study and agree involvement. Prior to the observation, the researcher (where possible) identified members of staff scheduled to work during the study period and invited them to participate. The researcher also attended team briefings or similar to explain the study to staff members, and sought guidance from the ward sister or other managerial staff for an appropriate time to recruit participants. Consent was taken by the researcher prior to any observation. The study was explained verbally to the participant, supported by a written information sheet (see Appendix J). Their right to withdraw from the observation at any point and their right to refuse to answer any questions was explained. The researcher answered any questions the participant had and asked them to sign the written consent form (see Appendix J). Particular care was taken to ensure staff consented voluntarily and did not feel they were being coerced into participating. All data collected were anonymised. Unique identifiers were used to maintain anonymity and to enable the cross-reference of the consent forms with the data collection.

152

5.3.3.1 Participants

Consideration was given as to whether it was most appropriate to observe the HCPs undertaking medicines reconciliation tasks or the patient’s journey and experience through their hospital stay. It was felt that by observing the HCPs as the participant, it would capture the experiences of the patients without the difficulties of consenting patients. The observation of HCPs eliminated considerations such as mental capacity and provision of information in other languages which would have been necessary through the observation of patients. As this part of the study involved the observation of staff and did not directly involve the observation of patients, service users or their carers were not directly involved in the research process. Although by the nature of the study and the patient being the common factor in all the tasks observed they were indirectly involved.

Staff were recruited from three specialties: care of the elderly, renal and medical (see Chapter 3 for rationale), based on their job roles i.e. if they undertook an aspect of medicines reconciliation as part of their role, and their willingness to participate. Staff were also recruited in A&E within one hospital as it is known that for 20% of patients this is the start of their journey through hospital.607 Participants included doctors (all grades including consultants), nurses, pharmacists and pharmacy technicians within A&E and on the hospital wards. No HCPs refused to participate. The grade of the doctors was recorded as their freedom to act is linked to their grade, whereas with the other HCPs observed the same competence level is expected irrespective of their period of employment or level of experience.

5.3.4 Observation schedule development and pilot

5.3.4.1 Deductive approach

A Secondary Care Observation Schedule (SCOS) was developed, using findings from the literature review to evaluate tasks associated with medicines reconciliation within secondary care. Evidence-informed practices use the best available research and practice knowledge to guide program design and implementation. This informed practice allows for innovation while incorporating the lessons learned from the existing research literature. Where evidence was

153 available evidence was used. Where national policy was available national policy was used. Lastly where expert opinion and consensus was published this was applied. Failing this, professional judgement and knowledge was applied by the researcher to enable a pragmatic approach to the observation. It incorporated ten sections: organisational policy, training, audit, admission, taking a history, prescribing the IMC, verifying the medication history, transfer, discharge and patient education (see Appendix K). Each section was divided into various standards or criteria which should be incorporated into a medicines reconciliation process, informed by the literature. The structured approach allowed systematic collection of data, and increased the consistency of data collection and analysis across organisations. The organisation’s policies, protocols and procedures were gained in advance and their content compared with the evidence based observation schedule to identify and include any relevant items not on the schedule (see Appendix L for e-mail request). In fact, no additional items were found in the policies, protocols or procedures. The SCOS was piloted through the observation of a pharmacist and consultant (for 4 hours in total) who were not likely to be involved in the study to ensure it was manageable and feasible to complete during observation. No amendments to content were made, but following the pilot the document’s sections were saved as separate files to allow easier navigation rather than as one large document.

5.3.4.2 Inductive approach

During the pilot and early observations it became apparent that there were non- process related observations which potentially impacted on the medicines reconciliation process and information gathering; nuances and ways of working which pertained to the individual. Thus, any observation associated with medicines reconciliation which fell outside the parameters of the schedule was recorded using field notes. Field notes also included observations related to tasks which were not incorporated into the schedule due to lack of current evidence in the literature, for example, nurse administration and ward round.

154

5.3.5 Observation process

Wards were observed Monday to Friday between 7.30am and 6.00pm. The researcher arrived on the ward for nurse handover which usually occurred at 7.30am. The study was explained to everyone working on the ward that day. Handover was observed followed by nurse administration of medication. From this point forward the researcher would either join the ward round or observe pharmacy staff verifying medication histories depending on what was happening on the ward at the time; consenting HCPs as appropriate. Writing discharge summaries, transcribing and clerking usually happened post ward round, often in the afternoon. Patients were often discharged late afternoon once their medication had arrived from pharmacy. A typical observation day would usually last 8 hours with half an hour lunch break. Lunch was usually taken at the same time as the doctors as the majority of medicines reconciliation observations were reliant on them being on the ward.

Observations within A&E were conducted weekdays and weekends at various times throughout the day between 8.00am and 10.00pm. These were chosen based on discussion with the A&E Matron to observe the busiest periods to obtain maximum data. The sister in charge was notified on arrival, then the researcher chose an area of A&E to observe depending on activity. This included the high dependency unit, triage, minor injuries, and main A&E.

On the wards and in A&E, when observation was not taking place, the researcher stood near the nursing station to wait for the next opportunity to observe. The researcher chose what to observe based on their knowledge of hospital processes. The routines on the ward, e.g. ward rounds and medication administration, were most often at fixed times, compared with pharmacists verifying medication histories and discharge advice notes which were ongoing throughout the day. The researcher also looked for potential red flags based on their clinical knowledge e.g. if they had observed an alteration to the patients medication on a ward round then observation of their discharge summary preparation would be observed, if possible. Often the staff would notify the researcher if they were about to undertake a relevant task which facilitated observation.

155

The researcher followed and observed each HCP conducting separate stages of medicines reconciliation. Each observation of an individual varied from 0.5 to 4 hours, and for no longer than 8 hours in total per individual over the course of a week. The same participant was observed single or multiple times, conducting the same task or different tasks, for example, a doctor may be observed admitting a patient to the ward, then observed discharging the same or a different patient.

During observation HCPs were asked non-leading questions to clarify and document the process accurately, their replies documented in field notes, verbatim where possible, or described in detailed notes if the response was lengthy and required clarification and/or probing. There was potential inconvenience to the participant being observed as they had to answer questions between caring for patients. This was minimised through grouping questions together, not interrupting important tasks and keeping questions to a minimum. In theory, participants may have found being observed stressful but were able to opt out of the observation at any time if they found it intrusive or felt it may affect the care of their patient. This did not happen during the observations.

Whilst observing, criteria were marked as met or not met and other identified practices, together with answers to specific questions were noted. Participants often vocalised and shared their thoughts with the researcher which gave a useful insight into the thought processes and allowed some conclusions to be made on potential contributory factors and defences to poor medicines reconciliation practice.

Patients were asked by the HCP being observed if they would object to a pharmacist researcher present during their consultation. No patients refused. There were several occasions where the researcher felt it appropriate to remove themselves from the consultation with a patient, for example, during intimate examination. During this time the conversation could still be heard, and was duly documented if pertinent to medicines reconciliation.

Within the inductive part of the study, observation was conducted until theoretical saturation of information had been achieved and no new themes

156 emerged.552 This was when the information was adding to the bulk of the coded data rather than the theory. If new categories emerged half way through and these became theoretically saturated then it indicates nothing has been lost by not having coded this theme in the first half of the data.559

5.3.6 Recording of data

Data on the process were recorded against the SCOS using a combination of tablet with handwriting recognition and paper-based notes. The researcher also contemporaneously recorded anything of note which contributed or potentially contributed toward successful or unsuccessful medicines reconciliation using field notes on the reverse side of the SCOS or in a separate Word® document if recording electronically. Field notes were unstructured other than ensuring the date, HCP being observed, and setting of the observation was noted. Abbreviations were used where possible to make recording easier. Also, where appropriate, any relevant verbal data were recorded verbatim during the observation e.g. where HCPs explained their rationale for conducting a task a certain way. These included the participant thinking aloud or offering information which they felt helpful, either spontaneously or in response to questions. The time the medication history was started and finished was also recorded. The researcher asked further questions to increase clarity and extract further information from the participants, also paraphrasing information to ensure information had been interpreted correctly.

5.3.7 Data analysis

The data were analysed and categorised using three frameworks. Firstly, using the SCOS at individual stages of three medicines reconciliation process (a deductive approach). Secondly at a theoretical level, using a human factors approach, specifically the Yorkshire Contributory Factors Framework (YCFF) developed by Lawton et al.608 The YCFF takes a human factors and systems approach, identifying potential factors which may contribute to error e.g. patient factors and physical environment etc. It summarises factors (active and latent failures) contributing to patient safety incidents in hospital settings (see figure 5.1 and table 5.1).608 Their ultimate aim was to review the literature and to identify potential areas where interventions could be targeted.

157

Figure 5.1 The Yorkshire Contributory Factors Framework608 (reprinted with kind permission)

158

Table 5.1 Yorkshire Contributory Factors Framework (YCFF) Criterion Definitions608 (reprinted with kind permission) Factor Definition Active Failures Any failure in performance or behaviour e.g. error, mistake, violation of the person at the ‘sharp-end’ (the health professional) Communication Effectiveness of the processes and systems in place for the Systems exchange and sharing of information between staff, patients, groups, departments and services. This includes both written (e.g. documentation) and verbal (e.g. handover) communication systems Equipment and Availability and functioning of equipment and supplies. Supplies External Policy Nationally driven policies/ directives that impact on the level of Content quality of resources available to hospital. Design of The design of equipment and supplies to overcome physical Equipment and and performance limitations. Supplies Individual Factors Characteristics of the person delivering care that may contribute in some way to active failures. Examples of such factors include inexperience, stress, personality, attitudes. Lines of Existence of clear lines of responsibility clarifying accountability Responsibility of staff members and delineating the job role. Management of The appropriate management and allocation of staff to ensure Staff and Staffing adequate skill mix and staffing levels for the volume of work. Levels Patient Factors Those features of the patient that make caring for them more difficult and therefore prone to error. These might include abnormal physiology, language difficulties, personality characteristics (e.g. aggressive attitude). Physical Features of the physical environment that help or hinder safe Environment practice. This refers to the layout of the unit, the fixtures and fittings and the level of noise, lighting, temperature etc. Policy and The existence of formal and written guidance for the Procedures appropriate conduct of work tasks and processes. This can also include situations. Safety Culture Organisational values, beliefs, and practices surrounding the management of safety and learning from error. Scheduling and Adequate scheduling to manage patient throughput minimising Bed Management delays and excessive workload. Staff Workload Level of activity and pressures on time during a shift. Supervision and The availability and quality of direct and local supervision and Leadership leadership. Support from Availability and adequacy of central services in supporting the Central Functions functioning of wards/units. This might include support from Information Technology and Human Resources, portering services, estates or clinically related services such as radiology, phlebotomy, pharmacy. Task Factors related to specific patient related tasks which may Characteristics Make individuals vulnerable to error Team Factors Any factor related to the working of different professionals within a group which they may be able to change to improve patient safety Training and Access to correct, timely and appropriate training both specific Education (e.g. task-related) and general (e.g. organisation-related).

159

The YCFF was used to categorise codes identified by the researcher through thematic analysis. As well as using the YCFF to categorise contributory factors, the factors were also used to categorise those elements which may act as a defence e.g. patient factors such as hearing impairment may act as a contributory factor but involving the patient in the consultation may act as a defence. Some of the qualitative data were quantified to contextualise the observations, for example counts of the numbers of discrepancies observed or reasons for stop, starting altering medication. Thirdly, a range of interpersonal communication was explored and a health professional-patient consultation models was employed as a framework to organise themes associated with specific medicines reconciliation communication between the healthcare team and their patients (see 5.3.7.1).

5.3.7.1 Communication systems - Verbal and written information analysis

It is well established that lack of communication, amongst other issues such as inaccurate information, contributes to error, in particular medication error. It is one of two contributory factors within the YCFF (the other being ‘safety culture’), that intersects all other factors (see figure 5.1). This is because communication issues can arise at an individual, team or organisational level and may be influenced by an individual’s actions or by the team and their organisational and team cultures. Lingard et al provide a typology of communication failures within the healthcare setting, however, the contributory factors or defences which may contribute to these failures are not discussed.609 Furthermore the YCFF itself does not provide a detailed exposition of communication or safety culture because the systematic review on which it is based did not generate evidence to underpin this. For data analysis in the current study it was considered necessary to gain a deeper understanding of the relationship between communication and medicines reconciliation.

A number of communication models were considered and from these a framework was developed to categorise the inductive data relating to communication thus expanding on the YCFF. The types of communication observed during medicines reconciliation were analysed (see table 5.2) to determine which communication models may be appropriate. Two types of

160 communication were observed: written and verbal. Written communication was one-way (i.e. provided in one direction), for example to provide information to the hospital (from primary care), within the hospital or from the hospital (to primary care). A detailed analysis of written information was not undertaken as the focus of the observation was to gain broad contributory factors which could then be explored through further work. It was not always practical for the researcher to see the notes during the observation as the researcher was trying to remain unobtrusive. There is potential for the written communication observed in the study to be evaluated through information seeking models. The investigation of the notes may be better through a more focused study using documentary analysis with think aloud/cognitive processing techniques to understand the rationale and interpretation of assumptions being made whilst reading them. This was outside the scope of this thesis.

Verbal communication was two-way, to either provide or gather information, and involved an interaction between two people. This was either face-to-face or on the telephone. HCPs communicated with either patients or with other HCPs that were either within their team and/or discipline or outside.

Interaction between the HCP and the patient occurred at all points in the medicines reconciliation process. There were times when this interaction was more structured e.g. during clerking, the ward round and verification of medication history by pharmacy staff. These interactions could be described as a consultation, thus consultation models were explored. The structure of clerking and the ward round follow the traditional medical model of consultation: history taking, examination, investigation, diagnosis, treatment and follow up. However, within these consultations the HCPs displayed nuances in content, approach and style, which potentially affected the quality of communication, not captured within the medical model. The majority of consultation models encourage patient involvement from the viewpoint of improving patient satisfaction and shared decision-making, rather than from a patient safety point of view, making it difficult to find a single model. Consequently, within the current study, a consultation analysis framework specifically relating to medicines reconciliation was developed based on the Clinical Observation Tool (COT) criteria610 (see table 5.3) and essential communication skills for a

161 medication history interview611 (see table 5.4) to assess the interaction between HCP and patient.

Table 5.2 Communication during medicines reconciliation Type of Communication by HCP

Written Verbal

One-way – provision of Two-way – provision & information receipt of information

(Face to Face or Telephone)

Communicated to the Aide Memoire Consultation Patient Discharge Summary Counselling on Medication

Communicated to the Discharge Summary Seeking Giving Advice HCP Information Patient Notes (Same/ different Case Discussion Medication Charts discipline/team) Referral Observation Charts Delegation Proforma Handover

Education/ Mentoring

162

It was more difficult to determine a suitable model to analyse the HCP-HCP interactions as this is not well explored within the literature. Thistlethwaite and Spencer suggest a number of interactions which occur outside the consultation room based on doctor interaction.612 These would be appropriate for any HCP but only categorise the types of interaction not how they may fail or be successful (see table 5.5). The frameworks were applied post-coding to determine whether the criteria within the frameworks were met during the consultation.

Generic communication models were also explored, including the linear model suggested by Shannon & Weaver613 and the interactive model suggested by Schramm614, though these only described the sender and the receiver rather than what was being communicated and how. Berlo’s model of communication was considered further as it had some potential to apply to one-way communication (see figure 5.2).615 However, during two-way communication the patient and HCP interchange positions between sender and receiver making Berlo’s model difficult to apply. It would also need the researcher to have more detail of the content of the documentation to analyse the message for content and structure etc. which was outside the scope of this study. This again could be explored through further work.

Table 5.3 COT assessment criteria adapted for medicines reconciliation 610 The HCP encourages the patients contribution at appropriate points in the consultation

The HCP responds to signals/cue that lead to a deeper understanding of the problem

The HCP uses appropriate psychological and social information to place the complaints in context.

The HCP explores the patients’ health understanding.

The HCP asked sufficient information to include or exclude likely medication.

The HCP appears to make a clinically appropriate working diagnoses.

The HCP speaks to the patient in appropriate language and avoids jargon.

The HCP makes effective use of resources.

The HCP documents anything that needs follow-up or review.

163

Table 5.4 Essential communication skills for a medication history

interview taken from Ellington et al611 (reprinted with kind permission) Skills Definition Example Formal Use patient's title and last name ‘Good morning Mr. Smith.’ form of address Rapport Use active listening skills to ‘It's not easy being in the hospital confirm interest in a patient and away from friends and family.’ help gain respect Active Reflect on a patient's feelings or ‘You sound unsure.’ OR`’Are you Listening/ identify a patient's underlying saying’ Empathic message Responding Open- Patient is free to answer in any ‘How are you taking your blood ended manner. This is useful when pressure medicine?’ questioning introducing a new subject. Closed- Patient can answer with a single ‘Do you take your blood pressure ended word, usually yes or no medicine in the morning?’ questioning Transition Verbally closing off one subject ‘We have just talked about the and introducing a new one allows prescription medications you take. the patient to make a mental Now let's talk about any transition nonprescription medications you may take.’ Verbal Technique to get the patient to ‘....dizzy spells?’ Following elaborate on a subject without asking more questions, but simply repeating the patient's last few words Avoidance Leading questions prompt the ‘You don't smoke, do you?’ of leading patient with a particular answer REWORD: ‘Do you use any questions tobacco products?’

Avoidance ‘Why’ questions can cause patients ‘Why were you taking the medicine of ‘why’ to get defensive. Rephrase in the morning?’ questions questions to start with, ‘for what REWORD: ‘For what reason...’ reason’ Timing Warn a patient that a series of ‘I am going to ask you a series of questions will follow questions now.’ Clarify Always accept the blame for ‘I must have written it incorrectly, I Conflicting inconsistent information that the thought you had said....’ Information patient may tell you or write Silence Allows the patient to show Maintain nonverbal facilitation and emotion, digest information, or stop speaking gather thoughts Answering Avoid definitive answers until a Patient asks, ‘Do you think I should Patient final drug therapy plan is devised stop taking...?’ Pharmacist says, Questions ‘Well, I'll make a note and evaluate it with Dr. Smith.’ Mentioning If a patient answers a question you ‘You mentioned earlier that you Previously were going to ask later in the occasionally take ibuprofen for Answered interview, jot it down headaches. Do you ever take Questions anything else for aches or pain?’

164

Table 5.5 Breadth of doctor interactions outside the consultation room (adapted from Thistlethwaite and Spencer, 2008)612 (reprinted with kind permission) Doctor-patient communication  Consultation by telephone  Letters and email to patients Doctor-patient's family/caregivers  Face to face with or without patient present  Telephone, letters, and email Doctor-doctor  Professional  Seeking or giving advice  Appraisal and mentoring  Doctor as patient Doctor-other health professional  Referral and delegation  Seeking or giving advice  Case discussion Within a team  Awareness of difference in professionals' language  Assume the roles and responsibilities of a team member and leader Writing medical records  Legal  Court appearances  Reports Working with media, authorities, other organisations and complaints  Against oneself including admitting error and apologising  About colleagues Whistle blowing Running meetings Teaching

165

Figure 5.2 Berlo’s SMCR Model of Communication616

5.3.8 Validity and reliability of methods used

The same researcher was used throughout the study to conduct the observation to eliminate inter-operator variability. Four different co-raters were used at various stages throughout the study to validate the data and the analysis of the data. The process is outlined below and summarised in figure 5.3. The co- raters were selected to bring expertise relating to medicines and patient safety theory. Three were qualified pharmacists and understood the complexities of medicines reconciliation and clinical processes surrounding medicines reconciliation. One co-rater was a health psychologist with experience of patient safety research (see table 5.6).

During observation on the wards, the researcher was observed by Co-rater 1 who independently observed alongside the researcher for four hours to enable validity and reliability of the observations to be assessed. Following the dual observation, Co-rater 1 and the researcher discussed and compared the observations to clarify interpretation. It was surprising to find no differences in observation were recorded. This in part may have been down to the structured observation schedule.

Once all data had been collected, emerging codes were recorded by the researcher using NVivo10®. A sample of data (approximately 5% - 27 patient encounters) was then coded independently by a second co-rater (Co-rater 2). 166

This included a cross-section of hospitals and tasks; ward round, verification of the medication history by a pharmacist, administration of medication and clerking the patient. The emerging codes were cross-checked with the researcher's codes and a consensus reached. The codes relating to specific tasks associated with medicines reconciliation were categorised using the YCFF as to whether they were a potential defence or contributory factor.

Approximately, 10% of the codes (83) were then also categorised by two co- raters (co-rater 3 and 4) to determine inter-rater reliability. The co-raters were provided with the definitions of the contributory factors as described by Lawton et al.608

Table 5.6 Summary of co-rater characteristics

Co-Rater 1 2 3 4 Profession Pharmacist Pharmacist Pharmacist Health Psychologist Area of Clinical University University Patient Safety Practice Pharmacist Senior Senior Research (Teaching Lecturer & Lecturer Hospital) & Clinical University Pharmacist Teacher (Teaching Practitioner Hospital) Worked in Yes Yes No Yes Secondary Care Experience Yes Yes Yes Yes of Qualitative research Experience No No No Yes of Patient Safety research and Human factors

167

Figure 5.3 Process of ensuring validity and reliability within the observation data

Co-rater 1 observes researcher to validate observation

All Data reviewed and emerging codes recorded by Researcher using NVivo 10

Sample of data coded independently by Co-rater 2. Codes cross checked and verified with researcher codes

Codes categorised by researcher using the YCFF

Sample (10%) codes categorised by Co-raters 3 and 4

5.3.9 Ethical considerations for Study 2

The design of the study was discussed with members of the pharmacy department at Bradford Teaching Hospitals NHS Foundation Trust (BTFHT) and the National Institute for Health Research (NIHR) Programme Grant Patient

168

Panel at Bradford Institute for Health Research (BIHR). The approach was also discussed and agreed with the NHS Hospital's Caldicott Guardianf. Ethical approval was sought and received from the relevant LREC and R&D departments as outlined within Chapter 3.

5.4 Results

The data will be presented in four sections. Where frameworks have been used the data will be presented following the structure of the framework:

Section 5.4.1 Overview: hospitals and participants

Section 5.4.2 Analysis using SCOS (deductive)

Section 5.4.3 Analysis using YCFF (inductive)

Section 5.4.4 Communication systems (inductive)

5.4.1 Overview: hospitals and participants

One hundred and forty-two HCP participants were observed during 674 patient episodes within four acute hospitals in West Yorkshire, between September 2011 and November 2013. The participants included doctors, nurses, pharmacists (see table 5.7) from a total of 15 wards and one A&E department. Non-qualified HCPs were also observed including both medical and pharmacy students. The total observation time was 374 hours. A patient episode was defined as an interaction which involved discussion with or about a patient.

f A Caldicott Guardian is a senior person responsible for protecting the confidentiality of patient and service-user information and enabling appropriate information-sharing within an organisation.

169

Table 5.7 Number and Type of Participants

HCP Number of Participants Consultant 24 Specialist Registrar (SpR) 4 Core Trainee (CT) 1-5 7 Staff grade doctor 3 GP Registrar 7 FY2 16 FY1 14 Medical student 4 Pharmacist 21 Pharmacy student 3 Pharmacy Technician 11 Nurses 28 Total 142

5.4.1.1 Organisational information

See table 4.1 Hospital Demographics Chapter 4.

5.4.2 Analysis using secondary care observation schedule (SCOS)

In this section, results relating to the SCOS (see section 5.3.2 and Appendix K) are presented.

5.4.2.1 Organisational policy

All hospitals had policies in place which comprised taking a medication history, reconciling medication, transcribing, prescribing, patients own medication and writing discharge information (see table 5.8). These were accessible to all staff on the hospital intranet. None of the policies clearly defined the responsibilities of different staff within the process, though the prescribing and transcribing was aimed predominantly at doctors and the verification of medication by pharmacy staff. Staff were required to sign to say they had read these policies and procedures in Hospitals B-C. 170

Three out of the four hospitals (B-D) used standardised documentation to record medication histories. These included space to record the name, form, strength and dose of the medication, allergy status, sources of information used, an indication of whether the medication was stopped started or changed at admission, and the name of the person completing the form. None of the forms had separate areas to record OTC or CAMs or any prompts within the form to ask about these. None of the forms had space or prompt to record indication, compliance or height and weight of the patient as detailed in the SCOS. In Hospitals B and C the medication history form was used throughout the hospital. In Hospital D, the form was only used on the admissions wards, where it was incorporated into their clerking documentation. On the renal ward the form was not used routinely; the pharmacy staff would occasionally use it. Hospital A compared sources directly against the IMC.

All hospitals had a target to ensure medicines reconciliation was conducted within 24 hours laid out by NICE & NPSA.6 Within hospitals B–D there was a clear person accountable for improving medicines reconciliation within the hospital. All hospitals had access to interpreting facilities if required although no staff used them during the observation, relying instead on family or other HCPs who spoke the same language.

Hospitals B-D all had dedicated admissions pharmacists and pharmacy technicians on all of their admission wards. Hospital A had a dedicated pharmacist on their elderly admissions ward, conversely, the general medical admissions ward did not have consistent cover and the role of the pharmacy technician on all wards within this hospital was being developed.

5.4.2.2 Training and education

Hospital A did not provide training for any of its staff on medicines reconciliation, Hospitals B, C and D had training programmes in place for their pharmacy technicians of which Hospital C and D’s were the most comprehensive. None of the hospitals’ training was multidisciplinary and all involved only pharmacy staff. During the consent process for the observation, many HCPs (with the exception of pharmacy staff) had not heard the term ‘medicines reconciliation’ or if they had were unsure of its exact meaning.

171

Table 5.8 Content summary of hospital medicines reconciliation policies

Hospital A B C D

Hospital has specific policy on medicines √ √ √ √ reconciliation (MR)

The MR policy is part of the overall medicines √ √ √ √ policy for the hospital

Contains a definition of MR √ √ √ √ vague vague

Describes 2 levels of MR – basic and X √ √ X full/comprehensive

The MR policy defines individual roles and √ √ √ √ responsibilities lacks detail Detailed table summary Specifies that more than one source should be √ √ √ √ used Including the Including the Including the patient patient patient Specifies limitations of the sources √ X X √ limited Specific questions to elicit commonly missed √ √ X √ medication e.g. eye/ear/nose drops. Contains details of specific information that should √ √ X √ be documented with certain medicines e.g. target International Normalised Ratio (INR) for Warfarin, day of administration for weekly bisphosphonate

OTC and CAMS OTC X X OTC

172

Hospital A B C D

How to deal with discrepancies √ X √ √ Lacks detail

Documentation of discrepancies in the notes √ √ √ √ Including the patient

Format and Layout Structured The MR policies is The MR policy The Hospital has a document which an overarching forms one section MR procedure and addresses major document with 2 of the Medicines SOPs associated aspects of MR with associated SOPs Code. The with MR. The practical ways to which provide detail structure follows the SOPs provide an take a medication on how to take 3 Cs approach by overview of steps history medication History the NPC but lack detail Allergies √ √ √ √

Audit X √ √ √

Questions on Compliance X X X √

173

5.4.2.3 Audit

Hospitals B, C and D, had previously conducted audit to assess their medicines reconciliation. The audits measured how soon post-admission the pharmacists were verifying the patient’s medication history in line with the WHO1 and the NICE & NPSA6 guidance of 24 hours. In Hospitals C and D this was re-audited routinely, with Hospital B more ad-hoc.

5.4.2.4 Overview of medicines reconciliation process

Different process approaches were found in the four hospitals, with some similarities which occurred within all hospitals (see figure 5.4).

Patients were usually admitted to hospital through one of 3 routes; via A&E, direct to an admissions unit (medical or surgical) or in the case of elective admissions via a pre-admissions assessment unit. Whichever route the patient took, a medication history was conducted by the junior doctor during the initial assessment or clerking or, occasionally by a pharmacist. None of the hospitals used nurses to take medication histories at acute admission to medical wards.

Following the medication history taking, medication was prescribed, predominantly by junior medical staff. Occasionally in Hospital D pharmacists would add medication which had been omitted if they were a qualified prescriber. All hospitals had pharmacist non-medical prescribers, although only hospital D had non-medical prescribers on the admissions wards. Throughout the hospital stay the patient’s medication was discussed on the ward round and medication amended and changed as deemed appropriate. Some patients had their prescription verified by a member of pharmacy staff at some point during their hospital stay. Patients were either discharged or transferred to a ‘downstream’ ward for example from A&E to the admissions ward or alternatively from an admissions ward to a longer stay ward or home.

The decision to discharge a patient was usually made during the ward round. The patient’s discharge medication was prescribed on a discharge summary sheet, transcribed from the patient’s IMC. A pharmacist checked the transcribing (where medication was being supplied), relevant medication was supplied to the patient (usually 7-14 days’ supply) and the patient was

174 discharged. Where no medication was supplied the discharge note was not checked by a pharmacist. It was unclear how often this was the case.

The patient moved to their own home, a care home or intermediate care, with their discharge information being passed to a number of different people, always the GP, and in some cases the district nurses, community pharmacist, social services, care home staff, or intermediate care.

This process was followed within each hospital unless specified otherwise.

175

Figure 5.4

176

5.4.2.5 Taking a medication history

The initial medication history was taken by the doctor during clerking. The medication history was then verified at a later point by a pharmacist or pharmacy technician. There was no system for allocating whether a medication history was taken by the pharmacist or pharmacy technician in any of the hospitals, with often the pharmacist starting at the top of the bed list and the pharmacy technician at the bottom and meeting in the middle or vice versa. The length of time it took to obtain a medication history varied from patient to patient. It was difficult to tell precisely how long the doctors took to take a medication history as it was often embedded in the clerking and medical history taking process. Pharmacy staff took longer than the medical staff to take a medication history due to a more thorough approach more suited to eliciting a medication history (average 15-20 minutes compared with 5-10 minutes). The medication history was documented in three out of the four hospitals on specific medicines reconciliation forms designed to standardise the process (used jointly by doctors and pharmacy staff). None of the hospitals held electronic records, had e-prescribing or stored information about medicines reconciliation electronically.

Sources

The number of sources used to verify a medication history ranged from one to four. The pharmacy staff used a wider variety of sources than medical staff who often only used the patient and a GP record or previous discharge summary in all hospitals observed. In all hospitals, it was often apparent from the IMC that the doctor had used something to inform them of the medication but the source was unclear. No HCPs asked about contraception. Only one doctor was seen asking about OTC medication and CAMs and no pharmacists. No HCPs routinely asked about compliance, if, though, there was a concern about compliance a discussion sometimes took place. Similarly, none of the HCPs asked whether the patient had any difficulties taking their medication or whether they needed any help with their medication.

Information was received from the GP in a variety of ways. All hospitals received information by fax or occasionally from the patient in the form of an admissions letter. Often, if the patient was deemed reliable then further sources 177 were not utilised. Hospitals differed in the consistency with which information received from the GP (either direct from the EHR or faxed) was filed in the notes. This was not mentioned in the policies. In Hospitals A and C faxes often had to be requested twice because the either had not been sent/received the first time or they had been misplaced.

In hospitals where e-discharge was used this was used more frequently as a source of information, certainly by the pharmacy staff, and on some occasions the doctors. One pharmacist ensured that they annotated the e-discharge with information on whether the medication had been reconciled so that, if it was used as a source in future, it was clear. There were only seven of 158 occasions when a primary care provider other than the GP was used as a source. These were the community pharmacist, district nurse and practice pharmacist. Community pharmacists were rarely contacted and where they were, it was predominantly for MDS patients. The name and address of the patients’ usual community pharmacy was only documented routinely in Hospital D. This was not prompted by existing systems, it relied on the pharmacy staff remembering to annotate the front of the prescription chart. In the other hospitals it was only recorded where the patient used an MDS. Sources were not always double checked with the patient and often GP lists were taken as being accurate.

Two of the hospitals (A and B) had full access to the patient’s primary care electronic health record which included access to the medication they were prescribed within primary care. Hospital D had access to the patient’s Summary Care Record (SCR). Where the hospital could not access the EHR, information was usually requested from the GP via the telephone, with some practices requesting a supporting fax request. This may have been because the hospital did not have access (in the case of Hospital C), the patient had opted out of SCR or the practice did not have SCR. Very occasionally a patient was admitted to hospital A or B from a different health economy where the EHR could not be accessed. In these instances the general medical practice was contacted.

178

In A&E the ambulance record was routinely looked at, however, once the patient was on the ward it was only observed being referred to once by the consultant during the ward round.

Patients were not consistently encouraged to bring their own medication in following admission in any of the four hospitals. All hospitals had lockers at the side of each patient bed to ensure the secure storage of the medication was in close proximity to the patient. This facilitated medication history taking when the patient’s medication was used as a source as the HCP did not have to go elsewhere to gather the information and be able to use it with the patient to confirm the history. During medication history taking there was variation in whether the HCP asked whether the patient had either brought a list of medication or their actual medication into hospital. Occasionally the patient would offer the information without being asked. Where present, family members and carers would sometimes assist with medication history-taking, intermittently contributing or interrupting the conversation.

Where there were discrepancies between how the medication was prescribed in primary care and how the patient specified that they took their medication, this was not documented nor communicated back to primary care in any cases observed. Nor was non-adherence ever explored with the patient. On one occasion a history of non-adherence was identified by the pharmacist in a renal patient who had just undergone transplant. This had the potential to lead to rejection of the kidney. It was documented in the notes to not discharge the patient until this had been discussed. The following week the pharmacist reported that the patient was discharged over the weekend with no discussion of her non-adherence.

On no occasion was a specific interpreter used with non-English speaking patients, however, healthcare staff interpreted on three occasions and family members on four occasions.

This part of the SCOS (Taking a Medication History) also included identification of certain communication skills which will be considered in section 5.4.3.6.

179

5.4.2.6 Prescribing medication

Within this study only doctors were observed prescribing on IMCs in Hospitals A-C, even though they all employed non-medical prescribers. Within Hospital D non-medical prescribing was observed, only to add medication that was omitted rather than the whole medication chart. The pre-admission medication was initially prescribed by the junior doctors (Foundation Year 1 [FY1] and Foundation Year 2 [FY2]) without consideration as to whether the medication was clinically necessary, e.g. seasonal medication such as antihistamines, or continuing need and on occasion without knowing why the medication was prescribed. These were identified later by the consultant who stopped the medication or were identified by the pharmacist during a clinical check. Decisions to stop, start or withhold medication were left to more senior members of the medical team, predominantly the consultant but on occasion, the registrar. However, initially all medicines were prescribed and the patient may have had a dose (of the ultimately changed/stopped medication) administered. Involvement of more senior medical staff depended on when the patient was admitted relative to the post-take ward round, which happened twice a day on admissions wards in Hospitals B-D. In Hospital A it only happened once a day on the elderly admission ward but twice a day on the general medical admission’s ward.

Prescribing decisions were not always documented in the notes and often lacked detail. Medication stopped at admission was not always documented, nor was the reason for it being stopped. Sometimes it was difficult to tell whether medication had been omitted, stopped or prescribed intentionally. Some medical staff would prescribe it then withhold it by annotating the chart, or cross through the prescribed item to demonstrate the patient was taking it prior to admission. Others would annotate the notes or the IMC, though, this was on the minority of occasions. None of the hospitals had pharmaceutical care plans in place for their patients. Sometimes, even where it was clear that medication had been stopped, the reason why was not clear to the HCP.

On several occasions where the prescriber was unsure about a dose or regimen they used relevant information sources such as the (British National Formulary) BNF or local hospital guidelines e.g. for antibiotic prescribing. The

180 duration of treatment was usually documented for antibiotics. The form or type of device was often omitted from the prescription, especially with inhaled medicines.

5.4.2.7 Verification of the medication history by pharmacy staff

Pharmacy staff attended the wards to undertake a variety of roles. For pharmacists, this included, clinical checking of the prescription charts, ordering medication, answering clinical queries and verifying the medication history. For pharmacy technicians, this included ordering medication, answering technical queries and also verifying the medication history.

Pharmacy staff verified the medication histories of patients following admission. In all hospitals this stage was referred to as medicines reconciliation, including within policies and procedures. Hospital D’s policy broke it down further into ‘full’ and ‘basic’ reconciliation to denote the verification undertaken by the pharmacist and the medication history taken by the doctor respectively. All hospitals had a full complement of pharmacy staff on the wards during the working week 9.00-5.30 but only Hospital D had pharmacists working on the wards at the weekend, even so this was still a diminished service. This meant that only a proportion of patients were seen by pharmacy staff, either during or outside standard working hours. Some patients were not reviewed at all, for others it was over four days from admission and others at discharge. Pharmacy staff annotated the front of the chart to indicate that a medication history had been verified. For three of the four hospitals (B-D) a complete medication history could also be found in the notes. Discrepancies identified during verification of the medication history were predominantly resolved with the doctor at the time of identification. Occasionally messages were left with the nurse where the doctor was not available. Some discrepancies were resolved with the patient or family member/ carer, the care home or the GP.

Queries were communicated in a variety of ways. These included writing in the patient notes, writing on the front of the IMC or speaking to the relevant HCP. Hospital C had a green communication slip which was attached to the front of the medication chart using sticky tape which notified the doctor of any queries or discrepancies. All HCPs including the nurses were aware of the form and

181 were alerted to it by its colour and prominence on the front of the chart. It would be noticed during administration rounds by the nurses who would then notify the doctors to resolve the query. During verification of the medication history, a clinical check was also often carried out (see also Section 5.4.3.7 Communication).

5.4.2.8 Transfer of patient care

When a patient was transferred to another ward within the hospital or to intermediate care, the medication chart was transferred with the patient for continuity in all four hospitals. The nurse also rang the destination to hand over information which they deemed relevant. This did not always involve information about medication.

5.4.2.9 Ward round

This was not included in the SCOS as there were no guidelines or empirical literature concerning the impact the ward round had on medicines reconciliation at transition points. However, because the literature shows that changes to medication in hospital are common,532,533 it was important to observe ward rounds. All ward rounds were led by the consultant except for two (out of 32); one of which was led by the registrar and the other by the FY1 and FY2 doctors. No decisions about treatment change and investigation could be done without the consultant’s approval, unless they were to stabilise the patient in an acute situation.

Stop, starting, changing and withholding medication

One hundred and thirty eight out of 193 patients had alterations made to their medication during the ward round or clerking; 63 had medication stopped, 34 started, 28 changed and 13 withheld. Medication was stopped started changed or temporarily withheld for a variety of reasons (see table 5.9). This usually occurred during the consultant ward round although it occasionally occurred during clerking by the junior medical staff where the patient was on medication which may affect their acute condition. Sometimes medication was stopped then restarted within the same admission and vice versa.

182

Table 5.9 Reasons for stopping, starting, changing or withholding medication

Stopping Medication Starting Medication Change in Medication Withholding Medication

(n=48) (n=36) (n= 26) (n=28)

e.g. increase, decrease dose, formulation change

Reason n Reason n Reason n Reason n Side effects e.g. 18 Treatment e.g. antibiotics, 17 Inadequate symptom control e.g. pain, blood 7 Awaiting results of an 10 Maelena/bleeding, Blood Pressure (BP) glucose levels investigation or blood test constipation, fall medication Acute course e.g. antibiotics, 7 Prophylaxis e.g. aspirin, 12 Dose not therapeutic 6 Temporary change in 8 steroids, low molecular weight statin, PPI, Calcium state of patient e.g. heparins (LMWH). supplements hypotension, hypoglycaemia, diarrhoea Prescribed in error at 4 Treatment of side effects 4 IV to oral 4 Overdose 6 admission Liver/renal impairment 3 Anticipatory medicines e.g. 3 Dysphagia 3 Temporary interaction 3 morphine Patient is palliative 3 Reduction due to renal function 2 Clostridium Difficile 1 infection Patient no longer taking/ 2 Side effects 1 refuses to take Toxicity 2 Interaction 1 Risk outweighs benefit e.g. 2 Microbiology results 1 warfarin Dysphagia 2 Toxicity/overdose 1 Interaction 2 Not working 1 Test/ investigation results 1 Improvement in symptoms 1 e.g. pain

183

Medication was crossed off the patient’s chart (stopped) for a variety of reasons. Most often to stabilise a patient in an acute situation e.g. hypoglycaemics or antihypertensives, or where the risk of continuing medication was considered to outweigh the benefit e.g. warfarin. A medicine would also be crossed off at the end of a short course of treatment e.g. antibiotics or if the patient needed palliative care and unnecessary medication was stopped. As well as stopping medication to stabilise a patient some medicines were also started. Often these medicines were for acute use and were stopped before discharge e.g. fluids or converted to oral/maintenance doses e.g. antibiotics (see table 5.9).

5.4.2.10 Medication administration

Medication administration was not included in the SCOS as there were no guidelines or empirical literature which looked at the interventions which could be made during medicines administration associated with medicines reconciliation or the nurses role in medicines reconciliation. During medication administration rounds seven patients highlighted to the nursing staff medication which had been omitted from the prescription chart or alternatively medication which had been stopped and the medication no longer taken. Two examples were seen on one of the elderly wards:

‘Nurse: Do you want some olive oil in your ears?

Patient: No I don’t use that anymore.’

Nurse 7, Elderly Ward, Hospital B

‘Nurse: You are written up for tramadol

Patient: I usually take that at home.

Nurse: Oh do you?

Patient: Yes.

Nurse: Did you bring it in with you? 184

Patient: Yes.

Nurse: Is it in your locker?

Patient: Yes. I also usually take quinine at night but I didn't get any last night.’

The nurse gets the medication out of the locker but is unable to administer it to the patient as they are prescribed tramadol 50-100mg when required whereas they usually take tramadol 200mg M/R which is what is in the POD locker.

Nurse 7, Elderly Ward, Hospital B

One nurse identified three discrepancies which were not communicated to the doctor for resolution following the administration round. These included the omission of a blood pressure medication highlighted to the nurse by the patient, the incorrect antipsychotic medication dose and the omission of eye drops.

5.4.2.11 Discharge

All hospitals produced a discharge summary to communicate medication lists to the next care provider. They were all completed by doctors, though pharmacists added to them where information was deemed incomplete. All observed wards wrote the discharge summary on the day of discharge except two wards, one in Hospital A and one in Hospital B. They were both elderly wards who were often waiting for patients to be assessed for social needs. Discharge summaries were written by a doctor on all wards as they acted as a prescription for the provision of discharge medication. In three Hospitals B-D, these were electronic making them legible. The legibility of the handwritten discharge summaries (in Hospital A) was variable, especially on the 4th carbon copy which was usually given to the patient.

All discharge summaries had space to record medication information (name, form, strength and directions), allergies, name of person completing the form and reason for admission. Adherence and any special formulation information was not recorded. All summaries also had space to provide information about medication that had been stopped, started or changed during the stay, whether 185 paper or electronic. This, however, was not always completed. Only one e- discharge system (Hospital C) used forcing functions to ensure it was completed for each medication the patient was taking prior to discharge.

On most occasions, it was the GP who received discharge information from all four hospitals. Infrequently the community pharmacy would receive a faxed copy if the patient was on a MDS. The district nurses on two occasions received verbal information via the telephone from ward nurses about the requirements of their patients post discharge; the administration of insulin and Vitamin B12.

The pharmacist’s role at discharge was to check the transcribing of the medication from the chart on to the discharge summary, clinically check the discharge summary and to verify the medication history in cases where it had not been done. If the patient did not need any medication supplying at discharge, pharmacy staff would not see the discharge summary before it was sent to primary care. It was unclear whether this led to error, but given the amount of discrepancies being identified, it is likely.

‘Blanket orders’ e.g. not itemising each item on a discharge summary and writing ‘No changes’ or ‘as previous’, were never used. No patients on ‘special’g medications or non-standard preparations were observed. If a medication prescribed at discharge was only intended for a specific time period e.g. antibiotics or steroids, the duration was more likely to be completed on the electronic discharge as it was prompted by the forcing functions in the electronic discharge software used in Hospitals B-D. In contrast, the handwritten discharge information relied on the prescriber remembering to add the information.

All hospitals ensured that the patient had at least one week’s supply of medication at discharge. The HCPs reported that discharge summaries were transmitted electronically in Hospitals B-D and via the post in Hospital A, although transmission was not observed during this study. There was no

g 'Specials' are special-order unlicensed medicines made to meet the needs of an individual patient.

186 specified timescale for sending and it was unclear how soon after discharge they were received in primary care.

5.4.2.12 Counselling patients

Patients were not advised during any of the observations to carry an accurate list of medication with them, however, during clerking and verification of the medication history the patient was often asked if they had brought the list with them which might serve as a prompt to the patient in future, although this was not standard practice. Discharge counselling was rarely observed because medication was received by the ward from pharmacy in an ad-hoc fashion making it difficult to observe the nurses providing the patients with their medication. It was only observed on four occasions. Nevertheless, the researcher asked the nurse on each ward to talk through the process. Patients were predominantly counselled by nurses at discharge in the form of a simple reiteration of the dose instructions documented on the discharge summary rather than specific counselling points relevant to the medication. Rarely, the consultant was observed explaining the medication changes to the patient on the ward round but this varied from consultant to consultant and consultant to patient. If medication was more complex, counselling would be conducted by the pharmacist e.g. warfarin or a specialist nurse e.g. insulin or inhalers.

All patients were given a copy of the discharge summary at the point of discharge. No other patient information was seen e.g. leaflets, aide-memoire etc. except on the renal ward in Hospital D, (post-transplant) where a comprehensive, personalised booklet on anti-rejection medication was provided. All patients newly initiated on warfarin were provided with relevant information including the NPSA yellow book.

No patients were observed being advised to speak to their community pharmacist regarding any medication changes although a pilot study was about to commence on one admission ward in Hospital A to provide the community pharmacist with a slip which indicated which medication had been changed during inpatient stay. Sometimes staff would ensure relevant family members were there before relaying information, especially in the case of warfarin. It was

187 not clear whether this was related to the specific medication type or the patient or some other reason.

Patients were rarely asked during any observations of any practical problems taking their medication. Patients were asked to demonstrate whether they could use their medication on two occasions – these were both related to inhaler technique in Hospital A.

5.4.3 Analysis using Yorkshire Contributory Factors Framework (YCFF)

The findings in the subsequent sections are presented according to each of the factors described in the YCFF. Each section discusses the individual contributory factors which may lead to medicines reconciliation-related error and or discrepancies followed by potential defences which may prevent such errors.

As described in section 5.3.8, 10% of the codes were categorised by two co- raters and the kappa coefficient calculated to determine inter-rater reliability. The kappa coefficient for agreement as to whether the code was a defence or contributory factor was high (0.88, P=0.00). Using the YCFF, there was some level of agreement (P < 0.05) for the following categories: communication systems, equipment and supplies, individual factors, patient factors, physical environment, policy and procedures, staff workload, and training and education. Agreement was strongest for patient factors, physical environment and staff workload (see table 5.10). (Some of these were limited by the number of codes falling into each category e.g. for equipment and supplies there was only one code categorised and all raters agreed.)

A kappa value (K) was also calculated to examine agreement between individual co-raters to determine any significant differences i.e. co-rater 1 and 2, co-rater 2 and 3 co-rater 1 and 3. The two pharmacist co-raters had a significant kappa (K=0.48) for policy and procedures (Which led to a significant overall kappa value for that contributory factor but it was not significant between the either of the pharmacists and the psychologist). This would seem reasonable given the potential unfamiliarity of the psychologist with pharmacy policy and procedure which came out in the discussions post-categorisation.

188

Table 5.10 Inter-rater reliability of categorisation using YCFF categories Contributory Factor Kappa P value Coefficient (K) Active Failures 0.18 0.01 Communication systems 0.40 <0.005 Equipment and supplies 1.00 <0.005 External policy -0.03 0.59 Design of equipment and supplies -0.02 0.59 Individual factors 0.38 <0.005 Lines of responsibility -0.03 0.66 Management of staff and staffing levels -0.03 0.59 Patient factors 0.65 <0.005 Physical environment 0.58 <0.005 Policy and procedures 0.13 0.02 Safety culture -0.04 0.68 Scheduling and bed management Too few variables specified Staff workload 0.47 <0.005 Supervision and leadership -0.03 0.63 Support from central functions -0.03 0.59 Task characteristics 0.07 0.15 Team factors 0.08 0.14 Training and education 0.25 <0.005

Conversely for the team factors (which did not produce a significant overall K value) there was more agreement between the pharmacist researcher and the psychologist (K=0.48) (both with more experience of human factors) compared with the pharmacist.

189

5.4.3.1 Active failures – Mistakes, lapses and violations

YCFF Definition: Any failure in performance or behaviour e.g. error, mistake, violation of the person at the ‘sharp-end’ (the healthcare professional)

Contributory factors

It was difficult to identify slips, mistakes and lapses through observation as the HCP’s thought processes cannot be observed and thus were unknown. It is probable that where HCPs have read the organisation’s policies and procedures and do not follow them that this is likely to be an active failure i.e. the HCP has either accidentally overlooked part of the task, or has purposefully not carried out that part (a violation), especially in the hospitals where they sign to say they have read the policies.

Some actions were definite violations, for example, using another person’s log on details to access the e-discharge system. Staff knowingly gave their logon details to facilitate access when other staff had been logged out. Other violations occurred because tasks were openly deemed too onerous:

‘Work out which insulin she's on and then reduce the dose – Don’t bother clerking. Just find out what drugs she’s on and then reduce them. Get X [FY1] to check with doctor [GP] after 9.00 am the patient’s dose. I don’t want the doctor to see them just confirm dose.’

Consultant 8, Elderly Admissions Unit, Hospital A

‘Often the nursing homes don’t send the patients with anything as they have to fill out a lot of paperwork and the ambulance may be in a rush to take them.’

Pharmacy Technician 1, MAU, Hospital B

Defences

There were no observations which were categorised as defences for active failures.

190

5.4.3.2 Situational factors a) Individual factors

YCFF Definition: Characteristics of the person delivering care that may contribute in some way to active failures. Examples of such factors include inexperience, stress, personality, attitudes.

Contributory factors

Different HCPs approached tasks associated with medicines reconciliation differently. Doctors always read the notes first, whereas some pharmacy staff would speak to the patient first and others would read through the notes. It was not always clear why the HCP chose to do things in a specific way but seemed to be a result of knowledge or experience that they had gained previously.

Whilst reading the notes some looked at the medicines first whilst others look at the patient’s context. One pharmacist explained their rationale:

‘If you look at the medicines first- sometimes you get a false picture. I first read through the notes to get a feel for what’s going on.’

Pharmacist 12, Elderly Ward, Hospital C

Nevertheless, all HCPs appeared to work systematically through the tasks they were conducting. Some were more conscientious than others for example ticking an item as it was cross checked. The HCPs would also double check and clarify information where needed to gain a fuller picture or to confirm their understanding. This happened irrespective of the task being undertaken or the HCP conducting the task. For example, some doctors would double check that they had transcribed the IMC correctly once complete and cross the items off as they were re-prescribed. Or they would count the number of items prescribed on the discharge prescription and check that it matched the number on the IMC. Some would use the patient in the checking process:

191

‘I will usually show the patient the PODs and go through each one and ask how they take it or whether they know how to take it.’

Pharmacy Technician 1, MAU, Hospital B

The way in which information was clarified varied depending on what required confirmation. Sometimes it was a discrepancy between two sources, in which case this was clarified with the patient or the owner of the source, most commonly the general medical practice. On two occasions, information was verified with other hospitals, following recent inpatient stays. Other specialties were sometimes contacted for information or a second opinion, often where the patient was already under the care of another specialty.

Most staff had an opinion on what was the ‘best’ way to manage reconciliation and pointed out what to look out for when taking a medication history, some of these beliefs were incorrect or not evidence based, for example:

‘I like to use more than one source. I usually go with the most recent source as the accurate one’

Pharmacy Technician 2, MAU, Hospital B

‘The GP history is a better source than dossett box’

Pharmacy Technician 5, MAU, Hospital C

‘I usually use two sources unless it is a MAR chart. The MAR chart indicates what the patient has actually been taking or is administered’

Pharmacist 7, Elderly Ward, Hospital B

HCPs also passed comment on the ability of other groups to conduct medication histories, stating their inferiority and lack of accuracy. This was usually the pharmacist but sometimes the doctors:

192

‘Some techs [technicians] pick up more than others, like this [medication not being ordered], I’ll have to order it. It takes up your time.’

Pharmacist 17, Elderly Ward, Hospital D

‘The doctors are not very good at completing where they have collected the information from. It is generally worse at night’

Pharmacist 1, MAU, Hospital B

‘The ambulance service can’t spell any drugs’

Doctor 16, Specialist Registrar, A&E

Staff were not always aware of the limitations of their knowledge or competence or the potential for this to lead to harm or may not have wanted to admit that they were unsure about medication. On one occasion the doctor prescribed medication without knowing its indication:

‘At the start of the ward round the junior doctor is interrupted by a nurse who has been administering medication.

Nurse 1: The patient is adamant that they want to take these can you prescribe them. Cetirizine 10mg [the nurse shows the junior doctor the packet] and Terazosin 2mg - she's on 2 of those [shows the junior doctor the second packet].

The doctor starts to prescribe them. The consultant questions the junior doctor asking if she knows what they are? She doesn't. The consultant gives the junior doctor a disapproving look to indicate that she should know what they are before prescribing them. The consultant asks the nurse to check that the patient’s BP [blood pressure] is ok before it is administered’

Field Notes, MAU, Hospital B

193

Other staff knew their limitations, for example some doctors chose to write notes during clerking so they did not forget details.

On 14 occasions the HCP made an assumption about the patient or their medication which may or may not have been correct. This was often about the number of medicines a patient was on or the ability of the patient to be able to recall their medication history based on age.

‘She’s only young so there is no reason she won’t be able to tell me.’

Pharmacy Technician 21, MAU, Hospital D

‘This means I'm probably not going to get much from the patient’. [The pharmacist pointed to the patient’s age – 88]

Pharmacist 7, Elderly Ward Hospital B

One pharmacy technician indicates how she uses assumptions to determine whether she has captured an accurate medication history:

‘I look at their age or whether they are pregnant and therefore assume that the patient isn't on a lot of meds. A 94 year old just on Simvastatin rings alarm bells.’

Pharmacy Technician 5, MAU, Hospital C

Other assumptions were made based on what had been documented by previous departments or staff (see also communication section 5.4.3.6).

‘You can see from this one that they are on lots of varying medication therefore it is likely to be a clear history but no sources have been documented’

Pharmacist 1, MAU, Hospital B

194

‘It assures me if the ambulance man has written thiamine and vitamin B and Senna that a full medication history has been taken’

D16 SP Registrar, A & E, Hospital A

On one occasion the HCP told the researcher that there was no point in observing clerking as A&E had recorded that the patient was not taking anything. The researcher explained that they would like to observe and continued to observe. A&E had omitted four medications including one for hypertension and warfarin. The HCP stated:

‘I suppose that illustrates your point as she’s on more medication. I never trust what A&E write.’

Doctor 51, MAU, Hospital D

Defences

Many of the HCPs observed had developed their own ways of working, based on their knowledge and previous experience which they believed helped them to elicit an accurate medication history.

‘It’s quite useful to look at previous discharge notes to get an idea of the patient’s past history. It’s also useful to tell you what medication people are on as often when patients come in they don’t bring in their medication and with the elderly they can be confused.’

Doctor 1, FY1,MAU, Hospital B

‘I always ask about herbal and OTC medicines. I went to X university and we had a lecture from the X unit and they stressed about the interactions. Ever since then I always check.’

Doctor 40, GP Registrar, Elderly Medical Unit, Hospital D

195

‘The way I work is I get all the information first before I speak to the patient. Different people do it differently but that’s the way I do it’.

Pharmacy Technician 21, MAU, Hospital D

Doctors would try to prescribe medication as soon as possible after clerking before the medicines ‘got lost.’ Two doctors in A&E made sure they gave the patient’s own medication back to the patient so that it got transferred safely to the ward. This was similar for pharmacy staff.

‘I try to write it when the dossett [MDS] is seen, as now it’s gone missing’

Pharmacy Technician 7, Elderly Admissions Unit, Hospital D

‘I ensure medicines are put back in the patient’s bag because sometimes they get left and the nurses have to pod them up to the ward’

Doctor 15, A&E, GP Registrar, Hospital A

‘Sometimes tablets get left on MAU and we have to chase them.’

Nurse 7, Elderly Ward, Hospital B

Observation of pharmacy staff and pharmacists’ responses during observation showed how they act as a ‘safety net’ (or defence) in many instances. This relied on the pharmacist having knowledge of individual policies and procedures plus working systematically and not missing any cues or vital information. Some pharmacy staff described getting a ‘funny feeling’ that something was inaccurate.

‘I just get an inkling that something is not right, for example, if the patient has co-morbidities listed but are not on any medication.’

Pharmacist 8, Elderly Ward, Hospital A

196

Some HCPs were more observant than others in noticing the patients’ surroundings. This led to additional useful information being elicited, for example that the patient used an inhaler or that the patient needed their hearing aid to communicate.

The HCPs were not deviating from any of the hospitals’ policies and procedures despite their differences in practice. This in the main was due to lack of detail in the policies and the need within the policies to allow for interpretation and clinical judgement.

b) Team factors

YCFF Definition: Any factor related to the working of different professionals within a group which they may be able to change to improve patient safety

Contributory factors

No contributory factors relating to team factors were observed.

Defences

There was inter and intra-hospital variation in systems which had been implemented to improve working and make things safer between wards, teams and disciplines. For example, the handover sheet for each ward varied depending on the information they found useful within their specialty. Different disciplines and teams had also developed their own methods of communicating, for example some wards had a ‘master jobs list’ for their doctors contributed to by all the doctors. Even then, ‘the master jobs list’ this would differ between wards with some annotating the ward patient handover sheet, some the bed board and others using a separate list. Some wards had developed aide- memoires to help them with tasks:

‘It’s a balance between providing enough information and not enough information. Some wards have a list up of what you need to include and what not to include [on the discharge summary].’

Doctor 3, FY1, MAU, Hospital B

197 c) Patient factors

YCFF Definition: Those features of the patient that make caring for them more difficult and therefore prone to error. These might include abnormal physiology, language difficulties, personality characteristics (e.g. aggressive attitude).

Contributory factors

Several patient factors impacted on the patient’s ability to be involved in a consultation with a HCP, which arguably contributed to the lack of contemporaneous information at reconciliation. These included hearing impairment (7 patients), visual impairment (4 patients), communication difficulties e.g. language difficulties (8 patients), the patient’s current condition (2 patients) and treatment and their capacity e.g. dementia diagnosis (14 patients). On one occasion on one of the elderly wards out of 15 beds only three patients were able to fully communicate. This occasionally frustrated the HCPs:

‘You should have observed the last one but there would have been no point as she couldn’t speak and I couldn’t take a history!’

Doctor 40, Specialist Registrar, Elderly Medical Unit, Hospital D

The patient might be asleep (7 patients) or unavailable (8 patients), for example they were away from the ward perhaps for an operation or investigation. When the patient was asleep the doctors would rouse them for a discussion, whereas, the pharmacy staff, in the majority of cases, would leave the patient to sleep and return later. In some cases the member of pharmacy staff forgot to return to the patient or ran out of time. It could be argued that the patient should be woken. Irrespective, it shows a difference in professional practice. Sometimes it was the patient’s condition which made them drowsy or confused, especially where they were taking sedating medication, had severe acute infection, were palliative or had previously been diagnosed with a condition such as dementia. Where the patient was away from the ward the patient notes were taken with the patient including the medication chart. In these cases, it was impossible for pharmacy staff to verify a medication history. On one occasion the patient was

198 unavailable because they had already gone home and a family member was calling back for the medication later. This highlights the risks with patients with communication difficulties. Hospitalisation itself may contribute to a patient’s state of mind, highlighted by one pharmacist:

‘These [renal] patients are used to being in a healthcare setting, whereas for some others it is the first time they have been in a hospital and we forget because we are so used to being here, but it can be quite daunting for patients.’

Pharmacist 19, Renal Ward, Hospital D

One patient seemed frightened as the doctors approached her bed stating ‘I am Gujarati speaking’ in broken English. When the patient was upset or agitated on occasions it distracted the HCPs from their usual work whilst they took time out with the patient. It also meant that the patient was distracted which potentially could affect the way they interacted with the HCP making it important to address the patients ideas, concerns and expectations to ensure a successful consultation.617

Patients sometimes had different concerns or priorities to the HCPs:

‘Following a long conversation with a patient who had just been informed by the consultant that she had had a heart attack, the consultant asked whether she had any questions. The patient replied ‘Yes, will I get breakfast?.’

Field Notes, MAU, Hospital D

Patients’ knowledge of their medication varied. Some patients could remember all their medication with strengths. Other patients were vaguer, remembering some, with a few not remembering any at all, or being able to describe them but not remember their name. Some patients relied on a carer to administer their medication, therefore, had no recollection of the medication they were taking. Sometimes the patient would communicate irrelevant, superfluous information which the HCP had to filter to make sense of what the patient was trying to

199 communicate. Occasionally the patient would not understand the questions being asked:

Doctor: Are you eating as normal?

Patient: I haven't been for three days

Doctor: Are you drinking as normal?

Patient: Yes four cans a day

Doctor 8, FY2, A&E, Hospital A

Defences

The patient, relative or carer often acted as a defence when medicine-related information could not be ascertained from the patient during medication history taking, ward rounds and medication administration (sometimes invited to do so by the HCP, in other cases chipping in), e.g. clarifying doses of medication, highlighting any medication that had been recently altered and informing staff of delayed or omitted medication. A relative or carer was observed to add value in 15 consultations. They added information which could not be drawn from another source, for example, the way in which the patient took their medication.

Some patients and/or their carers raised concern during various tasks, for example, during the administration of medication round, which could prevent errors occurring where acted on by HCPs. Others took active participation in decisions about their care. Five patients informed the nurse during the administration round that they usually took a medication that they had not been given. These included blood pressure medication, pain relief, PPIs and antihistamines. One patient also informed a member of staff that they had also been unable to take their medication because they had no thickener to take the medication with. On other occasions the patient provided information that they no longer took a particular medication:

‘Patient 294 had two IMCs. Aspirin was prescribed on the second chart. The nurse almost missed seeing it as she had not seen the second chart. 200

‘She administered it in a separate beaker. As it wasn't in with the other medication the patient asked what it was. The nurse responded “It's your aspirin.” The patient replied “Oh I don't take that it's been stopped.” The nurse discarded it. It was unclear whether she brought this to anyone’s attention later on to be crossed off the medication chart.’

Field notes, Observation of Medicines Administration, MAU, Hospital C

Self-administration of medication (mainly insulin and inhalers) was promoted on one of the wards to maintain patient independence and to ensure that the patient was competent and understood their regimen when they were discharged. d) Task characteristics

YCFF Definition: Factors related to specific patient related tasks which may make individuals vulnerable to error

Contributory factors

The task of collecting a medication history for a patient at admission to hospital requires gathering data from a number of sources because there is more than one source of primary care medication and variance in the patients’ adherence to treatment. Obtaining an accurate medication history not only depends on the content of the source used but also the HCP’s interpretation of the information within the source. This individual interpretation could be a potential basis of error. One pharmacist gave an example of a patient who had iron sulphate supplied in primary care three months previously. Some HCPs may assume that the patient was no longer taking it; however, they had received four months’ supply. Another example provided was of a doctor not taking notice of the date when the summary care record (SCR) was created, therefore using a record which was 3 months out of date. Each HCP taking a medication history chooses the sources they wish to use. This may lead to relevant information being overlooked if the wrong source is chosen. The tasks also rely on accurate contemporary information.

201

Defences

There were no observations which were categorised as defences for task characteristics.

5.4.3.3 Local working conditions a) Lines of responsibility

YCFF Definition: Existence of clear lines of responsibility clarifying accountability of staff members and delineating the job role.

Contributory factors

It was clear who was responsible for taking the initial medication history (doctor) and who was responsible for verifying the medication history. What was less clear was who was responsible for documenting the alterations in medication on the ward round in the medical notes. Within all hospitals tasks were delegated from one member of staff to other members of staff. Most often from a more senior member of staff to a more junior member of staff e.g. the consultant delegating to the junior doctor to write in the notes or to write the discharge summary. It was unclear in some cases whether the delegation was appropriate and whether the staff being delegated to were competent to undertake the tasks devolved or were clear that they had been asked. In two cases, in Hospital D, the doctor asked the pharmacist to take a medication history rather than the junior doctor. There was sometimes a difference between the staff who line managed an individual HCP and those who supervised a HCP undertaking a specific task e.g. pre-registration pharmacy students.

Defences

There were no observations that were categorised as defences under lines of responsibility.

202 b) Staff workload

YCFF Definition: Level of activity and pressures on time during a shift

Contributory factors

Staff workload varied from day to day and it was difficult to predict whether the day was going to be busy with either admissions or discharges. The admission wards have a constant turnover of new patients. On one day of observation the ward had 23 new patients out of 30 beds. Several participants from different professions specified that it was easier if you were present on the ward the previous day as the patients were more familiar, raising the importance of the continuity of staff. None of the pharmacy services ran a full complement of staff over the weekend, therefore a higher workload was experienced in all hospitals studied on Monday morning (see also management of staff and staffing levels). The extent of medical cover at weekends was not determined.

Defences

Workload was managed centrally through rotas (see management of staff and staffing levels) and also locally on the ward depending on the number of new patients, staff availability and other tasks. Where there was more than one member of a discipline or team on the ward there was often a negotiation or delegation of which jobs each member of staff would undertake (see also Lines of Responsibility 5.4.3.3(a)). One pharmacist described how a previous hospital he worked in managed their workload and improved their target for completing medicines reconciliation within 24 hours in line with the NPSA & NICE guidance:

‘I worked at X [hospital] previously. Pharmacists provided comprehensive ward cover between 8.00am and 7.00pm. Sometimes pharmacists saw the patient before the doctor. Patients also had medication dispensed on the wards. It provided quicker turnaround of patients.’

Pharmacist 7, Elderly Ward, Hospital B

203

The principal strategy for dealing with workload observed was prioritisation. Consultants on the admissions wards prioritised new patients so they could be seen with the night staff who clerked them. Staff also prioritised jobs deemed as ‘urgent’. Discharges were usually given a high priority, still, occasionally other tasks needed to be prioritised instead, for example, acutely unwell patients. The pharmacists also prioritised new patients (not only on the admission ward). On the long stay wards there was more time to plan different tasks, for example, the renal team in Hospital D did not review medication every day because their patients usually stayed at least 72 hours if not longer:

‘We did a big ward round yesterday and reviewed the medication so we won’t be looking at the charts as much today.’

Doctor 47, Specialist Registrar, Renal Ward, Hospital D

In Hospitals B-D the pharmacy technicians carried bleeps (rather than just the pharmacists) so they could be contacted and work assigned or delegated rather than being left to consider their own workload. This also assisted prioritisation. c) Supervision and leadership

YCFF Definition: The availability and quality of direct and local supervision and leadership

Contributory factor

It was left to the doctors to decide when to go for lunch. Doctors would go for lunch breaks together, with no staggering and often without notifying the nursing staff. This left wards with no medical cover which impacted when pharmacy staff required a doctor to make amendments to an IMC. Within Hospital A there were pharmacy technicians allocated to some wards and not to others. The amount of time they spent on the ward was variable and they were still trying to establish the role they should undertake. This was potentially due to lack of supervision and leadership with no one person taking responsibility for the role of the pharmacy technicians on the wards.

204

Defence There were no observations that were categorised as defences for supervision and leadership within this study. d) Management of staff and staffing levels

YCFF Definition: The appropriate management and allocation of staff to ensure adequate skill mix and staffing levels for the volume of work.

Contributory factors

Pharmacists often had to cover other wards (other than their usual ward) due to holiday and sickness, or do ‘on-call’ taking time away from their regular ward commitments and diluting their input on the wards. Pharmacy staff in all hospitals complained about not being able to ‘do a good job’ when they were covering staff shortages. Pharmacists also experienced a high volume of patients to see on a Monday morning due to lack of weekend rota with no increased staffing levels to compensate. At times the junior doctor ward cover was also lacking. Junior doctor training was scheduled for all junior doctors at the same time leaving some wards with no medical cover. As mentioned previously, it was left to the doctors to decide when to go for lunch, which again left wards with no medical cover. This impacted when pharmacy staff required a doctor make amendments to an IMC and potentially also reflects a lack of supervision leadership and management (see section 5.4.3.3(c)).

Defences

There were both central and local approaches to the management of staff and staffing levels. In each hospital, each HCP had a rota to indicate where they were working on a daily basis. For the nurses this was done at ward level. For the doctors it depended on their grade; the junior doctors being organised centrally and the more senior doctors within their specialty. The pharmacy staff rotas were devised by the pharmacy department to ensure the service had some level of cover at all times. Each ward had at least one pharmacist allocated to it for some part of the day. The length of time the pharmacist spent on the ward varied within each hospital depending on the pharmacist’s role and the other duties they were expected to conduct during the day, for example, aseptic or dispensary slots. Hospitals B-D all had pharmacy technicians

205 routinely allocated to the wards. Within Hospital A there were pharmacy technicians allocated to some wards and not to others. The amount of time they spent on the ward was variable and they were still trying to establish the role they should undertake. At least one pharmacist was allocated to each admission ward in all hospitals; in the morning, this was sometimes doubled together with a pharmacy technician. In Hospital D in the afternoon a co- ordinating pharmacist carried a bleep for the four acute admissions wards to respond to any orders and processing of discharge summaries.

In Hospital D, there was a higher number of doctors allocated to the admissions wards and higher grades than the other hospitals.

Table 5.11 Staff allocation on medical admissions unit (MAU), Hospital D

MAU Bay 1 Bay 2 Bay 3

(8 beds) (8 beds) (8 beds)

Core Trainee 1 1 1 Doctors Foundation 1 1 1 Year Doctors Qualified Nurse 1 1 1 Healthcare 1 1 1 Assistant

They had six doctors, three core trainee doctors (CTs) who were senior and three foundation year doctors (FYs) who were junior. The doctors are allocated two per bay one junior and one senior (to 8 patients). Each bay also has one qualified nurse and one healthcare assistant (HCA). Only CTs and above clerked the patients, compared with other hospitals where lower grade staff were allowed to clerk patients (see table 5.11). Despite Hospital D having more allocated staff, many staff expressed how they still felt the pressure of workload:

‘Monday mornings are horrendously busy seeing new admissions’

Pharmacist 15, Elderly Admissions Unit, Hospital D

206

‘I can’t do four things at the same time, I have to prioritise’

Pharmacy Technician 8, MAU, Hospital D

‘At the weekend there is one doctor on call to 160 beds’

Doctor 45, FY2, Elderly Ward, Hospital D

One doctor, in Hospital C, described how different the UK staffing systems and skill mix were compared with US systems. He said:

‘We never took blood - the phlebs came down. There are no social issues to sort, someone else does that and Drs don't write discharges they just get paid to think, someone else scribes.’

Doctor FY1, Elderly Ward, Hospital C e) Equipment and supplies

YCFF Definition: Availability and functioning of equipment and supplies

Contributory factors

HCPs relied on IT to undertake a number of tasks for example computers for e- discharge, faxes for receiving information from general medical practices. Hospitals B-D had implemented electronic discharge systems all of which were different in the way they functioned. This variation itself, due to lack of familiarity, could potentially be a contributory factor to error as doctors rotate around the different hospitals within the region. Reliance on IT caused a hindrance if the systems failed as medication histories could not be completed. This included inability to log on, and being locked out of computer systems which forced staff to violate information governance policies. Staff were still getting used to the new ways of working which came with the introduction of IT:

‘In some ways it’s easier and in some ways it’s not – for example it’s easier to cross check with a paper version to make sure you have got all the medications, it is more difficult on the screen. Paper ones [discharge summaries] used to go missing though and you’d have to re-write them.’

Doctor 46, GP registrar. Elderly Ward, Hospital D 207

Variance in the completion of discharge summaries between HCPs was observed on both paper and electronic discharge in all hospitals, with some HCPs (especially the seven GP trainees) providing more information than others. The electronic discharge increased the amount of information the HCP provided, nevertheless, it did not appear to eliminate the variance in quantity and quality between prescribers, for example, some would add in test results which they felt were relevant and others would not.

There were mixed feelings and lack of understanding by some prescribers of the potential added safety of electronic discharge versus paper discharge. Some preferred the method, despite the increased length of time taken to complete the electronic discharge summary. Occasionally, discharge summaries were amended by hand which led to inaccuracies in the electronic system. Where the prescriber did not logout correctly from the computer system they were unable to re-access the prescribing software for 30 minutes, disrupting work flow and resulting in the sharing of logon details and passwords between prescribers.

Understanding the e-discharge software was sometimes confusing for staff. On one system, in Hospital D, the name of the person who originally wrote the prescription appeared against the medication, yet, when it was amended and changed by anyone including doctors and pharmacists their name then appeared which could lead to difficulties in interpreting accountability. One pharmacist described the complexity of amending e-discharge summaries once they have been written:

‘At the outset most pharmacists felt it took more time then it got easier although it still takes time. You can’t just get the doctor to add things on to it, you need to send It back. Doctors don’t always understand this especially when you have it open on your screen and they want to hop on and change it and you have to tell them to log out and log back in so its not on your log on. It is better for patients on a large number of items like 15 meds as the doctors don’t try to squash them all on to one sheet’

Pharmacist 17, Elderly Ward, Hospital D

208

Electronic discharge allows the hospital formulary to be embedded within the software to promote cost effective prescribing. However, incorporating only formulary medicines caused the prescribers to free type non-formulary medication details e.g. the name, dose, form and strength; this could lead to typing errors and the compounding of prescribing errors.

Defences

Three out of the four hospitals (B-D) employed electronic discharge summaries. A number of advantages and disadvantages were observed relating to electronic discharge summaries during the study. Electronic discharge software incorporated forcing functions i.e. mandatory fields, to ensure the prescriber completed all relevant fields on the discharge summary. This included amendments and changes to medication with the reasons for change, creating an advantage over paper discharge and improving patient safety through better communication with primary care. Safety is further improved as the electronic discharge summaries were always legible and could be completed at a computer terminal away from distractions i.e. not written during the ward round (although it was not always the case that the terminals were away from distractions). E-discharge software also allowed discharge summaries to be written in advance and saved, without the risk of multiple versions existing as it did with the paper-based system.

‘We try to start them and update them as we go along as it is like a “Journal”. It's more suitable for an elderly ward as patients tend to hang around longer. When a patient has been in a long time, trying to follow what has happened to them is difficult especially if they have stopped and started medication, for example it may say stop codeine but later on in the plan, it is re-started!’

Doctor 5, FY1, Elderly Ward, Hospital B

209

Hospital D had introduced a single access portal, various e-forms and databases to their intranet for easy access to information and to allow key performance indicators (KPIs) to be recorded. The hospital portal allowed all correspondence and test results for a specific patient to be accessed through one logon screen, including discharge summaries and outpatient letters. This was observed on numerous occasions being utilised by the whole MDT. One consultant described it as ‘revolutionary’.

HCPs felt that there was a greater potential for IT within the hospital especially electronic patient records and electronic prescribing.

‘Electronic prescribing would make transcribing not an issue’

Doctor 6, FY2, Elderly Ward Hospital B

‘Electronic notes/prescribing would be better especially for medicines as it could flash up interactions and tell you which cause hyperkalaemia, for example we had a patient last week who was on a number of medicines and we had to trawl through the BNF to determine if anything may have caused it. One of the items was Fortisip and we couldn't determine what was in it’

Doctor 5, FY1, Elderly Ward, Hospital B

5.4.3.4 Latent organisational factors a) Training and education

YCFF Definition: Access to correct, timely and appropriate training both specific (e.g. task related) and general (e.g. organisation related)

Contributory factors

All HCPs observed were registered with a professional body. The majority of HCPs observed had undertaken an undergraduate programme or equivalent in either nursing, pharmacy or medicine. The exception was the pharmacy

210 technicians who were either ‘grandparented’h or had undertaken a pharmacy technician qualification (they are still registered with a professional regulatory body). The extent to which medicines reconciliation was embedded into any HCPs programme of study was unknown. The medical and pharmacy students observed during the study were all part way through their programme of study. They seemed to have received very little educational preparation regarding the importance of medicines reconciliation and how to elicit a medication history within their studies. One student clarified:

Medical Student: You get taught nothing [about medicines history taking], except that you have to ask about medication.

Resea rcher: What do you get taught to say?

Medical Student: Are you on any medication?

We are told to structure our clerking as follows:

Presenting Complaint

History of Presenting Complaint

Past Medical History

Drug History

Family History

Social History

And that’s it!

Final Year Medical Student, Hospital C

h The ‘grandparenting’ arrangements for pharmacy technicians were put in place to ease the transition to regulation for pharmacy technicians. Up until midnight on the 30 June 2011 an individual with the relevant work experience and qualifications could apply to register as a pharmacy technician.

211

Post-qualification the HCPs gained education and training from a variety of places; some mandatory depending on their role e.g. GP trainee programme and some optional. All the pharmacists observed had undertaken or were undertaking a post-graduate diploma in clinical pharmacy or had equivalent knowledge which was a pre-requisite to them working in hospital.

None of the hospitals provided medicines reconciliation training to their pharmacists; Hospitals B-D did so for their pharmacy technicians. The training covered taught workshops on taking an accurate medication history and a competency assessment. Hospital A had no formal medicines reconciliation training. None of the hospitals provided medicines reconciliation training to their doctors e.g. how to take an accurate medication history or how to write a discharge summary. Training for this was learned ‘on-the-job’ by trial and error, or with help from more senior colleagues; often the FY2s teaching the FY1s. This led to intra-professional variation as their knowledge was learned from peers. Both bad and good habits were passed on. This was observed during the FY1s’ first day as a doctor on the wards and during supervision of medical students:

‘It is the FY1s first day, post qualification. They have not been shown how to use the e-discharge software and are navigating their way around it themselves. The FY1 has been informed how to fill in the discharge summary from the previous FY1 who has progressed to being an FY2. One FY1 has been told to copy test results on to the discharge summary the other hasn’t’

Field Notes, Observation of Writing Discharge Prescription, Elderly Ward, Hospital B

‘A fax is received from a general medical practice. The medical students document the medication the patient is taking from the fax. The junior ward doctor advises the medical student to check the last issued date to check it has been recently issued.’

Field notes, Observation of Clerking Patients, Elderly Admissions Unit, Hospital A

212

The complexities and multiple scenarios of medicines reconciliation make it difficult to teach and to learn. One pharmacy technician explained:

‘It is difficult when learning as you can't always follow a patient from start to finish. There are also lots of different scenarios which you can come across’

Pharmacy Technician 5, MAU, Hospital C

Training, whilst important, also interrupted the management of workload on the wards as the doctors had to leave the ward to attend their teaching, on one occasion, in Hospital A, this resulted in the consultant being left to do the ward round alone. One pharmacist felt there should be re-training or re-assessment of competency following return to work e.g. post maternity leave or after long period of illness, for all tasks including medicines reconciliation.

Defences

All hospitals provided an induction for their staff when they started. The extent to which information relating to medicines reconciliation was incorporated was limited. Most provide information about prescribing, however, nothing specific to taking an accurate medication history.

Pharmacy technicians were provided with training in Hospitals B-D. This was a pre-requisite for working on the wards. Within hospitals B and D there were two different levels of pharmacy technicians on the ward. Those which were able to conduct medication histories and counselling (NHS Band 5) and others which provided a POD and stock checking service (NHS Band 4).

‘Different techs [technicians] are taught to different levels. I have had training on drug history training, discharge process, medicines reconciliation and counselling’

Pharmacy Technician 1, MAU, Hospital B

In Hospitals A and C only NHS Band 5 pharmacy technicians went on to the ward; their role, to take medication histories (see table 5.12).

213

Table 5.12 Level of ward pharmacy technician training

NHS Band 4 NHS Band 5 Hospital Present Duties Present Duties on Ward on Ward A X -  Medication History Taking B  POD and Stock  Medication Check History Taking, Patient Counselling C X -  Medication History Taking D  POD and Stock  Medication Check History Taking, Patient Counselling

The consultants and more experienced doctors sometimes used the ward round to teach their students, asking them questions and getting them to look up information; occasionally this would relate to medicines reconciliation e.g. withholding medication crossing out medication rather than crossing it off the IMC entirely. One GP registrar stated:

‘l always make students look things up. That way they remember what they have looked up and that they need to look things up.’

Doctor 15, GP Registrar, A&E, Hospital A b) Support from central functions

YCFF Definition: Availability and adequacy of central services in supporting the functioning of wards/units. This might include support from Information Technology and Human Resources, portering services, estates or clinically related services such as radiology, phlebotomy or pharmacy.

214

Contributory factors

There were no observations categorised as contributory factors within support from central functions.

Defence

The pharmacy departments themselves are central function which provide support to pharmacy and non-pharmacy staff regarding medicines management and the implementation of relevant policies and procedures (see 5.4.2.1 section). The pharmacy departments were responsible for any policies and procedures associated with medicines reconciliation, together with any training. c) Policies and procedures

YCFF Definition: The existence of formal and written guidance for the appropriate conduct of work tasks and processes. This can also include situations

Contributory factors

All hospitals had policies and procedures in place which covered varying tasks associated with medicines reconciliation. The content of the policies varied significantly with some containing more detail than others (see table 5.8).

All hospitals had a medication formulary in place, however, the formulary in Hospitals B-D was more restricted (i.e. had fewer items), which meant that at admission to hospital a patient was more likely to have their medication switched to a medication which was therapeutically equivalent e.g. PPI. This risked the medication not being switched back at discharge and being changed in primary care. This is important as certain medications may not be as cost- effective in primary care due to differences in drug pricing systems in primary and secondary care.

Defences

The policies led to standardisation of practice throughout the hospitals with pharmacists where possible, reconciling medication post-admission and at discharge. Hospitals B-D used standardised forms to collect medication history information and specified that a specific coloured pen should be used by the 215 pharmacist so that it contrasted. Various other standardised practice was also in place for example, signing the IMC to say medicines reconciliation was complete or using standardised handover sheets or feedback proformas. Hospital D also had a standardised renal medication regimen (see also section 5.4.2). d) Physical environment

YCFF Definition: Features of the physical environment that help or hinder safe practice. This refers to the layout of the unit, the fixtures and fittings and the level of noise, lighting, temperature etc.

Contributory factors

There were times when the wards felt calmer than others, the opposite being indicated by the HCPs running around or in the way in which they spoke e.g. hurriedly or calmly. There were distractions, interruptions or obstructions which occurred on the wards. This was increased on the admissions wards compared with the other wards due to many people coming and going (both patients and staff). Noise levels varied depending on the time of day and other activities occurring on the wards. Noise was greatest in the morning as numerous tasks were being undertaken simultaneously, for example breakfast, the ward round, medicines administration and washing and dressing the patients. It was the time when most healthcare teams came on to the ward (pharmacy staff, physiotherapists, occupational therapists, consultants, housekeeping). Noise was also greater at other mealtimes and during visiting hours. Gaining a medication history during visiting times could be very difficult due to the noise and relatives around the bed. Conversely it could also make things easier when the relative or carer was there to provide missing information.

On three occasions men from estates departments were on the wards conducting maintenance which involved noisy activities such as drilling or hammering. Cleaning, especially the deep cleans, could at times be disruptive. On one ward the medical team were barricaded when they emerged from a cubicle as all the beds had been pushed to the centre of the room, with no route through.

216

Often the only place to sit to undertake tasks such as transcribing and writing discharge notes (electronic or paper) on the wards was at the nurses’ station (see table 5.13). This was also where the telephones were located and where the doorbell sounded. Staff tended to congregate around the nursing station conversing. One doctor said:

‘The only place to sit is at the nurses station but I get constantly interrupted and find it distracting with the phones.’

Doctor 24, GP Registrar, Elderly Ward, Hospital A

On one ward, in Hospital B, the computers used to complete e-discharge were in the staff room where people sat to eat their lunch. On another occasion in Hospital A, a local TV channel were filming a news feature which was distracting for the doctors trying to work at the nurses station.

Table 5.13 Summary of ward layout in each hospital

A B C D Sites for Nurses station Nurses station Nurses station Each Discharge and during & staff room & staff room discipline has Summary ward round own Writing designated area Position of Next to bed End of Bay Next to bed Varied – next POD to bed, end of Lockers bay, moveable trolley with separate drawers for each patient

HCPs were interrupted during medicines reconciliation tasks by patients and other HCPs. The doctors were broken off to do other jobs e.g. to attend a deteriorating patient, to rewrite charts and to re-cannulate patients. These needed to be prioritised effectively and often the discharge summary writing had to wait. The pharmacists were also interrupted to answer clinical questions, for example, how to administer IV medication or whether medication can be crushed, to ask the whereabouts of discharge medication or to order medication. In the main, it was the nurses that were distracted mostly by the routine needs of patients during medication administration round, with patients

217 requesting blankets, to be moved, to go to the toilet or to have a cannula adjusted etc. On one occasion a pharmacist was interrupted mid-task to interpret for a patient who only spoke Punjabi.

On occasion tasks could not be completed, for example, when all the phones were being used, the line was engaged or the notes/medication chart was missing or the patient was unavailable or the HCP had left a message on the district nurses’ answer phone and was waiting for a call back. This interrupted the workflow and relied on good documentation to remember where each task had reached.

Overall, during the observation HCPs were interrupted a total of 13 times by bleeps going off in the middle of the task. In some instances the HCP would break off to answer the bleep, in others the HCP would wait until they had finished the task they were undertaking before answering. It was unclear how the HCPs decided whether they answered the bleep or not. On one occasion the doctor (FY2) was in the middle of documenting information in the patient’s notes when he was interrupted by his bleep. The doctor (FY2) passed the notes to another doctor (FY1) to continue writing.

Two pharmacists explained that the layout of the ward could also make it difficult to do medicines reconciliation:

‘It is difficult on this ward as the POD lockers are not next to the bed. It means that if you want to go through the medicines with the patient then you have to carry them all to the bed which can be difficult if there are a lot of them.’

Pharmacist 14, Medical Ward, Hospital B

‘It can be difficult to find charts – you end up running backwards and forwards between nurses station’

Pharmacist 4, Renal Ward, hospital A

All wards had side rooms to isolate patients for example due to infection control risk. These patients were often left until last during all tasks. The pharmacy

218 staff in Hospitals A-C would not enter the side rooms unless they felt they really must, usually reviewing the chart from outside the door without patient involvement.

Defences

For some hospitals the layout of the ward facilitated medicines reconciliation, for example, where the POD locker was next to the patients bed (Hospitals A and C). Hospital D had a separate place for each of the healthcare teams to sit i.e. doctors, nurses and pharmacists leading to fewer distractions and less struggle for space and equipment (see table 5.13). e) Scheduling and bed management

YCFF Definition: Adequate scheduling to manage patient throughput minimising delays and excessive workload

Contributory factors

Some wards contained ‘outliers’i, patients who were allocated to a different specialty ward, not always appropriate to their clinical needs, often due to a bed shortage on the correct specialty ward. The patient was looked after by the nurses and pharmacists who worked on the ward, though the medical team working on the ward did not tend to the patient who was the responsibility of the medical team who they were admitted under. This created problems if there were any errors on the medication chart as the doctors on the ward would not make any amendments and the doctors from the relevant specialty had to be bleeped to come to amend a chart. Sometimes the bleeped doctor arrived after the pharmacist had left the ward therefore liaison could not occur.

i Hospital inpatients who are allocated a bed on a hospital ward which is not clinically appropriate for their illness due to a lack of inpatient beds are commonly named ‘outliers‟, ‘sleep-outs‟ or ‘boarders‟.

219

On 12 occasions the patient was medically fit to be discharged but were awaiting social care input or an appropriate placement.

‘Some patients have been here so long there are no medical issues they are just waiting to be placed’

Pharmacist 17, Elderly Ward, Hospital D

This time period could provide opportunity for HCPs to ensure the patient had been counselled appropriately on their medication and knew which medication had been changed. However, this was not observed.

Mealtimes in all hospitals were protected, which meant that patients could not be approached during that time. Also all nurses were occupied serving meals. As pharmacy staff were sometimes scheduled to the ward 9.00am -1.00pm and mealtime was served between 12.00pm and 1.00pm the pharmacists’ work could not always be completed.

Defences

Hospitals counteracted bed management issues through prioritising discharges, employing discharge liaison nurses and trying to forward plan discharge where possible. Hospital A employed discharge nurses which covered all medical wards, compared with Hospital D who had discharge nurses specifically employed for the Elderly Admissions Unit. It was unclear what happened in the other hospitals, however, discharges were always prioritised in all hospitals observed.

5.4.3.5 Latent external factors a) External policy content

YCFF Definition: Nationally driven policies/directives that impact on the level of quality of resources available to hospital

Contributory factors

Little external policy content was observed having an effect as a potential contributory factor to error associated with medicines reconciliation. All hospitals were subject to the four hour Emergency Care Standard which added

220 pressure for A&E to clerk quickly and the wards to discharge patients in a timely manner.618 Hospitals also had no control over undergraduate curricula and the content taught to their medical, nursing or pharmacy students. There were a couple of occasions when poor communication related to medicines reconciliation processes led to readmission within 30 days.j The first was a patient who was not counselled appropriately at discharge resulting in the patient taking double the amount of blood pressure medication:

‘A patient was admitted to hospital with low blood pressure. When I reviewed (reconciled) the medication the patient brought in with them, there were two boxes of the same blood pressure medicine in different packaging. After a recent hospital admission the patient had been discharged with a box of the medicine but because it looked different to the box they received from their community pharmacy the patient thought they were different and had been taking them both at home. This was an unnecessary readmission that could have been avoided if healthcare professionals had communicated more effectively with the patient and each other.’k

Pharmacist 5, Elderly Admissions Unit, Hospital A

The second, was a patient who was discharged without communicating the medication changes to either the patient or the community pharmacy:

‘I had an issue a few weeks ago where a patient was discharged - They were a repeat dispensing patient and their medication was stopped whilst they were in hospital. No-one let the community pharmacy know and they supplied it when she was discharged. The patient did not know it had been stopped either. They were re-admitted to hospital’l

Pharmacist 1, MAU, Hospital B

j In April 2011 the Department of Health introduced a policy of non-payment for emergency readmissions within 30 days.264 k This quote is used twice in two different contexts, once as a latent external factor and the second as a communication factor. l This quote is used twice in two different contexts, once as a latent external factor and the second as a communication factor.

221

Both these examples are also discussed in section 5.4.3.6 Communication Systems.

Defences

All hospitals involved in the study were influenced by external policy associated with medicines reconciliation, namely the NICE & NPSA Technology Appraisal Technical patient safety solutions for medicines reconciliation on admission of adults to hospital6 and the NPSA Rapid Response Report Reducing harm from omitted and delayed medicines in hospital.619 As the implementation of these alerts is mandatory, all hospitals had policies in place regarding delayed or omitted medicines and a list of critical medicines for which there should be no delay. These contained medication such as anti-epileptic medicines and anti- Parkinson’s medication.

Treatment plans were influenced by national guidance, in particular NICE guidance. When patients were nearing end of life care the Liverpool Care Pathway (LCP) was followed and all patients’ non-essential medication was stopped. The hospitals also had various Commissioning for Quality and Innovation (CQUIN)m targets in place which impacted on medicines reconciliation negotiated with the Clinical Commissioning Groups (CCGs)n e.g. Hospital A was about to introduce a target to send changes to medication electronically via SystmOneo to general medical practices post-discharge. b) Design of equipment and supplies

YCFF Definition: The design of equipment and supplies to overcome physical and performance limitations.

m The aim of the Commissioning for Quality and Innovation (CQUIN) framework is to support improvements in the quality of services and the creation of new, improved patterns of care. n Clinical Commissioning Groups (CCGs) are NHS organisations set up by the Health and Social Care Act 2012 to organise the delivery of NHS services in England. They replace some of the functions of primary care trusts (PCTs) which ceased to exist in April 2013. o SystmOne is an electronic health record (EHR) predominantly used in primary care, mainly general medical practices. Its unique feature is that it allows sharing of the patient’s EHR between HCPs and organisations.

222

Contributory factors

As electronic discharge was new it was not embedded into practice; staff were still adapting to new ways of working which may make them more prone to error. Despite three hospitals using e-discharge systems there were still differences between processes observed, for example, different fields within the different software systems.

Defences

All e-discharge systems used forcing functions to ensure relevant fields were populated. However, this only ensured the box was populated not that the quality of information was high. This was most likely because they were unclear what was useful. Hospitals had also designed their own IT solutions which assisted medicines reconciliation related tasks, for example, a local database for renal patients shared across the region and a database of community pharmacy and general medical practice addresses within Hospital D.

5.4.3.6 Safety culture

YCFF definition: Organisational values, beliefs, and practices surrounding the management of safety and learning from error

Contributory factors

The junior doctors’ rotations meant they worked in all four hospitals through rotating around all the hospitals in the study, yet they admitted that they did things differently according to the hospitals that they were in. Sometimes the same task was done with less rigour between hospitals due to the combined effect of lack of policy and procedure in place and the more senior doctors in the team being seemingly less concerned with how tasks which impact on medicines reconciliation were conducted.

In Hospital A, discharge summaries were written on the ward round as the patients were being discharged because there was an organisational drive to discharge the patients quickly. The junior doctors were observed to write the discharge summary balanced on the top of the patient’s IMC, flicking backwards and forwards between the chart and discharge summary. Rarely would they access the notes and then only at the end of writing the discharge note to 223 confirm the diagnosis and co-morbidities. They were also readily subjected to interruptions at that time.

On four occasions when the pharmacy team (in Hospitals B and D) were working in the notes they were interrupted by others for access to the notes. On three occasions the notes or IMC were taken away from the pharmacy team as they were needed by the ward round. In all hospitals staff using the notes would regularly be interrupted for a patient addressograph sticker. On one occasion the ambulance crew arrived to take the patient before the nurse had time to handover to intermediate care which was then rushed, and consequently little information shared.

Defences

A greater safety culture and organisational leadership within Hospital D was apparent. The organisational demonstrably placed high value on improving quality and practice. This was highlighted through observation of the infrastructure and the way HCPs within the hospital acted. One member of staff explained:

‘Pharmacist 17: It is important that we document things to prove our worth. Many of our activities contribute towards KPIs and with all these job cuts in the NHS it is important that we show what we do. The message comes down from the top.

Researcher: From Pharmacy?

Pharmacist 17: Yes, we listen to the people on the ground and what they want and what they feel will work for them. What we do is often backed up with incidents which support the way in which we work for example we had an incident a while back with warfarin. The patient was admitted with an INR of 24 and had taken a 1, 3 and 5 mg tablet. Luckily the pharmacist had recorded that she had done a level 3 review and counselled the patient as he [subsequently] said he

224

hadn’t been counselled. This shows the importance of documentation. You notice the team-working more on the acute floor especially with the MDT [multidisciplinary team] meetings where everyone attends and you have your say and they ask your opinion.

[Pharmacist 17 checks next IMC]

For example, this man hasn’t had his VTE prophylaxis for two days. There was another incident where a patient was admitted with a PE [Pulmonary Embolism], when they looked through his notes he had missed five days prophylaxis and the pharmacist had checked the chart every day, so now we all check.

Researcher: It sounds like incidents influence you a lot

Pharmacist 17: Yes, I think it is important that we learn from incidents especially where pharmacy is involved.’

Field Notes, Observation of Verification of Medication History Taking, Hospital D

This demonstrates a culture of learning and sharing from incidents, plus striving to improve and deliver services. The KPIs discussed were associated with quality and safety, for example, the percentage of medicines reconciliation undertaken within 24 hours post-admission. There was also an increased level of respect for each different HCP within the MDT and an understanding of each other’s roles and responsibilities. It was the only hospital to have a daily formalised MDT (although only within the elderly specialty). There was less hierarchical structure compared to that seen in other three hospitals within the study. There was increased communication and liaison between and within the disciplines with everyone asking everyone else for their input into patients, for

225 example, one pharmacist rang another to hand over a complex patient who was prescribed clozapine (an antipsychotic) and needed a blood test.p The walls of the corridors were lined with key messages and patient improvement campaigns, for staff and visitors to see, for example:

‘When we all work more closely together patients will benefit from higher quality and less duplication and a better well rounded service’

‘We need to get the communication right between the community and the hospital so people can get the care they need at the right time’

One ward had been involved in testing cooked breakfasts for patients on their wards. The results were clearly displayed. Although this is not directly related to patient safety it demonstrated their ability to use PDSA cycles plus learn and share their findings.

Hospital D also had various KPIs to drive quality and had developed a computer programme to record and monitor these. Plus they had introduced a single access portal, various e-forms and databases to their intranet for easy access to information and to allow these KPIs to be recorded (see also section 5.4.3.3 (e) Equipment and Supplies).

The staff in Hospital D understood the consequences of their actions and ‘doing a good job’:

‘If I do a good job on admissions then it is much easier on the base wards, this hospital puts a lot of emphasis on [verifying medication histories] admissions.’

Pharmacist 16, MAU, Hospital D

Although one pharmacist felt the high standards set a precedence for the amount of medicines reconciliation which should be achieved:

p Patients who are prescribed Clozapine require routine blood monitoring to ensure the medication is not affecting their immune system. Clozapine is tightly controlled and patients who have not had a blood test are not supplied with medication. If two days medication is missed the dose must be re-titrated.

226

‘I feel that all I do is medicines reconciliation rather than have clinical input. We have set a standard and now we can’t go back on it.’

Pharmacist 15, MAU, Hospital D

Although the other hospitals did not demonstrate as much in the way of safety culture, Hospital B did have a far greater amount of standardisation throughout the hospital. This was in the form of standard operating procedures and proformas. This was not limited to medicines reconciliation, but extended throughout all their processes. This allowed staff to work across different wards and be familiar with their processes and procedures.

5.4.3.7 Communication systems

This section presents the data concerning communication. The YCFF does not explore communication in great detail as the systematic review did not uncover evidence to underpin it. Therefore the HCP to patient interaction i.e. the consultation was analysed as described in section 5.4.3.7.1. using a consultation analysis framework specifically relating to medicines reconciliation based on the COT criteria610 and essential communication skills for a medication history interview611 to assess the interaction between HCP and patient.

5.4.3.7.1 HCP to patient interaction – the consultation

The extent to which the patient was included in the consultation or task varied between HCPs. Practice between individual HCPs ranged from actively involving the patient on most or all occasions to talking in front of the patient to colleagues without inclusion.

Consultant 1: Is he on regular meds they seem to have been omitted. He needs to be given his regular meds.

Patient: [to the HCPs standing around the bed] I can tell you what I take

[The HCPs ignore the patient]

Junior doctor (D2): I got the list from the GP printout. 227

Patient: Would you like me to tell you what I take?

[The patient is ignored again]

Consultant 1: Does he have his meds with him?

[The consultant speaks over the patient]

Nurse 2: You mean these?

[The nurse points to his tablets held in a first aid box above his POD locker]

The junior doctor writes the patient up for regular medicines without clarifying with the patient. I checked with the technician later and the patient had been able to give a full history and indicate what he does and does not take.

Field Notes, Observation of Ward Round, MAU, Hospital B

The processes were also approached differently both inter and intra- professionally. Doctors were more likely to use a systems based approach to eliciting the patients medication history during the medical history taking compared with the pharmacist who would ask questions specifically relating to the patients medication using a wider variety of sources. This was a direct result of undergraduate training. During clerking the doctors asked questions regarding medication in and amongst their past medical history using questions which filter for information, for example:

‘Do you suffer from Diabetes? Heart Disease? Epilepsy’ [yes] ‘And do you take any medication for that?’

Doctor 1, FY1, MAU, Hospital B

There were various ways in which the HCP gained rapport with the patient, most commonly smiling, occasionally passing comment on how well the patient was looking. During clerking the doctor often did not acknowledge the patients

228 passing comments if they were irrelevant which appeared to be a tactic to keep the patient on track and not be distracted. Sometimes this felt cold and as if the HCP was uninterested. Many of the HCPs would come down to their level, or sit on the bed to talk to the patient despite the latter being against hospital infection control policy.

HCPs used a mixture of open and closed questioning to elicit information from patients, for example, ‘Do you take any medication?’ or ‘Can you tell me which medication you take?’ Probing was used to follow up partially answered questions and to clarify information. Where closed questions were asked the patient sometimes answered yes/no answers and on other occasions would add other information; sometimes useful and sometimes not. Occasionally HCPs used leading questions and interrupted the patient before they had finished. For example, ‘Do you take these twice a day?’ rather than ‘How often do you take these?’. Overall, on six occasions, the HCP asked multiple questions, not allowing time for the patient to answer, for example:

‘Have you been using your nebules more recently? Do you use them all year round or just in summer?’

Pharmacist 4, Renal Ward, Hospital A

The elderly patients often needed more time to process the questions and respond. This required the HCP to be patient and remain silent whilst the patient answered. Some HCPs were better at this than others. Information was never repeated back to the patient to confirm understanding and occasionally relevant patient information was ignored with patient concerns not addressed (15 occasions). Only one doctor used the medication with the patient to try to prompt them to remember their medication history. This was done more frequently with pharmacy staff, although not routinely. Pharmacy staff also often prompted the patient to remember non-oral medication by asking them direct questions e.g. about inhalers, eyedrops etc. On eight occasions the HCP prompted the patient when they could not remember the name of their medication and on only two occasions this was appropriate, conversely, on the other occasions there were a range of options it could have been rather than

229 the one offered which may have led to the wrong medication being prescribed, for example:

‘Pt 461 Has had a medication history taken. The doses of digoxin, warfarin and bendroflumethiazide are missing. Consultant 17 examines the patient and takes a history. The consultant asks the patient which dose of digoxin they usually take. The patient tells the consultant they do not know. The consultant prompts the patient asking ‘Is it 62.5 or 125 [micrograms]?’ The patient responds ‘125 I think?’. Consultant 17 fails to offer the 250 dose as an option. It is unclear whether the patient is just choosing the higher of the two dose as they know it is not 62.5 micrograms.’

Field notes, Elderly Medical Ward, Hospital D

The HCPs occasionally made wrong assumptions about the ability of the patient to be involved (see also section 5.4.3.2 (a) Individual Factors):

‘It is lunchtime and patient X has not had an insulin dose yet today. The pharmacist rang the DN [district nurse] this morning to confirm the patient’s dose and left a message on the answer phone. The DNs have not returned the call. Eventually the pharmacist tried asking the patient. The patient knows the dose name and dose. It was assumed he would not know as he has a MDS and the DN usually administers his insulin. This is because the patient is blind.’

Field Notes, Observation of Pharmacist Verification of Medication History, MAU, Hospital C

There were several occasions where the HCP did not pick up on cues that the patient either did not understand, had concerns that needed addressing or offered information which may have been important.

‘The pharmacist was counselling the patient on warfarin, however, the patient did not respond or acknowledge the pharmacist. After the

230

pharmacist had been talking for a few minutes the patient said- ‘l have a hearing aid somewhere.’

Field notes, Observation of Patient Counselling, Elderly Admissions Unit, Hospital A

HCPs often did not realise that they were using terms that the patient may not understand for example ‘SystmOne.’ One pharmacy technician, expressed a need for standardisation to improve communication and understanding.

‘Patients often do not understand what we are asking when we ask if we can look at their GP record as everyone asks it differently. I think there should be some set phrasing so that we all ask it the same so that the patients will understand.’

Pharmacy technician 2, Elderly Medical Unit Hospital A

Several previously discussed contributory factors hindered the communication within a patient consultation e.g. patient factors such as confusion, dementia, hearing impairment (see section 5.4.3.2(c)); physical environment such as noise on the wards and interruptions (see section 5.4.3.2(d)). Equally some previously discussed defences facilitated communication within the patient consultation e.g. relatives or carers supporting the patient and acting as a source of information.

Communicating Medication Changes to the Patient

Communication with patients about their medication was inconsistent. Most patients were not informed about changes to their medication regimen during their stay or at discharge. On 51 of 138 occasions when changes were made to a patient’s prescription, the HCP did not clarify with the patient which medication had been stopped, started, changed or withheld, or they told the patient that some medication had been changed but did not explain which or why. If the HCP communicated changes to the patient this was done verbally. On two occasions the lack of information provided to the patient led to serious error:

231

‘Patient 671 is hyperglycaemic but is not acidotic. They have been put on variable rate insulin.…….. The consultant explains to the FY2 to administer half of his long acting insulin and to remove the sliding scale. She also asks the FY2 to explain to the patient. The FY2 hasn’t understood the consultant so I [the researcher] explain, half now and half tonight then back to the patient’s usual dose tomorrow. I also ask the consultant whether the patient has been told to use his short acting analogue as normal. The consultant snaps at me ‘well he will do’. Later on the nurse finds the junior doctor to explain that the patient has a CBG [capillary blood glucose] of 24 – this is because he hasn’t had his short acting insulin with his lunch.’

Field Notes, Observation of Ward Round, MAU Hospital D

Pharmacist 5, Hospital A

On five occasions the information the HCP gave the patient was vague and unhelpful requiring further clarification, for example:

Consultant to junior doctor: ‘Stop metformin. Stop bendroflumethiazide. Restart Losartan at half dose. Continue on simvastatin.’ q This quote is used twice in two different contexts, once as a latent external factor and the second as a communication factor.

232

Consultant to patient: ‘We are going to stop some of your medicines and change some of the doses’ [does not specify which]

Consultant 12, MAU, Hospital C

On one occasion it was only because the patient asked the HCP if there had been any changes to their medication that information was given:

Consultant: ‘Is there anything you want to ask about your medication?’

Patient: ‘Is there anything that’s changed?’

Consultant: ‘Yes prednisolone and omeprazole the brother to lansoprazole’

Patient: ‘What about iron?’

Consultant: ‘We don’t need to give you that because you can have it at dialysis’

Pharmacist 19, Renal Ward, Hospital D

Even when it was known that a medication had been changed, the rationale was not always clear to the patient or a HCP other than the one that made the change. Often the HCP was left to make an assumption about the alteration:

‘Consultant 17 is unclear why the patient has had their bisoprolol stopped. He then notices she has had two falls recently. He does not tell or ask the patient. He annotates his handover notes for the MDT.’

Field notes, Observation of Ward Round, Hospital D

5.4.3.7.2 HCP to HCP communication

Intra and inter-professional communication regularly occurred between pharmacists, nurses and doctors (GPs, junior doctors and consultants) regarding medicines-related issues. These included both verbal and written

233 communication (electronic and handwritten), at handovers, on prescription charts, in patient notes, on discharge summaries and observation charts.

HCPs liaised with a variety of other specialties, most commonly palliative care and psychiatry but also neurology, oncology, the pain team, ophthalmology and specialist nurses such as tissue viability, COPD and alcohol nurse.

Verbal communication

All HCPs including ambulance staff, nurses, pharmacists and doctors were involved in relaying information about patients to other HCPs usually of the same profession; mainly within secondary care but occasionally to primary or intermediate care.r This occurred either at the beginning or end of the shift, during the ward round, when a patient had been reviewed by a different team/ HCP or when a patient transferred care setting e.g. either to a new ward or intermediate care. Nursing staff would speak to district nurses to arrange medication administration in the community or care homes to ensure they had medication supplies. Handovers were most often verbal, although, occasionally, this would be written in the notes, and in Hospital B on one ward the nursing handover was audio recorded.

All nurses used the SBARs approach to handover information to their nursing colleagues. Despite this, there was variation in the amount of detail which was communicated. Some nurses and wards provided more detail about medication than others. Where medication was specifically mentioned during handover it was usually to provide information on medication which was regarded as potentially dangerous e.g. insulin and warfarin, medication which was being used in an acute situation e.g. Intravenous (IV) antibiotics and analgesia, or medication with which there was an issue e.g. the patient was addicted or

r Intermediate Care is the term used for a range of services that deliver short term care, support or rehabilitation to patients. It bridges the gap between hospital and home, helping patients recover more quickly, so they can cope in their own home, preventing either unnecessary admission to hospital or allowing earlier discharge to patients who are medically fit but need some rehabilitation. s SBAR (situation, background, assessment, recommendation) is a structured method for communicating critical information that requires immediate attention and action contributing to effective escalation and increased patient safety

234 needed it prescribing. One ward confirmed which ‘when required’ medication had been administered during the night and another also conducted a bedside handover with the IMC to handover medication administered. No other systematic methods of verbal communication were observed.

Written communication

The majority of wards had a typed handover summary, all of which contained allergy status and some of which contained medication related information especially of those patients which may need monitoring e.g. insulin and warfarin. On one admission ward all the handover information was noted down by the individuals as the ward staff felt patients did not remain on the ward long enough to maintain typed information up-to-date. Other HCPs as well as nurses would often use the ward handover sheets as a reference source and annotate them when they had conducted relevant checks on the patient or to annotate jobs which needed doing e.g. transcribing medication chart. Handovers between pharmacists included more detailed reference to medication plus an indication of which patients had had their medication history verified. For all HCPs, communication usually included documenting that a task had been or needed to be undertaken e.g. medication review, medicines reconciliation or blood monitoring, alternatively that an intervention had been made e.g. dose adjustment or discrepancy identified.

Several previously discussed defences facilitated written communication between HCPs within secondary care and between secondary and primary care. These included equipment e.g. IT and discharge boards, plus standardised documentation and proformas e.g. handover sheets and discharge summaries, individual and team factors (see sections 5.4.3.2(c) & 5.4.3.3(d)).

Communicating medication changes to other HCPs (Both Primary and Secondary Care)

Most frequently any changes to medication were communicated to HCPs through written documentation; on the IMC, in the notes or on discharge summaries. HCPs would check the IMC and the patient notes to determine any changes to treatment since they last reviewed the patient. The extent to which

235 medication changes were documented, however, varied. HCPs reported that it was often difficult to know why things had been stopped, started or changed:

‘It’s not easy to work out why things have been crossed off as it is poorly documented in the notes. It’s very difficult if you haven’t been on the ward round.’

D33, FY1, MAU, Hospital C

‘It's worse when you don't know the patient - you don't know why things have been started or stopped’

Pharmacist 12, Elderly Ward, Hospital C

Information communicated was often inconsistent, incomplete or ambiguous (e.g.  L – Dopa). Frequently, changes to medication would either not be indicated in the patient’s notes or would lack detail such as the name or strength of the medication, stating, e.g. ‘increase diuretic’ or ‘decrease BP medication’. Rarely was the reason for the change explicitly stated. On occasion the medication was temporarily omitted on purpose e.g. antiplatelet medication when a gastrointestinal (GI) bleed was being queried or simvastatin when the patient was prescribed a macrolide antibiotic. This was recorded in different ways. Sometimes it was withheld on the IMC using 6’s, other times being crossed off completely or only being recorded in the notes. One consultant told the junior doctor:

‘Don't cross it out completely or it will never get restarted.’

Consultant 1, MAU, Hospital B

All discharge summaries had a place to communicate changes in medication, yet, on occasion the HCP would complete this without referring to the notes or seemingly being sure that there had not been any changes. This seemed more apparent in Hospital A where there was a box which the doctor signed to confirm that there had been no changes. On occasion this box was left blank. If noticed by the pharmacist they would prompt its completion or complete it themselves, though it wasn’t always noticed. The electronic discharge system prompted boxes to be completed using forcing functions. 236

Discharge information was routinely communicated to the GP, yet other primary care providers receive the information ad-hoc. This could lead to potential error if relevant HCPs do not receive infomation. One pharmacist described an error which had occurred through the lack of communication with a community pharmacist:

Pharmacist 1, MAU, Hospital B

A further incident was observed where the patient had been taking an antibiotic for six months which usually would have been short term. On investigation the GP had accidently continued the medication following a previous admission:

‘The pharmacist P21 is checking a discharge summary and notices a patient Pt 660 has been on Doxycycline BD for 6 months. He looks back through the patients discharges to try and find it. He rings the GP to check the indication. The doctor cannot clarify exactly he says it was added following instructions on the discharge letter in May. It was on the patients discharge in May as medication that they had come in on and the patient had been told to continue it at discharge. It was not listed in the medications that had been stopped started or changed. The pharmacist explains that the patient probably came in on it - therefore was discharged on it.’

Field notes, Observation of Discharge Summary Verification, MAU, Hospital D

t This quote is used twice in two different contexts, once using the SCOS for analysis and the second using the YCFF.

237

Some hospitals had standardised systems for communicating about medication e.g. proformas, e-discharge or specific places to leave messages for the MDT, which improved communication.

The discrepancies were communicated to the prescriber in a variety of ways including annotating the front of the medication chart or completing a green communication slip and attaching it to the front of the prescription chart. Pharmacists in Hospitals B-D also routinely wrote in the notes acknowledging that something had been done or needed doing. Pharmacists rarely wrote in the notes in Hopsital A. On two occasions relevant information written in the notes by a pharmacist was overlooked on ward round.

GP Registrars tended to consciously complete the discharge summaries more fully and provide plenty of information as they had ‘seen the difficulties at the other side when the patient is discharged’ and felt it ‘obstructive’ and ‘frustrating’ if things were documented incorrectly.

‘I feel I write more than others. I've seen the difficulties at the other side when the patient is discharged’

Doctor 40, GP Registrar, Elderly Ward, Hospital C

5.4.4 Errors and discrepancies observed

Throughout the observation various errors and unintended discrepancies (UIDs) were observed. (For definition, explanation and classification of UID see section 2.3.4 and Chapter 4) They were identified during ward rounds, nurse administration, and pharmacist verification of the medication history. They were then classified using the taxonomy described in Chapter 4. Medication omission was the most frequently occurring UID (67.3%), followed by addition of medication to the prescription (6.4%) and strength decrease (6.4%) (see table 5.14). A total of 50 patients had at least one omission on their medication chart. The number of omissions varied from one to six omissions per patient. The most common omission was inhalers followed by nutrition and blood with laxatives then pain relief next most common (see figure 5.5). On four occasions medication was omitted from the prescription chart that was not required whilst the patient was in hospital, nor were they recorded elsewhere within the 238 patients notes. These included antihistamines, Zoladex® injection and hydroxycobalamin injection.

Table 5.14 Number of discrepancies identified and categorised by type in all hospitals

Number Percentage Drug Omission 115 67.3 % Drug Addition 11 6.4 % Strength Decrease 11 6.4 % Therapeutic Class Substitution 5 2.9 % Strength Increase 5 2.9 % Other 5 2.9 % Difference in Timing 4 2.3 % Change in Brand 3 1.8 % Dose Increase 3 1.8 % Formulation/ route of administration 3 1.8 % Dose Decrease 2 1.2 % Frequency Increase 2 1.2 % Frequency Decrease 2 1.2 % Drug Duplication 0 0.0 % Total 171 100.0 %

The majority of discrepancies observed were not seen to cause patient harm. One, however, led to the patient suffering an epileptic fit.

‘The patient usually takes a combination of two antiepileptic medication, lamotrigine and phenytoin. The patient was given the dose of phenytoin at a different time in hospital and had suffered a seizure. The consultant had forgotten the patient had a previous diagnosis of epilepsy and was considering investigations as felt was as a result of stroke. On questioning the patient the consultant realised what had happened.’

Field Notes, Observation of Ward Round, Renal Ward, Hospital B

239

Figure 5.5 Percentage of medications omitted from inpatient medication charts (IMC) within Hospitals A-D

6 Percentage (%) 4

0 PPIs Creon Insulin Statins Antacid Quinine Inhalers Nitrates Baclofen Diuretics Laxatives Eye drops Emollients Pain Relief Antiemetic Antithyroid Antihistamine Corticosteroid Antidiarrhoeal Antidepressants Benzodiazepines Bisphosphonates Antihypertensive Antiplatelet medication

Nutrition and Blood Supplements

On another occasion a patient was discharged with her INR still elevated. The police had to contact her to bring her back.

There were eight occasions when either medication (two patients) or a medication list brought in by the patient had gone missing (six patients). On three occasions it was blamed on the paramedics, three the A&E staff and twice the ward staff. This mislaying of medication and lists was not unique to one hospital.

Documentation from the ambulance staff and A&E notes were often unclear. On three occasions they were completely illegible. Two doctors stated that they never trusted what A&E wrote. On one occasion an FY2 reported that he did not know that tinzaparin had been administered in A&E and consequently administered a second dose.

240

Allergy status

The mandatory allergy status box on the IMC had not been completed on four occasions. In all these cases the allergy box was completed by the pharmacist (apparent from green pen). There were nine occasions when the allergy recorded on the patients medication chart did not match other sources, such as, the GP record, and four occasions when the patient was taking medication they were allegedly allergic to – two of these were an allergy to a constituent part of the medication they were taking. For example, for one patient the GP record stated that the patient was allergic to timolol, yet, they were using Cosopt® (which contains timolol); and another allergic to paracetamol when they were taking co-codamol (which contains paracetamol). The extent to which these were communicated to primary care were unclear. On eleven occasions the reaction was not recorded. On four of these, the reaction had been completed by the pharmacist (again identified by green pen).

Comparison of discrepancy rates in Chapters 4 and 5

The discrepancy rates identified within this chapter have been compared with those identified in Chapter 4. Both quantitative (Chapter 4) and qualitative (Chapter 5) data collection determined that omission was the highest discrepancy type (see table 5.15). However, the types of medication omitted appeared to be different in the qualitative observational study compared with the quantitative survey (see figure 5.6). Inhalers remained the most commonly omitted medication across both quantitative and qualitative studies, although the percentage of laxatives, eye drops, anti-hypertensives, pain relief and nutritional and blood supplements all appear much higher in the qualitative observation (see figure 5.6). It is unclear what the potential reasons may be.

241

Table 5.15 Comparison of the types of discrepancies determined during the prospective observational survey and the observational study.

Baseline Data Observation Data

Chapter 4 Chapter 5 Drug Omission 237 57.4% 115 67.3% Dose Decrease 20 4.8% 2 1.2% Difference in timing 18 4.4% 4 2.3% Drug Addition 17 4.1% 11 6.4% Formulation/ route of 15 3.6% administration 3 1.8% Dose Increase 12 2.9% 3 1.8% Frequency Increase 7 1.7% 2 1.2% Other 6 1.5% 5 2.9% Drug Duplication 4 1.0% 0 0.0% Therapeutic class substitution 4 1.0% 5 2.9% Strength Decrease 4 1.0% 11 6.4% Frequency Decrease 4 1.0% 2 1.2% Strength Increase 2 0.5% 5 2.9% Change in Brand 1 0.2% 3 1.8% Uncategorised 62 15.0% 0 0.0% Total Categorised 351 85.0% 171 100.0% Total discrepancies 413 100.0% 171 100.0%

242

Figure 5.6 Comparison between type of medication omitted during prospective data collection and observational study

6 Percentage (%) 4 Baseline Data Observational Data 2

0 HRT PPIs Creon Insulin Statins NSAIDs Antacid Quinine Inhalers Nitrates Baclofen Diuretics Laxatives Dementia Eye drops Emollients Pain Relief Antibiotics Antiemetic Antithyroid Rubifacient Glucosamine Antipsychotic Antiarrythmic Anticoagulant Antihistamine Corticosteroid Antidiarrhoeal Mood Stabiliser Antidepressants Benzodiazepines Bisphosphonates Antihypertensive Contraceptive pill Oral Hypoglycaemic Parkinson's Medication Antiplatelet medication Anti-obesity medication Nutrition and Blood Supplements

243

5.4.5 Researcher interventions

As previously discussed, there were times when the researcher had to move from a level of moderate participation (acknowledging the patient and explaining the study) to active participation depending on the situation. This included intervention or collaboration with a HCP if the researcher noticed something which may have caused potential harm to the patient. The interventions or communications with patients were anonymously logged and are summarised below.

In total the researcher intervened in 17 out of the 674 patient episodes. In seven cases this was because the researcher knew information which the observed HCP did not. This was because the same patient had been previously seen by the researcher, for example, on a ward round. This illustrates lack of joined up approach and continuity (see also section 5.4.3.3 (b) Staff Workload). Clinical advice was provided, when asked, in 11 cases, for example, whether it was the most appropriate course of treatment; technical advice in five cases, for example how to complete a discharge form; and advice on illegible handwriting in two cases. In several cases, the patient would be asked a question by the HCP being observed, however, would answer the question to the researcher or look for reassurance from them.

5.5 Discussion

5.5.1 Policy, process and practice

The overall patient journey within each of the hospitals observed was fundamentally the same for medicines reconciliation, although great inter and intra hospital variation in the detail of the systems and the processes was observed. This is similar to findings by McLeod et al who found great variation in medication systems within hospitals in England.384 Both the organisation and HCPs employed differing approaches to the process of medicines reconciliation. All organisations had policies and procedures in place which covered different aspects of medicines reconciliation related tasks plus standardised documentation to varying degrees. The inconsistency in approach within and between organisations was particularly evident within the process,

244 communication and documentation, and whilst there was some standardisation there was definite room for improvement. (Standardisation is discussed further in section 7. 7.1).

All tasks were conducted methodically following the models suggested by the IHI2 and NPC80 (collection of the medication history, verification to ensure the history matches what is prescribed, documentation and communication). Yet, different emphasis was placed on different aspects of the process by the HCP dependent on their knowledge and experience. This led to the implementation of individual ways of working which acted as a defence against medicines reconciliation-related errors and/or discrepancies. The interpretation of the term medicines reconciliation within the hospitals varied, which is unsurprising given the lack of clear definition within the policies and literature. Whichever terms hospitals choose to use is irrelevant as long as all staff are clear of the meanings and tasks which it incorporates.

The accuracy of the medication history is reliant on the use and accurate interpretation of relevant and up-to-date sources. HCPs must recognise the reliability of each source and the potential limitations each may have.235 It is well-established that this should be part of the medication history, especially given that approximately 60% patients in the UK take OTC or CAMs.149,258 Yet, only one HCP confirmed whether the patient was taking any non-prescription medication (OTC or CAMs). There is suggestion that potential lack of knowledge and unfamiliarity with CAMs may prevent important discussions amongst HCPs and the patient.620 No HCPs asked about illicit medication use except doctors in A&E where it was suspected. Nor did anyone ask specifically about the contraceptive pill. This is similar to findings by Collins et al who found that the contraceptive pill was frequently omitted in surgical patients.76

The introduction of the NICE & NPSA guidance in 2007 placed greater emphasis on ensuring medicines reconciliation was conducted within 24 hours of admission and that there was pharmacy staff involvement.6 This external policy influenced all organisations, demonstrated through the conduct of audit in all hospitals studied. Although hospitals were trying to achieve the 24 hour target, the extent to which the hospitals routinely collected the data varied.

245

Greater emphasis on meeting the 24 hour target is needed; nevertheless, the need for quantity should not override quality.

5.5.2 The impact of human factors

To err is human published in 1999 recognised that human factors are essential to understanding patient safety and why errors occur in the healthcare setting.11 Since then, the emphasis on understanding human factors and human error in healthcare has grown and is now more common-place within the UK. This has been embraced by the NPSA who also believe that understanding the role of human factors is vital to reducing human error and harm to patients.22 Within this study various factors were identified which could and did contribute to error associated with medicines reconciliation. These were countered by a number of defences which either existed or were performed by individuals. These are summarised in table 5.17 and discussed below.

Individual and Team Factors

Whilst all organisations had policies and procedures in place, the tasks conducted relied on the HCPs interaction with the patient and their interpretation of the information they were presented with. The reliance on a HCP to interpret information adds a human element which introduces vulnerability in the system. The main difference in approach was observed between the doctors and the pharmacists; the doctors using a body-systems based approach and the pharmacists using a medicines-based approach. This was not a lack of systematic approach by the doctors by and large, but, what appeared to be an ill-fitting model to elicit an accurate comprehensive medication history; the doctors main aim being to stabilise the patient and gain a diagnosis. The differences in HCP knowledge and experience impacted on individual and team performance. An individual’s way of performing a task develops and evolves over time through their experience. It also reflects an individual’s training prior to practice. Some staff appeared more conscientious and thorough, with some recognising their limitations and others being unaware of what they do not know (unconscious incompetence).621 This was observed through participants openly explaining the processes they were undertaking to

246 the researcher which they believed were correct or best practice, not always recognising their inadequacies.

Unconscious incompetence comes from the conscious competence learning theory or matrix (see table 5.16) which describes learners or trainees competence journey. It begins at stage 1 ‘unconscious incompetence’, passing through stages 2 and 3 ‘conscious incompetence’ and ‘conscious competence’ arriving at stage 4 ‘unconscious competence’. There is the potential for HCPs who reach stage 4 to become complacent and automated in their actions, whilst those who remain at stage 1 may conduct tasks without insight into potential issues that may contribute to error. This is supported by Abadel and Hattab who found variation in the way graduates self-assess their performance in comparison with the way they were assessed by experts. Those in a younger age group in particular overestimated their clinical skills and competency.622

Table 5.16 Conscious Competence Learning Model623

Competence Incompetence Conscious Conscious competence Conscious incompetence (Stage 3) (Stage 2 Unconscious Unconscious Unconscious competence (Stage 4) incompetence (Stage 1)

Throughout the observation there was a sense that the HCPs personality impacted on the interaction with the patient, demonstrated through the extent to which the HCP developed rapport with the patient or the willingness they had to ‘do a good job’. Unfortunately, this could not be measured within the study as the SCOS and other frameworks employed did not take this into account. Nor did the data recorded provide enough depth for this to be analysed. The impact of personality traits on job performance has previously been cited in the literature and its impact on patient safety is starting to be recognised. 624,625 Studies have also found that empathy can make clinicians appear more competent and that the influence of gender of the person interviewing during medicines reconciliation can potentially influence the outcome of medication history taking with differences in the information elicited depending on the same sex or opposite sex interviews. This was potentially due to a different set of

247 expressive, interactive and interrogative skills.626 Lively also found that interviewers with a higher risk taking propensity were the most effective at medication history interviews.627

Occasionally, two types of assumption were seen within this study. The first stereotyping and the assumption the patient would or would not be able to participate or provide information; the second assumption due to lack of clear communication either written or verbal. In 1620, Francis Bacon proposed that the human mind assumes order and regularity to function daily.628 This maybe goes someway to explain the number of assumptions made by staff. It can, however, be potentially dangerous as not all cases will fit the assumption. It is therefore important that HCPs have an awareness of the assumptions they make. Assumption is cited as a cause of error and potential harm.629 The medicines reconciliation process is complex and relies on staff working systematically, interpreting information correctly, and not missing any cues or vital information. Due to the complexity of the process and the lack of embedded defences it is unsurprising that error and discrepancies occur.

Teams and groups of HCPs such as doctors, nurses and pharmacists can exhibit different values, beliefs, attitudes, customs and behaviours630 which can manifest in ‘turfism’u or working within ‘silos’.v Inter-professional collaboration has become increasingly necessary631 and is especially required in terms of medicines reconciliation as aspects of the medicines reconciliation process are managed by several professions throughout the patient’s journey. On occasion, more so within hospitals A-C there was a sense of ‘turfism’ and territoriality between wards and professions, especially towards A&E and ambulance staff. Barriers between professions need to be broken down to increase interprofessional collaboration.

u Turfism – Also known as territoriality v Silo – To work in isolation from others

248

Table 5.17 Summary of factors which contribute towards or prevent medicines reconciliation related discrepancy and/or error

Factor Contributory Factor Defence Active Failures Staff purposefully violating policy and None identified procedure e.g. providing logon details to another member of staff Communication Poor consultation skills e.g. multiple Appropriate consultation structure and Systems questions, leading questions consultation skills e.g. active listening Not picking up on cues/ red flags Standardisation of communication e.g. Use of jargon SBAR, proforma Assumptions/ lack of clarification Patient factors which hinder communication e.g. hearing impairment Safety Culture Learning from error MDT meetings Intentional rounding KPIs Latent External External Policy 4 hour A&E target NPSA Med Rec alert Factors Content Lack of teaching of medicines NPSA Delayed and omitted medicines reconciliation at undergraduate and rapid response report postgraduate level Routine audit CQUIN targets Design of Forcing functions IT Portal Equipment and Contact databases Supplies Latent Physical Noise Layout of ward Organisational Environment Interruptions Space for each discipline factors Lack of workspace Poor ward layout

249

Factor Contributory Factor Defence Policy and Poor standardisation Standardisation of practice and Procedures Lack of detail documentation Scheduling and Outliers Prioritisation of discharge Bed Protected mealtimes Discharge liaison nurse Management Support from Pharmacy department Central Production of policy and procedure Functions Training and Lack of training on IT Staff induction Education Lack of medicines reconciliation training Pharmacy Technician Training Local working Equipment and IT – e-discharge Logon creates accountability conditions Supplies Lack of telephones Forcing functions Reasons for changes to medication documented Lines of Delegation of task e.g. writing in the notes None identified Responsibility delegated from a consultant to a more junior doctor Management of On-call Rotas Staff and Unpredictable workload Management of workload Staffing Levels Lack of continuity of staff Delegation/ negotiation of jobs No weekend cover Prioritisation of new patients No lunch cover Scheduled training Staff Workload Unpredictable workload due to Prioritisation unpredictable nature of admissions Rotas Supervision and Poor delegation None Identified Leadership

250

Factor Contributory Factor Defence Situational Individual Different approaches to task Systematic and conscientious ways of Factors Factors Awareness of limitations working Assumption Clarification of information when uncertain Pharmacists as safety net Prescribing immediately after clerking so Observation of environment medication is not lost Cues and red flags ignored Conscientiousness Different approaches to task Sixth sense that something isn’t correct Patient Factors Hearing impairment Involvement of carer Visual impairment Self-administration Knowledge of medication Patient raises concerns abut medicines Language difficulties they usually take Acute condition Chronic condition Patient asleep Anxiety due to hospitalisation Recall of medication Task Medicines reconciliation relies on Characteristics interpretation and processing of information by HCPs (clinical decision making) which cannot be standardised or automated Team Factors None identified Systems implemented by team e.g. handover sheet, jobs lists

251

Patient Factors and Patient Involvement

Training and Education

Within this study only the pharmacy technicians had received training on how to take an accurate medication history. It is important that all staff involved in medicines related tasks understand the impact that it may have on reconciling medication, plus an awareness of the knock-on effects lack of communication can have on error and patient safety. This could be achieved through multidisciplinary training and education at both undergraduate and postgraduate level. Successful interventions within undergraduate programmes which specifically focus on medicines reconciliation have been implemented in the US and have led to improvements in confidence, knowledge and skills.93,96,328 In the UK, The British Pharmacological Society have identified learning outcomes which should be incorporated into the undergraduate medical prescribing curriculum; many associated with medicines reconciliation.338 The extent to which these have been embedded and/or adopted into the medical curriculum, though, is unclear. Training and education, however, are not always the solution to bad practice (This is discussed further in section 7.10).21

Electronic discharge has many positive aspects, yet still it creates a number of further issues which need addressing, some of which are due to the software itself. Others can be attributed to the way in which IT is used by individuals, thus highlighting the need for standardisation and training. Further work must be conducted to evaluate the role of IT within medicines reconciliation especially electronic discharge communication. Three of the hospitals had access to primary care records (through SystmOne and SCR). These have been shown to facilitate medicines reconciliation.235 Nevertheless, like any sources these need to be interpreted correctly. SystmOne was contemporary, however, the extent to which the SCR was up to date varied (in some cases up to three months old) which could potentially introduce error.

Safety Culture

Developing a culture of safety is key to improving patient safety and quality within the healthcare setting.11 The NPSA explains that safety culture is

252 achieved when staff within an organisation have a constant and active awareness of the potential for things to go wrong. Both the staff and the organisation are able to acknowledge mistakes, learn from them, and take action to put things right.22 It arises from a shared set of values, beliefs and norms related to patient safety. A review by Weaver et al found that interventions to improve safety culture differed, however, often included team training, team communication tools, executive walk roundsw and interdisciplinary roundingx, some of which were seen in Hospitals B-D (but not all).632 The YCFF framework used, like communication, did not expand on safety culture nor explore its potential contributory components or ways of identifying safety culture within an organisation. Further work is needed, potentially using markers of high reliability organisations to measure the safety culture. It is difficult to observe safety culture through tangible measures, therefore triangulation of observation with other methods such as interviewing, focus groups and questionnaires to elicit individual beliefs and values may be more appropriate. (Safety culture is discussed further in section 7.9.)

5.5.3 Communication

The communication observed could be divided into that which occurred between HCPs and that which occurred between the HCP and the patient, including both verbal and written. The appropriate use of open and closed- ended questions in relevant context is important during information gathering. Open ended questions allow the respondent to express themselves freely provided they feel comfortable, resulting in a larger wealth of information. Closed questions provide information which lacks depth and variety. There is greater chance of bias if closed questions are used as they may force the patient down a certain path and can condition the respondents thinking. Therefore, it may not reflect their true feelings. Multiple questions can be misleading and make it difficult to know which the patient has answered.

w Visits by hospital executives to patient care areas to discuss patient safety issues with providers x The planning and evaluation of patient care with health professionals from a variety of other health disciplines

253

Communication regarding medicines-related issues was variable. Consistency and understanding of information communicated could be improved through greater standardisation of procedures and documentation e.g. IT systems with forcing functions for mandatory fields, although this can potentially lead to assumption that the IT will act as a ‘safety net’.633 In addition to system changes, HCPs need to understand how the next person in the medicines management pathway is affected by the completeness and quality (or otherwise) of information provided and how error can contribute to preventable harm. Where possible, patients should be given the opportunity for increased involvement in discussions regarding changes to their medication regimens to help to improve medicines optimisation.45 Feedback to primary care was lacking especially regarding the patient’s non-compliance and incorrect allergy status, preventing issues being addressed once the patient returns to their usual place of residence.

The communication frameworks used for analysing the HCP-patient communication provide a useful model for teaching and observing medicines reconciliation interaction, although it does not examine the content of the consultation, just how it occurs. To analyse the interaction fully the communication frameworks must be used in conjunction with a further framework which also explores the content of the consultation. There is potential to combine the medication history taking section of the SCOS with the medicines reconciliation consultation framework to provide a comprehensive tool for analysis of the medicines reconciliation consultation. (This is discussed further in section 7.11)

5.5.4 Error and discrepancy

Similar to other studies omission was seen as the most frequent discrepancy during this part of the study and matches findings from Chapter 4 and in the literature. 77,109,125,145,166,171-183,201 No errors or discrepancies were seen to cause harm to the patient. Several, however, resulted in an ADE. It is surprising given the number of discrepancies and current practice that more events do not result in patient harm. This has been recognised by Boockvar who looked at the number of ADEs occurring from medication discrepancies. He determined that of 1350 discrepancies, only 65 (4.8%) resulted in ADEs.159 254

The verification of allergy status is integral to the medication history taking, to determine the appropriateness of subsequent medication. This sometimes lacked description of the reaction and was not always verified by the pharmacists. Where it was, the information often conflicted with medication already prescribed. This mirrors international studies which show that often the allergy is not described appropriately in both adults634 and children.635 As with any conflicting information source it is important that this is fed back to the source owner, in many cases the general medical practice, to resolve such discrepancies (see also communication).

Outliers

‘Outliers’ were more likely to have an unresolved discrepancy or have to wait longer for a discrepancy to be resolved. The doctors on the outlying ward refused to make alterations to the patient’s treatment and the doctors from the patient’s admitting specialty had to be contacted to amend a chart. There is little evidence looking at the increased risks associated with staying on an inappropriate ward whilst in hospital, however, previous work has determined that outliers are disadvantaged as they usually receive worse management517 and less timely and thorough medical review.636 Goulding found that outliers doubled in the winter months and were more common in medical and elderly specialties.636 Within Goulding’s study outliers were also usually significantly older than other inpatients and would endure a significantly longer hospital stay and an increased number of internal transfers.636 This was not observed in within this study. It was often the most challenging and unpopular patients including those that suffer confusion or dementia.636 It can be difficult to obtain a medication history for these patients, without the complication of them also residing on an inappropriate ward. Older patients, and those having longer stays in hospital, have been shown in some studies to be more likely to experience a medication discrepancy, potentially increasing the risk further if they are also an outlier.111,126,173,180,189,212

5.5.5 Frameworks

The SCOS provided a useful framework through which to measure the quality of current practice. The SCOS was limited in that it did not contain sections on

255 nurse administration of medication or ward round. For future use, these could could be added based on findings from this study, including standards to ensure where medication discrepancies are identified during administration of medication they are flagged to a relevant HCP and during ward round that medication which is stopped, started, altered or withheld is appropriately documented in the notes and communicated effectively to patients and other HCPs, including primary care and the community pharmacy. Adaptation of this framework could be used to provide recommendations and as an audit tool to aid improvement. (This will be discussed further in section 7.10.)

The YCFF provided a framework to categorise contributory factors which may lead to medicines reconciliation error. This work also provided the first opportunity to test the framework in an observational study within a clinical setting. It identified contributory factors which could and did lead to error in medication information gathered. Defences observed could also be categorised using the contributory factors framework. Through identifying contributory factors and defences, potential solutions and interventions can be implemented to improve the accuracy of information. It did not necessarily follow that a defence which counteracted a particular contributory factor would fall under the same heading, for example, lack of completion of a form may have been an active failure or individual factor, yet, a counteracting defence may be an IT solution with forcing functions i.e. design of equipment and supplies.

An observed practice may appear to be both a contributory factor or a defence depending on its context. For example, isolating a patient in a side room may be an infection control defence, however, it may be a contributory factor which leads to poor communication with the patient. For example patients who were staying in side rooms were often left until last on the ward round as the HCPs had to wear personal protective equipment (PPE) such as aprons and gloves to enter the room. These patients were often then rushed and not all HCPs chose to enter the room to communicate with the patient because of the extra infection control barriers needed. There is often overlap between those elements considered to be communication factors or culture, and other contributory factors. On many occasions, communication or safety culture (according to the YCFF) appear to be secondary to other underlying factors e.g. a health care

256 professional not completing the discharge summary in full may be due to an underlying contributory factor of lack of equipment e.g. inaccessible IT or a lack of leadership which results in poor safety culture may be down to individual factors. This results in a communication factor or safety culture factor but communication or culture cannot be confidently isolated as an original causal factor. There is interdependency between the factors on the YCFF. In order for a process to be robust all the potential contributory factors must be considered.

One might argue that the most robust system has fully engaged staff therefore if staff (and patients) fail to engage with the system appropriately, there is a clear potential for error. Staff engagement is one of the aims of high reliability organisations (HRO).y These will be discussed further in section 7.9.

Communication was analysed separately due to lack of expansion of the YCFF using consultation models. Both of which have previously been used in the analysis of observation and which were directly applicable to medicines reconciliation.

5.5.6 Strengths and limitations of Study 2

As the strengths and limitations of the observational work are pertinent to both primary and secondary care data collection these will be addressed in Chapter 6 (section 6.5.6), unless specific to secondary care. These will now be discussed.

Researcher intervention

The researcher intervened on a number of occasions, to act in the best interests of the patient. This potentially affected some of the data, however, gives a useful insight and raises questions regarding what would have happened if the researcher was absent. Many interventions occurred because the researcher was present during both the ward round and the verification of medication history which incidentally supports the argument for having pharmacists present on ward rounds.74 The researcher needed to maintain a balance between

y An organisation that has succeeded in avoiding catastrophes in an environment where normal accidents can be expected due to risk factors and complexity.

257 remaining a neutral observer and gaining rapport. This was achieved through courteous acts such as pulling the curtain round a patient’s bed, saying ‘hello’ to the patients, passing items when asked and relaying messages e.g. if a patient needed to go to the toilet. The researcher also engaged in answering clinical queries if posed by HCPs, as again it was in the best interest of the patient but also maintained rapport. The researcher additionally ensured the correct demeanour and appearance, as well as adhering to any hospital policies e.g. infection control and dress policy.

Co-rating

Using a mixture of pharmacists to co-rate the data could potentially introduce disciplinary bias. Yet, the employment of a pharmacist demonstrated the importance of using someone who understands the medicines reconciliation process, whilst the health psychologist had a broader understanding of human factors. This potentially demonstrates the importance of a MDT in application of the YCFF especially the involvement of clinicians. An inherent weakness of applying the YCFF to observation is that not all of the contributory factors are as easily observable as others e.g. It is difficult to observe whether something is an active failure, however, others are more easily measurable e.g. the environment or layout of the ward. Medicines reconciliation is an objective process bound in social interaction and behaviours, therefore a combination for example, observation with follow up interviews, may be appropriate. Think aloud/cognitive techniques may be a different option although this would decrease the spontaneity and flow of the HCPs actions. However the documentary analysis did provide further insight into what the organisations intended the process to be. This may have accounted for some of the low Kappa co-efficients.

Some argue that secondary analysis of data loses the original researcher’s connectedness with the data and other coders are unlikely to interpret the nuances in the data in the same way as someone who has directly experienced the field.564 Nevertheless, on balance it was felt that the increased inter-rater reliability and decreased interpretation would outweigh the above, to ensure that the researcher did not place more emphasis on certain aspects than others. Also, the most effective application of the YCFF as a data analysis tool 258 demanded that the person observing also carried out the analysis so that the context of the observation was understood. This was demonstrated within the study as the co-raters asked for more context in certain circumstances to be able to categorise the data. For example the co-raters asked for more detail regarding the isolation of patients in side rooms, the junior doctors writing in the notes and the patient having more than one IMC.

5.6 Conclusion of Study 2

This chapter has identified the key stages and roles of HCPs within the medicines reconciliation process, identifying where current practice lacks quality. The medicines reconciliation within the hospitals observed was inconsistent and relied heavily on individual interpretation of information with lack of cultural driver and no real standard approach. Communication from HCPs to other HCPs and HCPs to patients was unreliable and often lacked essential detail. This included varying provision of information to primary care. From the observations it was unclear how communication to primary care was being received in practice. The next chapter will explore the receipt of information in primary care.

259

CHAPTER 6 STUDY 3: FINDINGS FROM PRIMARY CARE OBSERVATION

Previous chapters have focused on secondary care. They have identified the scale of medicines reconciliation issues at admission (Chapter 4) and described secondary care processes and HCP roles within the process (Chapter 5). This chapter explored the receipt and management of medication-related information within general practice. It looked at those members of general practice staff involved and how they interpret and process the information.

6.1 Introduction Chapter 2 highlighted the difficulties which occur with medication post-discharge with the receipt, reconciliation and processing of medication information within primary care. There is often lack of clarity whether discrepancies post- discharge are a result of error on discharge or inaccurate medication history taking at admission. If the GP EHR (used as a source) is inaccurate, it can lead to an inaccurate medication history at admission. The inaccuracies will still be present at discharge. This demonstrates the spiral nature of medicines reconciliation and the inter-dependence of subsequent stages on the accuracy of previous stages. Unintentional changes made with medication, especially dose instructions, omissions and additions, occur in supplies obtained post- discharge,379,637 with medication being accidently discontinued or added in primary care, especially where the repeat medication list on the GP system is not updated appropriately.379,430 Despite most of the evidence on poor communication at discharge being over 15 years old, there is little to suggest that it has changed. This is supported by a systematic review by Bayoumi et al from 2009 who concluded that there was no good quality evidence demonstrating effective medicines reconciliation interventions in primary care434 and that further studies looking at this area were needed. The continuing problem and lack of evidence make a clear case for empirical study and in turn this next part of the thesis.

In common with the observational study in secondary care, this chapter uses an evidence-informed observation schedule but for primary care (Primary Care Observation Schedule [PCOS]) through which to observe and analyse the 260 processing of correspondence from secondary care. The processing of discharge summaries within general medical practices is examined to identify good practice and any issues which arise, to determine areas where intervention may be appropriate. It uses the information processing model by Crowe et al as a template on which to base the PCOS, which in turn was then used as a framework to analyse the results.429 This involves six stages: receipt of information, dissemination, amendment of records, reading and effecting recommendations, scanning information and informing the patient (see figure 6.1).

Figure 6.1 Information Processing Model429 (reprinted with kind permission)

261

6.2 Aims and Objectives The aim of this chapter was to examine the current medicines reconciliation practice within primary care together with the role of HCPs, administrative staff and patients within the process. The specific objectives:

 To review the systems of reconciling medication within general medical practices in primary care.  To determine the roles of primary care staff with respect to medicines reconciliation.  To identify any practice which may lead to error and compromise patient safety.  To identify any practice and solutions which enhance patient safety related to medicines reconciliation in general practice.

6.3 Method 6.3.1 Overview

The methods used within this part of the study were similar to those in Chapter 5; a deductive framework plus inductive observation were combined to determine current systems within general practice associated with medicines reconciliation. The same rationale for using observational methods in the secondary care observation applied to this part of the study, including lack of previous qualitative evidence in this area plus a need to establish the context of reconciliation. The methods described here are based on the methodological theory described in Chapter 3.

6.3.2 Access to general medical practices and recruitment of staff

A letter including information about the study and consent information was sent to all general medical practices within one PCT area to ask them to participate in the study (see Appendix M). The PCT was served by all the hospitals which participated in the previous parts of the study (seen in Chapters 4 and 5). Study information was also placed in the PCT newsletter and emailed to key people for dissemination, for example, the Medical Director, Deputy Medical Director, and general managers to circulate amongst relevant groups. Following each practice’s agreement to participate, a date was arranged for observation and

262 they were asked to send their relevant protocols relating to medicines reconciliation to the researcher prior to the practice visit. Consent information was sent in advance for participants to read and ask any questions. At the practice, staff were recruited for observation based on their job roles i.e. if they usually conducted an aspect of medicines reconciliation as part of their role, and if they were willing to participate. The staff were identified through a conversation with the practice manager on arrival at the practice. The study was explained to the participant and written consent obtained.

6.3.3 Ethical considerations in Study 3

Ethics approval was gained via the LREC and relevant R&D departments as described in Chapter 3 (see Section 3.9 Ethical Considerations). The same ethical principles regarding the observation of participants are relevant within this section of the study as those outlined in Chapter 5.

6.3.4 Development of primary care observation schedule (PCOS)

The reviewed literature described in Chapter 2 was used to develop an evidence-informed PCOS of good practice for processing medicines related information within primary care (see Appendix N) , using the same approach as for the Secondary Care Observation Schedule (see section 5.3.4). The PCOS is divided into two sections:  Part A looks at practice details including policies and procedures, education and training and audit  Part B covers observations of processing information about medication, including receipt, dissemination, review, amendment, scanning, informing the patient and repeat prescribing through observing the relevant members of staff. Each section was further divided into criteria which should be met according to the literature e.g. where there is uncertainty information must be clarified. The participating organisations were asked for any policies, guidelines, protocols or documents that related to the way in which information was processed following discharge. This may include opening post, processing repeat prescriptions, scanning information etc. These were emailed in advance to the researcher.

263

This structured approach allowed systematic collection of data, and increased the consistency of approach within practices. Where organisations had policies and documents in place associated with medicines reconciliation, these were obtained prior to observation and the content compared with the evidence based PCOS. Where a practice’s written policy and/or protocols include relevant items not already included in the PCOS these were added. Due to the limited amount of evidence on medicines reconciliation within primary care, this part of the study tested the observation schedules in practice, serving as a small scale observational study, hence these forms were not piloted in advance of entering the practices.

6.3.5 Data collection

Criteria included in the PCOS were marked as met, not met or partially met. Other identified practices, plus answers to questions which clarified practice were noted during the observation. The researcher determined the most relevant person within the general medical practice to assist with answering information on the demographics of the practice and the most appropriate people to observe with assistance from the practice manager or administrative staff managers. The study was explained to the HCP or administrative staff member and consent obtained. The participant was observed and the PCOS annotated as appropriate. Where further observations were made which fell outside the PCOS, notes were made within the comments box or on the reverse of the framework, ensuring the corresponding criterion number was written by the side of any comments. During observation the researcher asked non- leading questions to clarify and document the process accurately. The length of observation varied from 30 minutes per person in practices where several members of staff were involved in the process to four hours where one person processed all the information. Observations were conducted on weekdays during September, October and November 2011. This is a busy time of year for admissions to hospital, due to winter pressures, thus providing the potential for increased intervention of processing information post-discharge.638 The same researcher was used throughout the study to eliminate inter-operator variability. No patient identifiable data were recorded during the observations.

264

6.3.6 Data analysis

There is a small amount of evidence within primary care that looks at the processing of information, which allowed the development of the PCOS, thus allowing framework analysis to be used for the majority of data produced. There is less evidence highlighting the difficulties with interpreting and reviewing secondary care correspondence, therefore where relevant observations fell outside the framework thematic analysis was used to interpret the data. Observation was conducted until the research reached theoretical saturation when the data confirmed the analysis rather than provide any new information.559

6.4 Results

6.4.1 Demographics of general medical practices

Eight out of 88 general medical practices within one PCT in the health economy being studied agreed to participate. Two practices subsequently withdrew due to limited staff capacity, leaving six (Practices A-F). Their practice populations ranged from approximately 3500 to 24,750 patients per practice and they were served by all four hospitals involved in the study to varying degrees (predominantly Hospitals A and C with some correspondence from B and D) (see table 6.1). Due to the number of employees in Practice C it had its own human resources and finance department. The length of time the staff had worked at the practice ranged from two to 24 years. The majority of participants were administrative staff with some GPs and pharmacists involved (see table 6.2).

6.4.2 Policy

Each practice had general policiesz or protocolsaa in place for dealing with all communication received. The processes conducted within each practice matched all the policies that were in place, however, the majority of documents

z Policy - A statement of intent by an organisation to follow a particular course of action. aa Protocol - The accepted or established code of procedure or behaviour in any group, organisation, or situation

265 lacked content. The policies varied in style, content and layout with some being more comprehensive than others and two including a visual flow diagram (see table 6.3). The number of pages per protocol varied from one to six sides of A4. None were structured in the same way and all chose to focus on different aspects of the process.

Table 6.1 Practice Demographics

Practice Total Number of Number of Number of Number of Hospitals Number Registered GPs Hours Branch Liaised of Hours Patients employed Practice Surgeries with Observed (WTE)* Pharmacist Employed per Week

A 6.0 6495 5.4 4 0 ABD

B 5.5 7122 5.5 4 0 ABD

C 6.5 24748 15.9 4 3 ABCD

D 6.0 10833 8.1 4 0 ABD

E 7.0 12137 10.8 4 1 ABD

F 5.5 3508 2.0 4 0 ABCD

*Whole time equivalent

Table 6.2 Number of participants in study categorised by role

Role Number of Participants

GP 4

Pharmacist 2

Administrative staff 14

Total 20

266

Within the policy three included step-by-step instructions on how to process information focusing predominantly on how to navigate the computer system. The other three provided a general overview of what happened with incoming information. Three specifically mentioned the processing of discharge information and only one medication, the others focusing on the processing of incoming mail in general. None specifically included details on how medication information is processed. Three protocols briefly mentioned who conducted which stage of the process but did not specifically state who could and could not process information about medication. No policies were available on how information on medication is provided to other HCPs when requested e.g. on admission to hospital. All practices had repeat prescribing policies in place due to a previous PCT KPIs. The repeat prescribing policies described the roles and responsibilities of those involved in the repeat prescribing process e.g. prescriber, receptionist, pharmacy staff with reference to which members of staff can initiate or authorise medication within the practice. The repeat prescribing policies had a date approved and review date with an appropriate review period, however, none of the other policies received contained this. There was no documentation which indicated whether staff had read the policies or whether they would adhere to the content.

6.4.3 Training on medicines reconciliation related processes

None of the practices specifically provided training for their staff on medicines reconciliation or processing of discharge information. The participants described having completed ‘on-the-job’ training through shadowing others. Several staff had been on training courses regarding medical terminology, and summarising notes especially where it was deemed that this was required by their role.

6.4.4 Audit

Three practices had conducted an audit of their processes; two looking at the number of discharge summaries processed and the timescales, the third a more detailed audit looking at the error rates on discharge information. No other aspects had been audited.

267

Table 6.3 Content of policies and protocols for processing correspondence within general medical practices

Shred

efore it efore is

Description B

Discharge Discharge and

eviewed R entioned epresentation ischarge ummary ummary canned pecifically M Practice Pages of Number Diagrammatic R When is Summary D S S is Roles of responsibilities Staff Practice Medication S A 1  5 days Yes X X Step by step visual representation of entire process B 1.5 X Immediately No X  General overview of process

Details on how to read code, scan and attach to Not C 6  Yes X X patients’ records. mentioned Diagrammatic representation for dealing with post Section on dealing with discharge summaries, Not Not D 2 X   referral letters and palliative care diagnosis with mentioned mentioned examples. No mention of scanning and read coding Not Not E 1 X X X Includes how to scan and attach to the patients mentioned mentioned record Handling and sorting post, who is responsible for F 4 X Immediately No  X each stage, how to deal with electronic mail

268

6.4.5 Receipt of information

Each practice received information from a range of different healthcare organisations. All practices received information from secondary care hospitals, in the form of discharge advice summaries and also outpatient or consultant letters. Four reported receiving information from the community pharmacy, this was usually MUR forms and repeat prescription requests. Practices would occasionally receive discharge information and repeat prescribing requests from the patient or representative e.g. relative, care home, in advance of the discharge advice summary being received by the hospital. All practices reported that the patient would have a carbon copy of the discharge advice summary, which was often illegible as it was the last carbon copy of a quadruplicate. Practices sometimes received information regarding patients who were not registered at their practice and one practice stated that on two occasions the hospital had asked a patient to deliver someone else’s discharge summary to the practice because they were registered at the same practice, potentially raising data confidentiality issues. As well as information from the hospital, information was also received by other care providers such as tertiary care, mental health, or specialist screening services.

The information was predominantly received via post (hard copy) twice daily. The amount of post received per day varied from approximately 30 items in the smallest practice (F), to 300 in the largest practice (C). Approximately half to two-thirds of these were discharge summaries. From some specialty areas, however, e.g. diabetic retinal screening, A&E attendance and information from Women and Children’s services was received electronically. As this was new, practices were cross-referencing the electronic with paper copies to ensure all information was received; ultimately the aim was to be paperless once secondary care was satisfied that the system was working. Very occasionally a faxed discharge summary or letter were received. The faxes varied in legibility.

In all practices, the information was received by administrative staff and the paper-based letters opened and date stamped. Processes within practices varied and often depended on the size of the practice and the amount of mail received. Post contained information related to a variety of different clinical

269 issues not just relating to medication e.g. discharge summaries, letters, radiology reports, blood results, diabetic retinopathy and cytology reports. Three practices sorted their post into piles of similar sorts. These were all processed as a bundle of all the same type. In one practice where photocopies, faxes and carbon copies were received they were kept on one side until the original is received then cross-referenced. A practice may receive information from more than one healthcare organisation for each of their patients, especially where they had been referred to tertiary care or for specialist screening e.g. echocardiogram (ECGs) and bowel cancer screening. The layout of the discharge summaries observed varied from all four hospitals. The mental health discharge summary did not have anything to indicate whether any changes had been made to medication.

The largest practice (C), which had four branch surgeries, acted as a ‘head office’ for their other branch surgeries, processing all of their information and booking their appointments over the phone. The administrative service ran in a similar way to a call centre and processing unit.

The length of time between the episode of care and the information arriving at the practice varied. A&E attendance was usually received within 24 hours. Practices reported receiving discharge summaries within a couple of days post discharge with follow up letters taking four to six weeks depending on factors including the workload and efficiency of the consultant and secretary, or whether the consultant is on annual leave. On two occasions (out of 333) discharge summaries were observed arriving 10 days post-discharge. The practices sometimes relied on the follow up letters to clarify issues in the discharge summary, however, these were not routinely linked or compared on arrival. Occasionally duplications were received in the post. Practice A described compiling a ‘flak’ list and ‘task’ list. A ‘flak’ list for problems or issues e.g. a patient requesting an item before the discharge summary arrived and a ‘task’ list for prescription requests which were urgent and could not wait.

Overall post was usually processed in 4 stages - Opening post, scanning, attaching to the patient record and read coding (new diagnoses or reason for admission). Read codes are a catalogue of codes which represent clinical terms for example diagnoses and investigations. They are used to record

270 patient findings and procedures in health and social care IT systems across primary and secondary care e.g. General Practice and pathology reporting. The codes enable records to be searched for activity and audited. Correspondence was read coded either before or after the review stage depending on the individual practice. In three practices, administrative staff read coded the information sent but did not make any amendments to the medication or patients record. In the remaining three practices, the read coding was conducted by the GP or practice pharmacist.

6.4.6 Scanning of information

In all practices, correspondence was routinely scanned to the computer by administrative staff. In Practice D the correspondence was annotated with a ‘s’ to denote it had been scanned. The information scanned satisfactorily if it was on white paper and not a carbon copy. Scanned documents were poor and illegible if on coloured paper which did not scan well. It varied from practice to practice whether scanning was completed before or after review and amendment by a relevant clinician. One practice said:

‘We scan first in case they get lost’

Administrative staff, Practice E

All practices except one (Practice B) scanned the same type of correspondence together. One practice articulated:

‘You need to ensure sheets go through individually. If the letter is two pages they needed to be linked together within the EHR to ensure that they do not get missed.’

Administrative Staff, Practice F

Once scanned, the file is attached to the relevant patient’s notes. This was initially done manually, however, during the observation period the EHR system (SystmOne®) used in all practices within the PCT observed had a software upgrade. This included writing recognition software, which recognised the name and NHS number of the patient and linked it to their record. Once

271 attached to the patient record, some manual tasks were conducted, for example, a date entered so it appeared in the record at the appropriate time and assigning the relevant type of correspondence e.g. discharge summary or x-ray report etc. so it was easily identifiable. Different practices entered different dates to indicate the episode of care. Most used the date the discharge summary had been signed to represent the date of discharge. This could, however, have been before the date of discharge if the discharge summary was prepared in advance. One practice used the date the discharge summary was received which could be much later than discharge. Practice A linked the new correspondence to any previous communication e.g. the referral letter so that the trail of information could easily be found. This was the only practice to do this.

Once scanned practices would shred the paper correspondence. The amount of time the letters were kept for before shredding varied between practices. Some shredded straight away, others kept for a month to allow time for the patient to order a new repeat. Practice E annotated the EHR with a blue flag once it had been scanned and assigned to ensure none were missed. In the smaller practices the scanning and read coding was usually done by the same person and all staff shared duties, for example, in one practice the nurse was also the summariser, and the receptionist a medical secretary and phlebotomist.

6.4.7 Dissemination of information received

The way in which information received was distributed between HCPs depended on the practice. Two practices divided the work equally between the GPs, one allocated it to the doctor who last saw the patient or referred them to the hospital. If a doctor was not specified then it was allocated to the doctor ‘on call’ or ‘hot-doc’. One practice disseminated all the letters received that day to the on-call doctor. In Practice C the majority of letters were addressed to the main partner therefore it was felt it would be an unfair allocation if he was to review all of them. Practice F distributed to the relevant doctor (named on discharge), some still coming through to a doctor who had previously retired. In one practice the GP reviewed the information from a diagnosis point of view and to ascertain any follow up required, then, signed the corner and passed to the 272 pharmacist to process the medication related information. This practice and one other practice allocated anything medication related to the practice pharmacists for review although this was sometimes actioned by a doctor when a pharmacist was not available.

One practice indicated that if the doctor was on maternity leave, one person is allocated to deal with their mail. In practice C where the information was distributed between several doctors the member of staff indicated that

‘Sometimes it is difficult to know who to send it [the information] to.’

Administrative staff, Practice C

In the practices where the information was scanned before review, the workload was shared amongst all GPs using a task allocation function on the EHR. The work was shared equally, irrespective of whether they were part time or not. Sometimes the information was sent to one clinician for info and another for action.

6.4.8 Review of information

Who In the majority of cases communication about medication was reviewed by the GP although occasionally this was also done by either a nurse or the practice pharmacist. The nurses had varied roles within the process. In practice F, the nurse was also a notes summariser and in another (Practice D) the community matron who looked after a caseload of elderly patients reviewed discharge summaries.

How Two practices separated the discharge summaries into two piles; those that need actioning and those that had no actions. They relied on the hospital doctor completing the discharge summary accurately and indicating whether there had been any changes to the medication. The rest of the practices reviewed all of the discharge information. One pharmacist queried:

273

‘I don’t know how other practices do it? They just look at the [discharge summaries with] changes and take them [the discharge summaries] away’

Practice Pharmacist, Practice D

The clinician reviewing the discharge summary cross-referenced the medication on the patient’s repeat template in their EHR with the information on the discharge summary. One pharmacist estimated that approximately 2/3rds of discharge summaries did not have any actions as the patient was young or had been admitted to the hospital as a day case.

The clinician reviewing the information used their clinical judgement to determine whether any changes needed to be made to the patient’s medication. They looked at the diagnosis to see if the changes look reasonable e.g. addition of aspirin, statin and beta-blocker post myocardial infarction (MI). One clinician added a review date to specific medications, for example, PPIs so the patient did not remain on a treatment dose and was reduced to a maintenance dose. Another stated:

‘Where cardiovascular medication has been changed and they have seen a cardiologist I assume that they know what they are doing and the changes are correct.’

GP, Practice C

Five of the six practices had an electronic formulary embedded into their EHR which identified more cost-effective alternatives when medication was being prescribed for a patient. The other practice relied on the clinician using their judgement and remembering to consider cost-effectiveness.

Some clinicians were more thorough than others, checking blood results & monitoring. The pharmacists took a more holistic approach than the doctors, reviewing other issues e.g. compliance. The practice pharmacists would send tasks to the GPs to remind them to monitor patients e.g. urea and electrolytes (U&Es). One of the pharmacists annotated the correspondence to say it had been reviewed and also read coded the fact that the hospital information had been reviewed. It varied within the practices as to whether the read codes were 274 checked by a clinician. Practice B electronically highlighted on the scanned information what needed read coding. One GP from Practice A, routinely double checked the read codes were correctly allocated. Often it was unclear whether the choice to review information in a specific way was unique to the clinician or the general practice.

Two practices reviewed the allergy status recorded by the hospital to check it matched with their records. Where the practices checked the allergy status they found that sometimes the two did not match or the patient was taking something their records said they were allergic to, for example, on one patient’s record it said they were allergic to codeine but they had been taking co- codamol. Often the nature of the allergy was missing.

Error The discharge summary often contained omissions or changes in strength of medication with no indication on the discharge summary that anything had changed. Of 25 discharge summaries observed, a total of 24 omissions were observed on 10 discharge summaries (ranging from one to six per summary). Three contained one addition each and three had a dose that was either increased or decreased compared with before admission. For example, the discharge summary of one patient who used a MDS made no mention of a diuretic the patient took routinely, showed their morphine dose to be halved and a statin which they had not taken for three months to be restarted. There was no indication on the discharge summary that there had been any changes to the patient’s medication. In another case only two out of the eight medicines on the discharge summary matched the patient’s record. Again the discharge summary was signed to indicate there were no changes. The discrepancies included omitted insulin and the addition of one medication the patient had not had for six years. Occasionally, the clinician reviewing the medication noticed obvious clinical omissions which were not in line with national guidance, for example the patient was prescribed triple therapy with no long term PPI or the patient had not been initiated on a statin post MI. Where correspondence did state that a medication had been stopped sometimes the reason was unclear. The clinicians observed offered some suggestions why errors may occur. One practice pharmacist indicated:

275

‘Occasionally medication which was prescribed acutely prior to admission is continued post-admission and added to the repeat.’

Practice Pharmacist, Practice A

A GP stated:

‘I do not trust hospital discharge information. Sometimes the patient has been in hospital a long time therefore medication changes are not documented as they move from ward to ward.’

GP, Practice B

There were also five occasions where a patient’s calcium and vitamin D preparation had been switched to an alternative which was stocked by the hospital and not amended back at discharge and one occasion where the PPI omeprazole had been switched to a different PPI, lansoprazole.

There was variation in the legibility of the information received, clinic letters, follow-up discharge letters (from all hospitals) and e-discharge notes (from Hospitals B, C & D) were typed which were easily read. Hospital A used handwritten quadruplicate carbon copy forms, which could be difficult to read, especially the last copy.

The information in outpatient letters from consultants were often the most difficult to interpret. It was sometimes unclear whether the letter was just for information or whether the doctor needed to prescribe a supply for the patient, for example, one letter emphasised that Equasym XL (a controlled drug) had been working for the patient but it did not state whether the consultant was to continue to prescribe for the patient or whether any had been supplied. This can also be the case when the patient is prescribed an unusual product e.g. special formulation. In other cases it is unclear whether it is an acute course or long term. One letter stated ‘Prescribe prednisolone’, but did not indicate strength or whether it was a short course or long term. Some consultants left the choice of the medication to the GP or gave the doctor a number of options e.g. try X or increase Y. Another letter said ‘should you decide to treat with a non-volume depleting drug, avoid a calcium channel blocker’ but was not

276 specific as to what should be used or what the alternative options were. If the letter included a medication history there was often inconsistency whether the form and strength were documented within and between letters. In one case, the medication which was specific to the specialty contained full details, however, for the others just the name was listed.

There were two cases of misinterpretation by the clinician reviewing the information where the researcher had to intervene. The first was misinterpretation of the annotation ‘POD’ on the discharge summary. The clinician thought that it meant that the patient had been on the medication prior to discharge when it was used to mean that it could be found in the POD locker (part of the patient’s bedside cabinet where the medication is locked away). The second, misinterpretation of ‘continue until 2/9’ written on the discharge summary. The clinician thought they had meant to put 2/7 (meaning 2 days) when they meant continue until 2ndSeptember.

There was one case where the hospital had increased the medication of a patient not realising that the patient did not comply with his treatment. The patient is regularly readmitted.

Clarification of information Correspondence varied as to whether there was sufficient information to review the patient’s medication accurately. Where there was uncertainty, information was not routinely followed up with secondary care. The clinician reviewing the information rarely contacted the hospital to clarify any discrepancies as it was deemed ‘too difficult’, the patient’s notes had often left the ward, the doctor who treated the patient was off shift or it relied on the hospital doctor’s memory to clarify the question. One GP highlighted:

‘It can sometimes take several hours to resolve queries from discharge prescriptions.’

GP, Practice B

More often the clinician would make a decision or occasionally ring the patient to clarify. In one practice, the discharge summary is put on one side until the

277 follow-up letter arrived, to see whether it added any clarification. The GP makes sure the patient has enough supplies until then.

6.4.9 Amendment

After the information was reviewed, the relevant amendments were made to the EHR. This was always conducted by a clinician in all the practices observed; either the GP, pharmacist or nurse all of whom had relevant access on their smartcards to enable this to happen. Practices had previously been sent a letter from the PCT advising them to review the access administrative staff had on their smartcards, limiting prescribing and amending of prescriptions only to clinicians. Although the practice managers had access rights to review and amend prescriptions, this was something which they definitively reported they did not do.

Some GPs reviewed the read codes assigned by the administrative staff and checked any new diagnoses given. In the majority of cases the practice did not need to assign brand new read codes, they usually linked the episode of care to existing read codes assigned to the patient.

A record of any shared carebb medication was annotated on the EHR as well as any medication the hospital prescribed. There was no specific place to put this on the system therefore each practice assigned it in a different place. If patients were issued warfarin by the hospital, Practices B, C and F put it on the repeat template, but made the quantity one tablet so anyone seeing the record would know and the system would highlight any interactions. This was not necessarily a robust method.

The time taken from receipt of information and making relevant amendments on the system varied depending on the way in which the information was received and who in the practice was reviewing the information. If it was received electronically, it was usually reviewed and amended the same day. Non- electronic took 24-48 hours. If the person reviewing was part time it could take bb Shared healthcare is a model of integrated healthcare delivery in which the collaboration among practitioners of different disciplines or with different skills and knowledge allows for the delivery of patient healthcare by the most appropriate healthcare practitioner

278 four days to a week. The largest practice reported it took one to two days to scan plus a week to make amendments.

6.4.10 Patient information

Patients were not contacted routinely in any of the practices to inform them of changes to their medication. Where they were contacted, it was more often by the practice pharmacist. One practice pharmacist said:

‘If they are elderly or likely to be confused I may ring or pass to nurse case manager. I try to ring the patient but, where there is no answer, I write.’

Pharmacist, Practice D.

In Practice F, the practice pharmacist put a note on the prescription to inform the patient. They also rang a nursing home to reinforce that medication had been stopped.

6.4.11 Repeat prescribing

Only one practice reported regularly running a report to identify items on the repeat prescribing template that had not been recently ordered. Each patient record identified was reviewed to check whether the item of medication should be removed including identification of any compliance issues. If anything was identified the patient was asked to call in for a medication review. This was carried out by the practice pharmacist. The pharmacist added:

‘Community Pharmacists often order on behalf of the patient and are not aware there have been any changes.’

Practice Pharmacist, Practice A

Practice B had a system in place to notify the community pharmacy of changes to MDS. They fax the changes to the community pharmacy and the pharmacy signs the bottom and faxes back to say they have made the amendments. On two occasions repeat prescriptions had been printed before amendments had been made to the repeat prescribing template post-discharge. 279

6.5 Discussion 6.5.1 Overall process

All models which describe medicines reconciliation contain at least three stages; collection of medication history, validation of any existing medication lists and clarification.2,3,80,340 They are clearly more focused towards medicines reconciliation within secondary care. The reconciliation undertaken within general medical practices does not incorporate all of these stages. At admission to hospital information is gathered from a variety of sources before processing during the medication history stage. However, although primary care receives the list of medication, they do not gather the information themselves, therefore the medication history taking stage is not applicable. It could be argued that there needs to be a different model for medicines reconciliation within primary care or acknowledgement within the models as to how they can be applied within different settings. Even the visual representations of the patient journey discussed in Chapter 2 considers admission, within a hospital stay and discharge but does not consider what happens in primary care.

Practices have a role in the provision of medicines information at admission to hospital whether this is planned or unplanned or on referral to an outpatient clinic. This was not observed whilst in primary care; furthermore, none of the practices had policies in place regarding this. The accuracy of the medication history taken at admission can affect the accuracy of discharge communication. If medication is not recorded on admission it may be accidently omitted by either the patient or GP when the patient is discharged as they assume it is not to be taken.379 Similarly if GP records are not kept up-to-date it can cause error at admission. It was sometimes difficult to tell whether medication had been purposefully stopped or whether they were never recorded on admission. This mirrors findings from over 20 years ago.379 Although practices all had their own individual approach to processing medicines-related information they all contained the key stages previously described by Crowe et al.429 It would not be appropriate to implement one-size-fits all approach to the processing, review and amendment of discharge information due to the differences in practice population, number of staff and amount of correspondence received. Nevertheless, as protocols lack information there is need for the practice 280 protocols to contain standardised sections and increased detail on all stages in the process. This could be in line with the stages of the information processing model: receipt, scanning, dissemination, review and amendment.429 Also incorporating potential pitfalls and good practice recommendations, for example which member of staff is responsible at each stage and the provision of information to hospital at admission. Increased detail within the policies including visual flow diagrams would be beneficial. Staff observed had little training, if any at all, on processing medication-related information. This suggests there is increased need for staff to have appropriate training, plus competency assessment. There were no records that staff had read the protocols employed. Staff should sign to say they have read relevant protocols and practices should maintain an up-to date record. Previous incentives to have robust repeat prescribing policies in general medical practices had proven successful within the area, with all practices having repeat prescribing protocols in place, based on an evidence-based template provided by the PCT. The extent to which the practice matches the policies is, however, unclear. The successful implementation of repeat prescribing protocols in general practices mirrors previous research which indicates that financial incentives improve quality and therefore potentially patient safety.639

6.5.2 Review of information

The GP is the central record holder in primary care and on receipt of discharge information will review and update the patient’s records. Ultimately this then may inform the patient’s medication history if they are readmitted. This study could not disprove or corroborate previous findings that medication is accidentally discontinued or added in primary care,432 or that medication is often added to existing regimens without previous medication being deleted or stopped leading to duplication of therapeutic effect. However, it is easy to see where this may occur as the reviewing clinician sometimes made assumptions regarding information within the discharge summary. This in turn may have led to the repeat medication list on the GP system not being updated appropriately, similar to findings by Cochrane et al and Burns et al.379,430 On occasion, within this study, general practices relied on the patient, their relatives or their carers to provide information, especially where discharge information is conflicting.

281

The patient can sometimes present before the discharge summary is received, similar to previous findings by Mageean and Harding in the 1980s.42,519 It seems unreasonable that nearly 30 years on, and with the technology available that discharge information cannot be received in a timely way in primary care.

It is important that post-discharge amendments are made before repeat prescriptions are requested. One Australian study showed that within 30 days post-discharge, 71% of patients will visit their GP and 86% will visit their community pharmacy within two weeks, making it important for primary care HCPs to receive and action the discharge summary within this time.428 Practices could add in filter questions when patients request repeats to determine whether there are likely to have been any changes, for example, ‘Have you been in hospital or had an outpatient appointment since your last preacription? Did they make any alterations to your medication?’

Work should be fairly allocated to staff on a pro rata basis. The NPC stipulate that discharge information must be reviewed by someone with appropriate skills. Crowe et al429 discuss the amending and updating of records post- discharge as a task exclusively carried out by GPs themselves, whilst this study found that in practice this is sometimes conducted by pharmacists and in some instances, nurses. This may have been due to the differences in timing and location of the studies. The NPC notes that in primary care, responsibility for the process is sometimes inappropriately devolved to an untrained person, for example a receptionist, who does not have appropriate competencies. Administrative staff were heavily involved in the processing of discharge information, however, no evidence of administrative staff amending the EHR was found. A systematic review by Burns et al in 2012 determined that accuracy of coding post-discharge lies between 80 and 90%, suggesting that routinely collected data is sufficiently robust to support its use for research and managerial decision-making.640 This still leaves 10 to 20% of patients with inaccuracies in their coding which could potentially lead to failure to identify them during routine audits. Clinician involvement in coding has been suggested to improve accuracy.641 The systematic review by Bayoumi et al434 highlighted the need for improved quality evidence demonstrating effective medicines reconciliation interventions in primary care. This study identifies specific areas

282 where intervention may be appropriate e.g. improvement of procedures and protocols.

6.5.3 Discrepancies identified

A number of discrepancies on discharge summaries were observed. Although there are few studies that look at the number of discrepancies within discharge information from a primary care perspective, there are studies conducted within the hospital that look at the number of discrepancies appearing on discharge information. The findings within this section of the study are in keeping with the evidence that the omission is always the highest discrepancy found during medicines reconciliation.74-79 It was often unclear which medication had been stopped, started or changed during the patient’s stay in hospital. There is a role for the hospital to improve the quality of their discharge summaries, to ensure accuracy and to clearly indicate changes. This would assist the practices with reconciling medication for their patients. There is also a need for more standardised outpatient letters, with clearer information as to whether the letter is for information only or whether the general medical practice must undertake an action. A summary of action points within the letter would be helpful. There is a continuum between medicines reconciliation and the review of medication at discharge as often the clinical appropriateness of the medication prescribed is reviewed by the clinician whilst they are processing the discharge summary.

Only two of the six practices checked the allergy status on the patient’s discharge summary. This is important and should be included in the review of discharge information as one US study showed that approximately one-third of patients had a discrepancy in their allergy status.512 This could lead to medication being prescribed to which the patient was allergic or alternatively the patient being denied appropriate therapy because the prescriber believes the patient is allergic.

6.5.4 Involvement of the patient in medicines reconciliation

There is need for any changes to a patient’s medication to be communicated effectively and for the patient to be involved in their own care. Cromarty et al demonstrate that the provision of the discharge summary to the patient at

283 discharge was valued by the patient and decreased the incidence of medicines related problems post-discharge.392 It is also important that anyone else involved in the patient’s care is also notified e.g. community pharmacist, care home, relative or carer; all who may order repeat prescriptions on the patient’s behalf.

6.5.5 Role of the hospital

The NPC specifies that discharge information should be clear, unambiguous and legible and should be available to the GP (or other primary care prescriber) as soon as possible after discharge, preferably within two working days.80 The legibility of discharge communication within this study varied; the quality of the handwritten carbon copies was often poor. This is similar to previous findings,41,401 especially where the correspondence is the third carbon copy of a discharge summary or a handwritten document.80 The majority of correspondence was received often before it was requested by the patient i.e. typically 11-14 days, however, there were occasions when this was not the case. Similarly, a quantitative study by Akram and Urban380 conducted in the same health economy found that 80% of discharge summaries were received within seven days and took four to nine days to be actioned by the general medical practice. Other international studies have found similar.373,426,427 This timeliness is not in keeping with the NPC two day guideline. The method of delivery of discharge summaries varied include posting, hand-delivery by the patient, faxing and electronically. This was also found in previous studies.40,103,366 Although the proportion of faxed and postal receipt is decreasing as the amount of electronic transfer increases. The move to electronic discharge summaries will reduce the issues with illegibility due to handwriting and quadruplicate carbon copies, plus reduce workload from opening post and date stamping plus other issues e.g. mis-filing and lost information. The implementation of a standardised formulary across both primary and secondary care would also minimise the potential for discrepancies, e.g. duplication, as on occasion patients had been switched to alternative medication for the same therapeutic purpose at admission e.g. PPIs and calcium and vitamin D preparations. The introduction of IT can potentially

284 introduce new issues as staff become familiar with new ways of working (see section 7.6 Information Technology)

6.5.6 Strengths and Limitations of Study 3 and overall observation

This section will discuss the specific strengths and limitations within this chapter and also the overall strengths and limitations of collecting observational data (relating to Chapters 5 and 6).

Observer bias There are potentially a number of biases which can occur through observation e.g. the Hawthorne effect, the Halo effect and observer bias. These have previously been discussed within Chapter 3. As discussed earlier, previous observers of health-related situations have noted that their impact is often minor due to the complex, demanding work needing to be undertaken in a timely manner.238 In a similar way, the medical staff did not appear concerned at being observed, stating they were ‘used to it’ because of various assessments and appraisals. Staff were also very busy and often admitted to forgetting they were being watched. On several occasions the participants asked the researcher ‘How did I do?’ or ‘Did I pass?’. One participant asked for personal feedback as they felt they did not usually receive peer review and thought it would be helpful to inform their practice. Staff were regularly unconsciously incompetent, and therefore would conduct their tasks not knowing what they did not know and assuming they were conducting tasks appropriately, therefore appearing unconcerned by the observation. Notably, the interpretation of the original work which uncovered the Hawthorne effect has been re-examined, having called into question the quality of the original methods used and the conclusions drawn, which ignites a debate as to whether the Hawthorne effect is a true 583 phenomenon. Observation schedules were also used which provided more objective, structured observations.

Data recording Observation alone cannot provide a full picture and only provides an illustration of what is happening at the time. Not observing for the whole day or ‘out-of- hours’ created a potential to miss key events which could have provided a more

285 comprehensive view. Field notes were used to record the observational data. It is argued that this may not be as accurate as other means as it relies on documenting everything seen – almost instantaneously, however, ethical considerations would make it difficult to have used video or audio recording. As has been explained earlier, observations were recorded against an observation schedule where possible, plus included verbatim accounts rather than the researcher’s construct of a general sense of what was said to prevent the researcher opinion influencing the reporting.573 Furthermore, notes were expanded as soon as possible post observation to improve accuracy.

Reflexivity The researcher needed an awareness of their own cultural assumptions and any preconceived ideas, especially given the researcher’s additional, ongoing role as a pharmacist. This was countered by being reflexive and recording any feelings and opinions which were later reflected on e.g the researcher remained aware that she may observe the pharmacists more favourably due to being from the same profession (halo effect). Nevertheless, the researcher’s role as a pharmacist enriched the information gathered as they used their prior knowledge and experience to gain access, build rapport, understand the process and pose appropriate questions to the participants. It also allowed the researcher to be flexible and make a judgement on what to observe.642 Miles et al 1998 have criticised the lack of collaboration between health service researchers and clinicians, suggesting a closer link is needed.643 Research- practitioners can go some way to reduce this gap and may assist access to different areas of healthcare.

Framework The SCOS provided a useful framework through which to measure the quality of current practice. There were some parts of the PCOS which on reflection maybe were not as relevant to medicines reconciliation, for example, some of the standards relating to repeat prescribing e.g. a designated area to process repeat prescriptions, although this is important in general for the governance and accuracy of repeat prescription generation. Adaptation of this framework

286 could be used to further observe general medical practice and also provide used as an audit tool to aid improvement and provide recommendations.

6.5.7 Conclusion of study 3 This chapter reviewed the systems of reconciling medication within general medical practices and determine the roles of primary care staff with reference to medicines reconciliation. It also identified practice which could lead to error and solutions which enhance patient safety related to medicines reconciliation in general practice. Practices vary in their approach to processing information on medication received from other organisations. This can depend on a number of factors including the size of the practice and the number of staff employed. Although it would be unrealistic to expect every practice to process medicines- related information in the same way, this study has highlighted specific aspects of the process which may contribute to error. Some practices and clinicians have embedded good practice into their processing, which could be shared with others. There is a need for standardisation and improved policies and protocols, together with improved electronic communication, improved accuracy and timeliness of discharge information and standardisation of outpatient letters. Improvements are, however, necessary within both primary and secondary care to improve the overall process and to make the transitions between primary and secondary care seamless, including patient involvement to ensure medication changes have been embraced by the patient.

287

CHAPTER 7 GENERAL DISCUSSION AND CONCLUSIONS

This thesis has utilised mixed-methods across multiple centres to examine and critically appraise the whole medicines reconciliation process. The study followed the patient journey from admission, through to discharge, and into primary care. Firstly, the study determined the type and severity of discrepancies occurring at admission to hospital. Then, subsequently, examined the context of these findings through a qualitative observational study of current medicines reconciliation practice in both primary and secondary care. Two evidence-informed, observational schedules were developed and tested, and used as a framework for analysis. The YCFF608 was also applied to identify potential factors which contribute to medicines reconciliation error/discrepancy, but also, outwith its intended role, to identify and categorise potential defences which could lead to the amelioration of such errors or discrepancies. This chapter will discuss the major findings within the study and their implications for practice in the context of other published literature relating to medicines reconciliation. The study limitations will also be discussed before considering a range of recommendations specific to practice and policy.

7.1 Definition and Taxonomy of Medicines Reconciliation

The literature review highlighted the lack of a standardised definition of medicines reconciliation, both internationally and within the UK. This point was reinforced through the observational work and analysis of the policies and procedures in both the hospitals and primary care. Pharmacy staff were familiar with the term, yet, during the consent process many of the medical, nursing and support staff asked for further clarification. This absence of a single standardised definition hinders the understanding of those undertaking medicines reconciliation, and consequently, makes it difficult to clarify process and the roles and responsibilities of the various HCP within that process. It also influences valid and reliable measurement. Furthermore, depending on the definition used, it could include or exclude certain stages within the process, some of which could be argued as essential, for example, reconciliation within primary care. It is also common in the literature to read ‘medication’ reconciliation in the US, Canada and Australia compared with ‘medicines’ reconciliation in the UK. A standardised, internationally accepted definition is 288 important to ensure consistency in, and management of medicines reconciliation. This study used the IHI definition2 of medicines reconciliation and the NPC’s model (collecting, checking, communicating),80 as a starting point, to carry out both the literature review and the empirical work. This model and definition, however, are limited to the secondary care setting. Any definition of medicines reconciliation must encompass and be applicable to all transitions of care, not just between secondary and primary care; for example the interface between general practice and nursing homes. This study has led to a new definition which encompasses all stages and professions to enable process, roles and responsibilities to be more clearly defined.

‘The retrospective assessment and evaluation of the patient’s recent medicines (e.g. inpatient medication chart or discharge prescription) at specific transition points in care and the resolution of any discrepancies found. The aim is to ensure the information transferred accurately reflects the patient’s current treatment requirements (including drug, dose/strength, form, route, frequency).’

Medicines reconciliation error is also poorly defined, although one could claim that medicines reconciliation error in part is ‘a failure in the medication history taking process’. However, this does not encompass the errors in communication due to poor documentation when medication is altered. A standardised taxonomy or classification system for discrepancy type (apropos definition) is imperative to be able to compare published work, measure progress and benchmark organisations, and ultimately increase quality. The amended taxonomy of errors used within the descriptive prospective study (Chapter 4) and the observational study (Chapter 5) (initially proposed by the Universities of Leeds, Sydney and Wisconsin [David Alldred. Senior Lecturer in Pharmacy Practice, University of Bradford. Personal communication 2011]), encompassed all discrepancies found during data collection. In the main, it was easily applied by the pharmacists collecting the data and would seem a suitable and fitting classification system for clinicians and indeed researchers. However, some of the difficulties found by pharmacists using this proposed taxonomy were understanding the difference between dose, frequency and strength. They also struggled to recognise when a medicines reconciliation discrepancy or

289 error differed from a prescribing error. For example, prescribing the incorrect strength of a medication due to poor medication history taking would be classified as a medicines reconciliation discrepancy, however, prescribing the incorrect strength of a new acute medication would be a prescribing error. The use of the taxonomy was assisted through guidance notes which accompanied the data collection tool, which could be further improved to clarify finer detail. This study was the first time the taxonomy has been applied in practice and has validated its use, enhancing confidence in its appropriateness. Moving forward, the next stage of development would be to test the definition and taxonomy further, initially with a multidisciplinary professional panel followed by pilot implementation in practice.

7.1.1 Defining Medication Reconciliation Error

To be able to define medicines reconciliation and related tasks accurately, as well as measure error and/or discrepancies associated with medicines reconciliation, it is important to define whereabouts in the patient journey medicines reconciliation starts and ends. The term ‘medication reconciliation error’ could be open to interpretation as many use the term medicines reconciliation solely to describe the act of verifying the medication history by a pharmacist. Thus ‘medication reconciliation error’ would then refer to errors made by the pharmacist in verifying the medication history. However, many of the definitions within the literature are much broader than this, also covering medication histories taken by doctors and nurses at admission, plus the reconciliation conducted in general practice post discharge. Therefore within this thesis medicines reconciliation was considered as any retrospective assessment and evaluation of the patient’s medication at transitions (admission, during stay, and discharge) with medication reconciliation error being defined as any error arising from this retrospective assessment and evaluation.

Medicines reconciliation error is independent of prescribing error. Prescribing error is defined as ‘a failure in the decision-making process’. Yet, with medicines reconciliation it is often less to do with the failure in the decision making process, rather the lack of awareness of the medications the patient was taking prior to the transition in care due to failure of systematic approach and failure to utilise and or interpret the sources of information correctly. 290

The term medicines reconciliation itself is problematic due to lack of clear definition and misinterpretation by HCPs. It should therefore either be made redundant or used cautiously as a broad overarching term with the specific stages referred to individually as 1) medication history taking, 2) verification of the medication history, 3) reconciliation at discharge. Consequently, to avoid confusion and to aid categorisation of so-called medicine reconciliation errors and prescribing errors, it would be easier to categorise the error according to the stage in the patient journey where the error arises. This would allow the error at each of the stages to be more easily described and measured as they would describe more specifically where they occur eg verification error, medication history error, writing the IMC.

7.2 Scale and Severity of Medicines Reconciliation Related Error and Discrepancy

Medicines reconciliation continues to be a worldwide problem. Between starting and completing this study more than 200 empirical studies were published on the topic. Most of the studies emphasised high rates of discrepancies with little improvement. Some countries are further on their journey towards reducing error and/or discrepancies associated with medicines reconciliation than others for example the US, Australia and Canada, and are engaged in international campaigns e.g. WHO.268 The UK is also committed to the reduction of medicines reconciliation discrepancies and error and, whilst the UK is not as publically involved in international campaigns, the NHS and NICE & NPSA are continuing to drive the improvement of medicines reconciliation forward. The rates of discrepancies found in the current study are in keeping with those found in other international studies and have been corroborated through both quantitative and qualitative data collection (Chapters 4, 5 and 6); omission remained the highest discrepancy type in all three studies, which is in keeping with the literature.77,109,125,145,166,171-183,201 The similarities in rate of discrepancy and their potential severity with the existing literature suggest the study conducted here has some generalisability and transferability.

There are sufficient studies which explore and determine the rate of medicines reconciliation between primary and secondary care; this is well established. There now needs to be further studies which explore the scale of discrepancies 291 between other interfaces, for example, care homes, outpatient departments, GP referral letters, dental visits644 as well as more rigorously designed studies which look at interventions to reduce error. This need for such studies, reflects an earlier literature which called for an increase in quality improvement and patient safety intervention especially in relation to medicines reconciliation.99,645

The effect of medicines reconciliation error on readmission rates is not well evaluated, nor has it been explored here, within this study, or cost evaluated. However, it is not difficult to see how an improvement in medicines reconciliation could potentially impact on both readmission and costs. The current study identified that there is a potential for some (although few) discrepancies to cause harm. Exactly how many would be difficult to quantify. This makes any potential cost savings due to intervention difficult to calculate. It has only been explored in one previous study; Boockvar et al looked at the actual ADEs occurring from medication discrepancies and determined that of 1350 discrepancies, 65 (4.8%) resulted in ADEs.159 A health economic analysis was not conducted, but would be beneficial to determine potential cost implications.

The frequency of errors or discrepancies is not necessarily the most important measure in patient safety research, rather the frequency of adverse events and severity of harm to the patient. For some events it is important that they never occur e.g. ‘never events‘cc.646 Although not listed in the NPSA Never Events Framework, it could be expected that those medicines reconciliation related ADEs classified as most severe, such as the omission of insulin, should never occur. And, indeed omission of insulin is listed in the NPSA Rapid Response Report, (Reducing harm from omitted and delayed medicines in hospital).619 As stated earlier the severity of discrepancies identified (Chapter 4) is in keeping with those found in previous studies. Difficulties lie with the subjectivity of severity scales and the numerous different scales used within the literature. Further studies exploring the rate of discrepancies and errors which go on to cause tangible or measurable harm are needed to be able to more accurately

cc Never Events are serious, largely preventable, patient safety incidents that should not occur if the available preventative measures have been implemented by healthcare providers. (NPSA, 2011)

292 look at potential cost and cost savings due to intervention (see section 7.8 further research).

7.3 Sources Used to Conduct Medicines Reconciliation

The accuracy of the medication history is reliant on the use and accurate interpretation of relevant and up-to-date sources. HCPs must also recognise the potential limitations each source may have and their reliability. There is a plethora of evidence which demonstrates the accuracy of different sources used during medication history taking and many that illustrate the importance of using more than one source of information. Yet, there is no evidence to suggest that any one source is more accurate than another. With this in mind, there is need to dispel this myth, concluding that each source adds value to an overall picture. Contrary to the NPC guidance,80 it is possible that MAR charts are more accurate than suggested as they are a record of what the patient has actually been taking. HCPs within this study felt they presented a truer reflection, although no studies have actually looked at their accuracy. The quantitative and qualitative data determined within this study both show that often more than one source is not used when taking a medication history, however, the reasons for this are unknown. It is clear from both the current study and published evidence that patient involvement is paramount, despite this only 50% of patients were used as a source of information during their medication history gathering and relevant cues and concerns raised by the patient were sometimes ignored. Patients were not necessarily seen as superfluous, rather the HCPs fixated on the process of completing the task and there was overreliance on the GP record as a source of information. Furthermore patients were not routinely encouraged to bring their medication into hospital, which supports previous findings,531 and when they did they were not routinely used as a source. Patient involvement is discussed further in section 7.6.1.

7.4 Patient Safety and Human Factors

This thesis has been underpinned by patient safety and human factors theory, drawing on the previous work of Reason which describes a person and systems approach.16 This section will provide a general overview of patient safety and

293 human factors and their role in medicines reconciliation. Further detail will be discussed in subsequent sections.

There has been international emphasis over the last decade on patient safety. With this has come an increased interest in the role human factors plays in healthcare and a recognition of their importance.11 Since then, the emphasis on understanding human factors and human error in healthcare has grown and is becoming more common place. Over a similar time frame, the drive to decrease medication safety incidents has turned focus to improving communication about medication at transition points, predominantly admission and discharge.276,647-650 Interest in the scale of issues plus interventions to improve medicines reconciliation have been sought; more recently being embedded into the medicines optimisation agenda in the UK.45

A human factors informed system should identify and remove system hazards from the design. Processes should be designed to diminish the impact of intrinsic difficulties which hinder the clinicians’ ability to perform their work and deliver safe care. It was acknowledged in Chapter 2 that the study of human factors in relation to medicines reconciliation has not previously been well explored and that any medicines reconciliation processes in place were certainly not designed with human factors in mind. This could create opportunity for error and compromise patient safety. Several dominant factors were seen to potentially contribute to medicines reconciliation-related errors and/or discrepancies. These included: individual factors, team factors, patient factors, education and training, IT, communication factors and cultural factors, the most significant of which appeared to be communication or lack thereof, especially written documentation within and between organisations.

By using human factors the YCFF and communication models the study has provided a fascinating insight into the medicines reconciliation process. This suggests other frameworks of analysis could be used to explore complex processes in medicines management. Carayon et al demonstrated the human factors link between system design and patient safety (SEIPS). Their work emphasises that a system not designed according to human factors principles can create opportunities for errors and hazards.651,652 SEIPS considers system in the light of human factors and the Donebedian model; structure, process and 294 outcomes.653 This has been applied to healthcare previously to improve patient safety,651,652,654 There are also plans to apply this to medicines reconciliation as part of the US based MARQUIS (Medication Reconciliation Quality Improvement Study); its aim, to operationalise best practices for inpatient medicines reconciliation, test their effect, and understand any barriers and facilitators of successful implementation.143

Despite there not being any previous literature which looks at human factors in relation to the medicines reconciliation process, Cohen in his book Medication Errors has described potential ‘at risk’ behaviours in medication processes which could contribute to error, all of which were discovered in this current study and are grounded in a human factors approach (see table 7.1).655

7.5 Process

7.5.1 Overview

Systems, processes and procedures

Medicines reconciliation forms one small, but extremely important, part of the overall healthcare system. Van Cott stated that healthcare is composed of a large set of interacting systems that are connected by intricate networks of individuals, teams, procedures, regulations, communication, equipment and devices that function with diffused management in a changing and fluid environment.656 Medicines reconciliation is an organisation wide process that crosses departments, professions and specialties as well as the primary- secondary care interface. As demonstrated in the literature review, previous studies of medicines reconciliation have studied isolated parts of the process rather than the process as a whole. This study has observed the process in its entirety and provided a further critical insight into how a range of healthcare systems co-exist with medicines reconciliation. This together with an understanding of its relationship to other external systems and people can provide a substantive contribution to resolving the issues associated with medicines reconciliation.

295

Table 7.1 At risk behaviours in medication processes relating to medicines reconciliation (adapted from Cohen p616)655 Patient information e.g. not waking the patient for assessments all medications Drug information e.g. writing incomplete discharge instructions, prescribing, dispensing, or administering medications without complete knowledge of the medication

Communication e.g. rushed communication, failure to speak up when there is concern, use of error-prone abbreviations (e.g. u instead of units for insulin doses), not reading back spoken orders, providing incomplete orders e.g. Incomplete details, illegible handwriting Environment staffing e.g. failure to supervise, inadequate staff, managing multiple priorities Patient education e.g. Prescribing, administration of or dispensing medications without educating the patient, disregarding the patients or ‘caregivers’ concerns about their medications appearance, reactions or effects, or other worries, discharging patients without proper education about their medications to take home Staff education e.g. Lack of a structured and ongoing staff competency programme related to medication use. Culture e.g. sacrificing safety for timeliness, failure to report and share error, blame culture Double check e.g. over-confidence in others work Teamwork e.g. reluctance to consult others or ask for help, lack of responsiveness to HCP or patient requests Technology e.g. failure to fully engage available technology

Within the system are a number of defined processes. A process can be defined as:

‘A sequence of interdependent and linked procedures which, at every stage, consume one or more resources (employee time, energy, 296

machines, money) to convert inputs (data, material, parts, etc.) into outputs. These outputs then serve as inputs for the next stage until a known goal or end result is reached.’657

Previously medicines reconciliation has been described as a system with many inter-related proceduresdd that make up one large scale process, which of course, crosses the primary secondary care interface (amongst others). The ‘output’ described in the definition could be a discharge summary from secondary care which is inputted once it reaches the general medical practice.

Healthcare professionals role in process

Medicines reconciliation involves different HCPs at different stages within the process to produce BPMH. HCP behaviours are a product of the systems and processes in which they work. Variation and inconsistencies in approach need to be addressed. This is difficult when there is no ‘gold standard’ approach to medication history taking and limited standards for medicines reconciliation.

There is a need to clearly define roles and responsibilities within the medicines reconciliation process. There is disagreement about who should take the lead and responsibility for embedding patient safety principles into medicines reconciliation processes within the US.339,340 This is clearer in the UK, as the NPSA has suggested pharmacy staff, although further work to identify HCPs’ perceptions of the roles they should play in the medicines reconciliation process would be beneficial to start to overcome any barriers which exist, for example, effective collaboration, teamwork and seamless care. The WHO suggests that healthcare organisations should assign clear roles and responsibilities to HCPs for all steps in the medicines reconciliation process to ensure accountability,1 however, the WHO do not make suggestion as to what these roles and responsibilities should be or to whom they should be assigned. It might be pertinent, rather than to assign a specific role to a specific group of HCPs e.g. initial clerking medication history to doctors, to provide a list of competencies that should be achieved by anyone who is to undertake the role. This has been achieved in other areas of healthcare in Australia658 and the UK.659 This may dd Procedure - A series of steps taken to accomplish an end.

297 take more effort than assigning the appropriate task to a appropriate HCP, especially as pharmacists already display many of the relevant competencies needed to accurately undertake a medication history (see section 7.4.1 At admission - Pharmacy staff).39,49,66,69,71,74,118,122,124,162,208,237,239,277,278,300-321,346-348

As discussed in Chapter 2 the patient’s journey through hospital includes three distinct stages: admission, during stay and discharge.160,289 The following sub- sections of this chapter will re-visit the process at each of these stages in the light of the study findings, plus post-discharge, within primary care, considering the role of each individual HCP at each stage.

7.5.2 At admission

Doctors

Following the patient journey, the medicines reconciliation process starts at admission to hospital. Traditionally, within the UK, it is the doctor’s role to clerk the patient and take the initial medication history. Yet, as demonstrated throughout the study and supported by existing literature, this leads to a high number of errors and discrepancies in the patient’s medication history. 39,49,66,69,71,74,118,122,124,162,208,237,239,277,278,300-321,346-348 There is a current lack of evidence to suggest which interventions would improve doctors’ performance during medication history taking. Various defences could be put in place to improve the accuracy of medication history at admission, for example, increased standardisation of documentation, training and education and increased inclusion of the topic in the undergraduate curriculum, which have already been shown to work.93,94,96,328,329,338 Training and education also has its limitations in comparison to system change. These are discussed further in section 7.6.5. The identification of appropriate interventions are needed, mindful that education and training is not an especially effective means of improving patient safety when used alone.660,661 Alternatively, it would be worth reviewing whether doctors are the most appropriate people to undertake the task and whether pharmacy staff would be better placed given their skill set. The observation in Chapter 5 showed that Doctors placed little emphasis on the medication during clerking and were more likely to use a systems-based approach to eliciting the patients medication history during the medical history

298 taking compared with the pharmacist who would ask questions specifically relating to the patients medication using a wider variety of sources. This is similar to previous findings which determined that clerking doctors spend on average 45 minutes on the admission interview with approximately 0-5minutes on the medication.662 Vashitz in two separate studies looked at the way in which doctors cognitively reconcile medication. They found that clinicans commonly arrange clerking information into two groups conditions and medicines assigning mental order in terms of systems. This goes some way to explaining the observations seen, although it is unclear whether this is a result of the way they have been taught or whether it is inherent.663,664

Pharmacy staff

Pharmacy staff acted as a ‘safety net’ identifying various omissions and discrepancies in both Chapters 4 and 5. The evidence base also suggests their increased accuracy at eliciting medication histories in comparison to doctors.39,49,66,69,71,74,118,122,124,162,208,237,239,277,278,300-321,346-348 This increased accuracy poses the question whether medication history should be their role. It was also identified by Kwan et al in a systematic review that pharmacists contributed the most to interventions which improved medicines reconciliation.665 In the current study, pharmacy staff mainly worked ‘office hours’, 9.00am to 5.00pm, on weekdays with lack of weekend cover. It could be argued that pharmacy cover should be increased if pharmacists are acting as a ‘safety net’ to verify medication histories. This, however, has staffing and cost implications raising the question as to whether priority might be accorded to specific points in the hospital system, although there may be some justification to consider this on admissions wards. The findings within this study and the personal view of the researcher is that pharmacists have many of the appropriate skills to undertake tasks associated with medicines management, however, where this proves difficult due to time constraints or cost implications other professionals involved should be appropriately competency assessed.

Through observation the quality of medication histories elicited by staff in A&E was found to be poor, mainly due to their focus on stabilising the patient rather than providing maintenance care. As previously discussed, for many patients the patient journey through hospital starts in A&E therefore it could be argued 299 that it is important to obtain accurate information from the start, moving the verification of medication histories by pharmacists into the A&E setting, shown to be effective by Mills and McGuffie who took the medication histories and prescribed the medication in this setting.66

The American Society of Hospital Pharmacists (ASHP) argues that pharmacists should take a leadership role in improving medicines reconciliation, acting as advocates and medicines experts to provide competence based training, inform systems and develop evidence based policy and procedure.340 They should identify barriers that prevent effective medicines reconciliation and guide workflow development, establishing roles and responsibilities.340 Studies have shown decreased error where pharmacists have prescribed existing medication at admission both in A&E66 and surgical pre-admission clinics.124,321 It would also be more efficient as currently the medication history is being taken twice; once by the doctor and again by a member of pharmacy staff, (although this does act as a ‘safety net’.) There is also evidence that non-medical prescribing by both nurses and pharmacists is clinically appropriate.354 However, the hours worked by pharmacy staff currently only allow for a proportion of medication histories to be verified. This has the potential to lead to error in those IMCs not reviewed, especially as pharmacists are acting as a ‘safety net’ to identify any unintended discrepancies. If pharmacists were to review every IMC and take the initial medication history, pharmacist hours would need to be significantly increased and the cost-effectiveness of this evaluated (see section 7.8 Further Research).

Other staff

This study predominantly looked at the role of doctors, pharmacy staff and nurses in the secondary care setting, administrative staff, practice pharmacists and GPs in primary care, and to a far lesser extent the community pharmacist and ambulance staff. Whilst it is clear that both community pharmacy staff and ambulance staff have a role in assuring accurate medicines information at admission, further integration is needed at transitions in care, which may mean developing studies to determine the added benefit that both groups can provide at admission. Such work should identify the barriers and enablers specific to

300 ambulance staff bringing patient’s own medication into hospital, and examine the accuracy and purpose of the medication histories they take.

7.5.3 During stay

During the patient’s stay in hospital they come into contact with many HCPs and are involved in many systems, processes and procedures. In terms of medicines reconciliation, the main associated processes are the prescribing, alteration and transcribing of medication by prescribers (mainly doctors), the administration of medication by nurses and the handover of patients.

Nursing staff

There is a lack of evidence recognising the role that nursing staff play in medicines reconciliation. Good practice guidance produced by NHS Greater Glasgow and Clyde suggests that nurses should maintain an overview of whether patients on their ward have had their medicines reconciled at admission, prompting relevant HCPs to take action as required; administer medication as directed on the admission chart; incorporate medicines reconciliation into handover procedures when patients transfer care and at discharge and where the prescribers are nurses ensure that prescriptions are accurately and legibly recorded.359 Based on the findings in this current study, nurses have a greater role to play as an educator and administrator of medicines; identifying omissions and other discrepancies during the administration process, ensuring that they pick up on any cues that the IMC may be incorrect. This stresses the importance of their role within the medicines reconciliation process. They are likewise well placed to identify errors and discrepancies on the medication chart through routine observation and listening to patients, especially if the patient queries medication. This has not been sufficiently acknowledged in previous literature or in practice.

Within Chapter 5, all handovers were found to be inconsistent, with lack of structure as to which information should be conveyed about medication at handovers both on the wards and when the patient is transferring care e.g. to another ward or intermediate care. Whilst handover is usually associated with nursing staff, it could be further standardised within all professions.

301

Doctors

Doctors are responsible during the patient stay for altering medication and transcribing IMCs. It is recognised that transcribing, as part of medicines reconciliation, carries a substantial risk of errors,52,78,177,363 with a suggested overall error rate of 1% in the UK associated with rewriting charts and writing discharge prescriptions.52 Transcribing errors may go unnoticed, and then be transferred to the next care setting. Transcribing was rarely seen within this current study, nor its impact on accuracy of information about medication. Additional work in the area would be useful to determine the accuracy of transcribing and the most appropriate HCP to undertake transcribing – although the introduction of electronic systems will eliminate the need for transcribing as a new IMC will not be required and information on e-prescribing systems will be automatically transferred on to the electronic discharge summary.

During the observation reported in Chapter 5, the majority of changes to medication occurred on the consultant ward round. These were frequently unclear and were not communicated (to other HCPs or the patient). There is little in the literature which looks at the alteration of medication within secondary or primary care. One US study however highlighted that the family doctor failed to communicate critical elements of medication use to the patient.666 There is a lack of evidence which looks at how well information about stopping, starting or changing medication is communicated either in the notes or verbal to HCPs and to patients. This needs studying in further detail to quantify the issue and to work out any potential barriers and facilitators to improving practice. This could be done using cognitive processing/think aloud techniques to be able to understand the information intended by the sender and the information elicited by the receiver of information.667

7.5.4 At discharge

Doctors, pharmacists and nurses currently have very different roles at discharge. It is common practice (and supported by the observations here) that the doctor writes the discharge summary; the pharmacy staff check the accuracy, then they ensure a supply of medication; then nurses (and occasionally pharmacists) counsel the patient. There is a lack of evidence on

302 the accuracy and ability of nurses in counselling patients and whether they are best placed to provide this role. As the number of non-medical prescribers increases there is the potential for the transcription from the IMC to the discharge summary to change; in some organisations this is already conducted by a pharmacist.362 Moreover, nurse and pharmacist prescribers have been shown, in general, to make clinically appropriate prescribing decisions.354 As previously mentioned, the increased introduction of IT could make transcription a redundant process as electronic means can facilitate the transfer of information from admission through to discharge (see section 7.6.2 Role of Information Technology (IT)).

7.5.5 In primary care

Chapter 6 presented the results of a small scale observational study to explore medicines reconciliation within primary care and to test the use of an observation schedule for the processing of discharge information. The overall model used to process information is the same within all practices and mirrors the information processing model suggested by Crowe et al.429 There is wide variation, however, in the finer detail of the processes used within each GP setting, some, but not all, of this appears to be due to the difference in demographics of each of the practices, for example, the larger practices may have a person allocated to undertake each stage of processing discharge information, whereas others may have staff which undertake several roles. It is important that a clinician reviews and interprets the information received in a timely way. This was achieved in several of the practices through a ‘rota’ system for the GPs. The GP is the primary record holder, therefore, accuracy is paramount as many rely on this as a source of information, especially at admission to hospital and with the increased drive to directly access the detailed care record and summary care records from the hospital setting. A different model of reconciliation is needed in primary care to reflect the differences in process and acknowledge that the patient journey does not stop once the patient is discharged. There is a great need to improve the quality of information emerging from secondary care to decrease the number of assumptions that are being made in order to make sense of the information

303 received. There is also need to improve the quality of data in the primary care EHR and to ensure data is contemporary.

Community pharmacy

Community pharmacies can provide medication counselling, support and reduce discrepancies at discharge.316,416,417 This is dependent on the receipt of relevant discharge information for their patients; often they are unaware the patient has been in hospital.360,421 The communication observed between community pharmacies and both secondary and primary care within this study was ad-hoc and tended to be where the patient was using an MDS. This is supported by previous findings, albeit in the same health economy.360 There is greater potential for the role of community pharmacists within medicines reconciliation at both admission and discharge than is currently happening.107,327 Some hospitals in England are encouraging patients to visit their community pharmacy for a Medicines Use Review (MUR) post-discharge, despite their being little supporting evidence.435

7.5.6 Closing the loop

Once the patient reaches primary care this may not be the end of their journey. Patients may be readmitted, for example, the renal and elderly patients seen within the study. Errors which have previously occurred may be compounded at re-admission with HCPs not questioning what appears to be well-established long term medication. And undeniably, as seen in Chapter 5, the discrepancy may be the cause of their readmission.

7.6 Strategies for Improvement

The quantitative and qualitative data highlighted variation in ways of working, omission and mis-communication of information with poor patient involvement, all which had the potential to lead to medication error or discrepancy. The data also showed that these contributory factors could be counteracted through improved patient involvement, increased use of IT, standardisation, optimising communication, training and education and improving organisational culture. These will be discussed in turn below.

304

7.6.1 Patient involvement in medicines reconciliation

The patient is the constant throughout the whole medicines reconciliation process and therefore can add value and information which cannot be gained elsewhere, for example, which medications they are currently taking or complying with.509,511-513 The results of Chapters 4 and 5 demonstrate that patients were not always involved in the consultation about the medication they were taking prior to admission (approximately 50% of consultations); their cues and concerns frequently ignored. This lack of involvement was identified as a contributory factor to medication discrepancy and error. Conversely, involving the patient created a defence. When they were involved the patients provided supplementary information at admission and were able to provide information to primary care on any changes to medication regimens at discharge. They can be a useful resource where information needs to be confirmed as seen in Chapters 5 and 6. Patients want to be involved in their care and can increasingly play a central role in patient safety.502,503 Patients should always be engaged even if they are limited in the information which they can provide but still should not be relied on as the sole source as there is potential for even the most competent of patients to accidently omit medication due to forgetfulness or their acute state.

Davis et al found that patients were keen to participate in safety related behaviours especially if encouraged by the doctor, although they were more willing to ask general questions about their care, rather than more challenging questions.668 Davis et al’s study, however, only questioned English speaking patients. Potentially different findings may have been found in non-English speaking patients or for patients who were more vulnerable. Davis et al suggests the use of social cognition models such as the health-belief model and the theory of planned behaviour model to define the key predictors of patient involvement in safety. They also suggest observational studies of patient involvement to gain an understanding of different factors that affect patient involvement such as socio-demographic factors and physical, social and environmental factors and the engagement of HCPs. This would be pertinent and relevant to medicines reconciliation.

305

The WHO policy specifies that patients and their families must be involved in the medicines reconciliation process to optimise treatment and improve safety. A competent patient is in the best position to be aware of all the medications prescribed by multiple caregivers and can also provide information on exactly how they use their medicines, information which cannot be obtained from other sources. Involving the patient can reveal a significant amount of extra information not provided by other sources.78,121 Within this study, the extent to which the patient was involved in medicines reconciliation associated consultations depended on the HCP and the patient; specifically their ability to communicate; their knowledge about their medication and their confidence to speak out. The HCPs tended to be process focused and be unaware of any nuances from the patient which may explicitly or implicitly suggest an intervention was needed. The ability and willingness of the patient to be involved in their healthcare varies from patient to patient. There can be difficulties obtaining accurate information from patients about their medication, which may be caused by their acute condition, sensory/cognitive impairment, lack of support or language barriers.6,76 This affects both their involvement in obtaining an accurate medication history and also their ability to participate in patient safety.669 Irrespective, the HCP should try to engage with every patient and facilitate their involvement, recognising their limitations and vulnerabilities.670

Over the last decade, there has been a national and international drive2,501 to empower patients, promote patient involvement and make them central to healthcare, ultimately to improve care and patient safety.502,503,671,672 Building a safer NHS for patients: Implementing an Organisation with a Memory recognised the need for patients to be involved in their care to achieve patient safety.27 Initiation of the NPSA, further promoted patient and public involvement, introducing numerous campaigns and guidelines.673 Patients can be involved in promoting safer healthcare in various ways, for example, being vigilant and speaking up to voice any concerns, encouraging involvement in care or identifying adverse drug ADRs;669 acting as an ‘extra barrier’ or ‘safety buffer’.674 An appropriate level of engagement is however, important, so as not to damage the patient’s relationship with the HCP or erode trust.670,675

306

Several generalised approaches have shown to encourage patient involvement in their own safety. These include inviting the patient to engage,675,676 providing appropriate amounts of relevant information677 and being honest and open.669 More specifically, in terms of medicines reconciliation, widespread engagement could be more effectively facilitated through the employment of an interpreter, where needed, or returning when the patient or their carer is present, rather than what was sometimes observed; an assumption regarding the patient’s ability or willingness to participate. Patients can also be engaged through self- administration of medication and counselling. Counselled patients have been shown to have been increased knowledge and to display fewer errors than uncounselled patients in recalling information about their medication82,383,395,424,537 which is imperative if patients are to be used as a source of information within the medicines reconciliation process. They must also be encouraged to bring their medication into hospital with them.

HCPs must recognise it may be the patient’s first visit to hospital and whilst the HCP is familiar with the environment the patient may not be. The patient may have varying concerns which differ from those of the HCP or those that the HCP assumes of the patient. Patients may also feel so unwell that interaction with the HCP is not of interest.82 Ultimately, HCPs must be further educated on the importance and potential value of patient involvement, making patients central to communication and process design. Organisations should also take responsibility for making the patient central to healthcare including reference to them in all policies and standard operating procedures (SOPs) and indeed potentially involving them in the design of services.678 Further work looking at the specific role of the patient in the medicines reconciliation process and their ability to be involved is needed. This could potentially be developed through an experience-based codesign with the patient.678

7.6.2 Role of information technology (IT) in medicines reconciliation

Recent advances in technology provide opportunity to improve the medicines reconciliation process and add safeguards to medicines reconciliation processes. IT has a significant role to play in the improvement of safety around medicines reconciliation processes. There were several contributory factors identified which would be eliminated through the use of appropriate IT. These 307 include legibility, the completeness of data and transcribing error. Ideally an integrated system in which the medication history is entered (potentially fed in from the primary care EHR), then feeds into the electronic prescribing system and subsequently the electronic discharge. This would contribute toward seamless care. Although, where electronic systems transfer the information from one stage to the next there is potential for any errors/discrepancies to be carried throughout the patient journey if generated or not identified at admission.

Even though there is increasing evidence that the introduction of IT can reduce many of the errors and discrepancies from inappropriate medicines reconciliation (as well as assist other patient safety issues), widespread use of IT-based interventions within hospitals is low and often limited to specific areas.385 IT is slowly being embedded into secondary care at varying rates through the implementation of e-discharge, EHRs, and e-prescribing.384,385 Although this can be ostensibly rather frustrating, the current situation provides an increased opportunity to improve the medicines reconciliation process. As the use of e-discharge increases so does its potential for being used as an accurate source of medication information at re-admission, provided that the medication was reconciled on the first admission. Data completeness is important for patient safety475 and data accuracy still relies on the accuracy of the primary care EHR and the completeness of discharge summaries. There is still the need for a human to interpret the information. Once implemented, IT can ensure information is uniformly available and can address active and latent failures,475 reducing error for example by using forcing functions to ensure fields are not left blank.478

The Health and Social Care Information Centre (HSCIC), formerly Connecting for Health (CfH), have published guidance in an attempt to standardise the quality of information within medical records and communications when patients are admitted to and discharged from hospital.679 They emphasise that discharge planning information should be started at the time of admission. They describe the information which should be incorporated into electronic discharge communication suggesting that information falls into three categories mandatory, mandatory with defaults or optional, employing forcing functions for

308 those which are mandatory. They go on to specify that a list of medications and any changes should be mandatory and specific medication recommendations mandatory with a default. Forcing functions are part of human factors engineering to force constraints on systems and processes.

Despite the positives introduced through IT, it should also be acknowledged that the introduction of IT systems brings with it other challenges, new ways of working and changes to systems and processes and new types of errors.489 Within this study, variance in the completion of discharge summaries between HCPs was still observed. The electronic discharge improved the amount of information provided, however, it did not eliminate the variance in quantity and quality between prescribers. It would appear that there were mixed feelings and lack of understanding by some prescribers of the potential added safety of electronic discharge versus paper discharge. Workarounds were employed, for example, amending the electronic discharge summary by hand which led to inaccuracies in the electronic system. Also, the sharing of logon details and passwords between prescribers following lock-out of the system was seen. Challenges have previously been seen elsewhere through the introduction of e- prescribing, leading to even less engagement with the patient due to the difficulties of transporting a computer terminal around the ward.680 In a five year analysis of technology-related medication errors 17.1% technology related. 98.1% of which were socio-technical errors i.e. due to a human interaction with technology, rule violation, slips, mistakes and lapses being the highest cause.681

7.6.3 Standardisation of medicines reconciliation-related processes

Bates suggested, in 2000, that healthcare, compared with the manufacturing industry, had little standardisation.682 Healthcare processes have grown and evolved over the years, adapting to circumstances and new knowledge and technologies rather than being designed with patient safety in mind. They are often complex, with no one person responsible for the entire system and simple mapping of the processes can be a difficult task. Various studies show that when care is divergent, it can impact on clinical outcomes, potentially leading to patient harm.25,683-685 Standardisation of tasks can reduce the complexity of working, and improve the receipt of correct medication.685 It is being increasingly recognised as a way of improving patient safety.686 309

Standardisation, or the routine application of an evidence based protocols are effective for reducing human error, and thus harm, in given situations.687 It reduces the reliance on memory and allows consistency in data collecting. It also allows staff to work across multiple areas (as seen in Hospital B), resulting in increased efficiency and potentially decreased costs.687,688 Standardisation of documentation or proformas have been shown to aid the medication history taking process within the UK101,442 and Canada,292 providing appropriate boxes and prompts to facilitate information gathering by ensuring consistency and completeness of information.401

Standardisation was observed during clerking and medication history verification as standardised documentation together with communication proformas were used, also electronic discharge. Standardisation was also observed during nursing handover through the use of SBAR, previously shown to improve communication.683,684 In primary care communication documentation was also used between one general medical practice and community pharmacies to notify them of changes to MDS. A higher use of EHR was seen in primary care which also contributed to increased population of data, however, like electronic discharge the complete population of fields does not always equate to the accurate completion of fields.

Despite the growing body of evidence and the success of some standardised parts of the process there is still room for further standardisation of policy, practice and procedure within medicines reconciliation to reduce the wide variation in practice within and between organisations. This could potentially be produced nationally through the implementation of policy standards and toolkit development to aid implementation. There are various toolkits which already exist which address current issues to varying degrees. These could be adopted and adapted to fit the UK.2,155,156,268 There is also the potential to increases standardisation through the implementation of IT e.g. through the use of mandatory fields and forcing functions. However, just because a box is completed it does not necessarily mean that the content is of quality.

310

7.6.4 Optimising communication

Communication is defined as:

‘the act or an instance of communicating; the imparting or exchange of information, ideas, or feelings’57

Communication failures are an extremely common cause of avoidable harm.689 The JCAHO in the US has reported that the primary cause of over 70% of serious injury or deaths is failure to communicate.689 Whilst it is unclear whether incidents have been analysed in a similar way in the UK, improving communication features widely in many of the NPSA alerts6,690-693 and research illustrating issues with communication about medication at transition points in care, within the UK, dates back over 25 years.40-42 This includes delayed untimely communication and provision of inaccurate information.

As previously discussed communication is one of two contributory factors within the YCFF (the other being ‘safety culture’), that intersects all other factors (see figure 5.1). However, the YCFF itself does not provide a detailed exposition of communication or safety culture because the systematic review on which it is based did not generate evidence to underpin this. This study demonstrates the complexities and interdependencies between coordination of care and communication about medication, including the fact that communication is both verbal and written. Both verbal and written communication occur throughout the medicines reconciliation process including the documentation of the medication history, prescribing, the query of discrepancies, instructions to alter medication, production of the discharge summary and alteration of the primary care record post discharge. Throughout this study, in Chapters 4, 5 and 6 communication, or lack thereof, has been seen as a continuous theme which has led to discrepancies and/or error. This was due to poor communication between HCPs or the HCP and the patient (both within and between primary and secondary care).

Simplistic models of communication describe the movement of information between a sender and receiver.613 Assumption was often made both in primary and secondary care, especially when interpreting documented information e.g. in the notes or discharge summary. This is a direct failure in communication in 311 terms of both the sender and receiver; the sender perhaps not providing sufficient information, e.g. in the patient’s medical notes and the receiver failing to clarify e.g. ambiguous notes. Some communication models describe communication as a cyclical process.614,694-696 Feedback completes the communication loop and is described as the most crucial element provided by the receiver to the sender.696,697 Feedback ensures the receiver has received the message and interpreted the meaning of the message in the desired way. Without feedback, the process of communication is not complete. This is more easily completed with verbal communication and ‘closing the communication loop’ has been shown to improve patient safety.694-696 it is more difficult to ‘close the loop’ with written information, especially discharge summaries and documentation in notes as often the ‘receiver’ is unable to clarify with the ‘sender’ when needed.

Effective communication involves a shared understanding and mental model of a situation. Leonard et al describe this as ‘getting everyone in the same movie’, which is potentially where efforts should be focused i.e. a shared understanding of medicines reconciliation.698 Reasons why this does not always occur can be best described using Gerbner’s model of communication.699 He emphasises the dynamic nature of communication and factors which affect its reliability. In essence, communication is the relay of an event (E) which has been perceived (E1) by a person (M) (see figure 7.1). A person is unable to perceive the entire event and thus selects and filters information to communicate onwards (content) using a certain method (form). The content relayed is then perceived by the person receiving the information, again filtering for content in a similar way. This could explain some of the mis-interpretation of information by HCPs and the failure to pick up on cues from the patient. Also some of the failure to provide sufficient content if the HCP assumed they had explained what had happened appropriately or that information provided had been received effectively.

312

Figure 7.1 Gerbner’s General Model699

The perception and communication can be affected by human factors including the attitude and feelings of the person at the time of communicating or receiving communication. This would be difficult to observe and could only be elicited through interviews or focus group. The contribution of human factors to communication failure is well recognised.608,689,700

Using Berlo’s communication model (discussed earlier in Chapter 5), communication is multifaceted. All components must be addressed to communicate effectively including the ‘source, message, channel and receiver’ (see figure 5.2).615,616 This includes effective communication skills of both the ‘sender’ and ‘receiver’ and a common culture, knowledge and attitude to be able to explain and interpret the message. The message must be structured and 313 have appropriate content and be communicated by the correct means. Similarly Lingard et al suggest communication fails for different reasons which could be used to understand some of the communication failures observed within the study.609 These include occasion failures; content failures; audience failures and purpose failures. For example, an occasion failure may be the lack of or delayed verification of the medication history by pharmacy staff; content failure, the failure to include relevant details in the notes e.g. the form or strength of medication; audience failures, the failure to communicate to the correct HCP e.g. not notifying the district nurse that they need to administer a medication; purpose failure, a discrepancy remaining unresolved following a communication.

Applying these communication models to medicines reconciliation, it is important that HCPs have a common understanding and knowledge of a) the medicines reconciliation process (including definition), b) the ways in which communication can fail and c) the information which needs communicating. This would ensure that everyone was ‘in the same movie’. This can be achieved, in part, through standardisation such as SBAR and use of proformas for documentation to prompt the HCP, as discussed previously in section 7.8.1. Standardisation is not the only strategy to improve communication. Training and education plus experiential learning may add benefit. This was seen in Chapter 5 when the GP registrars completed the discharge summaries more fully after seeing things from the ‘other side’ i.e. in primary care. It could also be achieved through the integration of the COT Criteria and Essential communication skills for a medication history interview to improve communication skills associated with medicines reconciliation into national standards and competency assessments within organisations.610,611

7.6.5 Training and education

Internationally, successful educational interventions have been recognised to improve the accuracy, knowledge and skills of medication history taking amongst HCPs, for example, curricula which specifically focus on medicines reconciliation at transition in care and medication history taking.93,96,328,329 Whilst these interventions were successful, there is often a misperception that training and education is a must to improve performance and increase safety. 314

Whilst it is important, it has been suggested that training and education are generally a weak safety intervention.21 This is demonstrated in a study by Peyton et al showed that although educational intervention of doctors and nurses increased the accuracy of medicines reconciliation, it only improved from 14.4% to 18.9%.118 Russ et al provide an overview of when training may or may not be appropriate to improve patient safety (see table 7.2).21 Applied to medicines reconciliation this means that training should be undertaken initially when first introduced to the concept, however, where the HCP continues to conduct a task incorrectly a different type of intervention is probably required.

In Chapter 5, the concept of conscious competence is discussed. In order to be consciously competent, staff must have an awareness of their own performance. Gallwey specifies that to achieve optimum performance the conscious, critical self and the performing self must be recognised without passing judgement.701 This is difficult to teach and may in part be solved by encouraging HCPs to be reflexive. Stevick suggests that whilst a student’s conscious self is developing, the teacher can ‘lend themselves’ as a conscious self, serving as a model for how the student’s conscious self may eventually act.702 This makes the idea of leaders, role models and champions for medicines reconciliation a must within any given organisation. One previous study has explored the role of champions within the medicines reconciliation process, demonstrating successful sustained change which led to increased compliance with new processes.703 It created engagement and ‘buy-in’ rather than feeling that change was being imposed.

Understanding the conscious and performing self ultimately requires experiential learning. Some guidance and instruction is required, however, what is more important is the understanding of how to apply the guideline. This is supported by Walton and Elliot who suggest that quality and safety is best undertaken in the workplace rather than the lecture theatres.661 Over the last few years there has been a shift from didactic teaching to competence based education.704,705 The teaching of communication skills are embedded within undergraduate programmes, however HCPs must be able to apply these principles to different scenarios, including medicines reconciliation. This could be assessed through the consultation frameworks used within Chapter 5.

315

Table 7.2 Overview of when training may or may not be appropriate as a human factor approach to patient safety21

Training is likely an appropriate human Training is likely an inappropriate human factors approach to patient safety if… factors approach to patient safety if… A. The goal is for individuals to become familiar A. The goal is for individuals to stop using with new technologies, tools or devices to learn technologies, tools or devices ‘in the wrong about the available options and functions (e.g., way’. (This is described as the ‘bad apple’ training a physician when s/he is first introduced fallacy. to an electronic health record; training when first learning how to use laparoscopic tools). Training should include knowledge about strengths and limitations of specific technologies. B. It allows individuals to develop and test new B. It is an attempt to change innate human techniques or practice evidence-based characteristics or imperfections (e.g., staff techniques in a safe, low risk environment (e.g., meeting to ‘be more vigilant’ unlikely to lead simulation of operating room to practice a team to sustainable safety improvements. communication technique that has been demonstrated to improve situational awareness.) C. It provides a mechanism for individuals to C. It is intended to address a type of error gain experience with specialised techniques that that is occurring across multiple people. involve sensorimotor skills (e.g., performing (This indicates the system design does not surgeries and catheter insertions with match human characteristics and that supervision or in a simulated environment). system changes, not training, are needed.) D. It is used to instantiate knowledge in realistic D. Individuals have been previously trained scenarios, such as to practice or test about the safety issue(s) and the problem procedures for emergency situations (e.g., rapid persists. (Additional training is unlikely to be response). effective. The phenomenon above indicates there is an issue with other system components.) E. Other system components are considered E. Training is the only safety intervention or first, redesigned, and addressed using human the primary intervention used, especially factors expertise and principles and no other when other system components have not system changes can possibly be made. been carefully considered and modified first.

316

Walton and Elliot suggest a shift in medical education to that which trains HCPs on the complexity and interdependence of systems (systems theory) so that they have a better chance of making sense of their workplace and the potential reasons why things go wrong thus preventing error. By understanding the nature of errors and application of quality improvement error can be reduced. In 2005, the Australian Council for Safety and Quality in Health Care published the National Patient Safety Education Framework (NPSEF).706 It identifies competencies which HCPs need to deliver safe healthcare and contains 7 domains including communication, the management of ADEs and near misses, working safely and continual learning; all of which are applicable to medicines reconciliation.

In summary, there needs to be a balance between the level of knowledge and skills gained as an undergraduate and those learned in practice in order to allow the HCP to be reflexive and apply patient safety, communication and medicines reconciliation principles in their practice. Competence and educational frameworks for HCP in both patient and medication safety supported by regulatory and professional bodies, similar to the NPSEF in Australia, could assist the standardisation of knowledge and skills throughout undergraduate curriculum. These could be developed based on existing literature, together with the findings from this study to try to improve the standard of communication at transitions and the reconciliation of medicines (see also section 7.8 Further Research).

7.6.6 Improving organisational safety culture

Safety culture is one aspect of the wider culture of an organisation.

‘The safety culture of an organisation is the product of individual and group values, attitudes, perceptions, competencies, and patterns of behaviour that determine the commitment to, and the style and proficiency of, an organisation’s health and safety management’ (HSC, 1993)707.

‘Organisations with a positive safety culture are characterised by communications founded on mutual trust, by shared perceptions of the

317

importance of safety and by confidence in the efficacy of preventive measures’ (HSC, 1993)707

Medicines reconciliation is a complex area of high risk, which although not always demonstrated, could lead to serious consequences. Recognising this, it is important that organisations develop a culture of safety such that the processes and workforce are focused on a clear goal.11 This was only seen in one of the secondary care organisations studied. To err is Human and Building a safer NHS for patients were clear from the outset that for healthcare organisations to improve safety they must provide clear leadership, make safety and risk everyone’s concern whilst recognising human limits and ability.11,27 This includes designing processes with safety in mind, avoiding reliance on memory, using forcing functions where appropriate, simplifying and standardising where possible, promoting effective team functioning, including the patient in design of care processes, improving access to information and creating a learning environment where everyone learns form error. This was seen in Hospital D, and to some extent in Hospitals B and C, but not in A. Lack of awareness by HCP of the extent to which errors occur and the potential causes are a barrier to improving patient safety. One example, seen within this study, was the difference in attitude between doctors who have spent time in primary care and seen discrepancies from the ‘other side’, the primary care perspective. They were found to complete the discharge summaries more thoroughly.

Culture shapes the clinician and staff perceptions about what is normal behaviour in a particular work area. It also influences a person’s motivation to engage in safety behaviours.632 There is emerging evidence to support the effectiveness of interventions to improve safety culture.632 Westrum suggests that leadership shapes culture, which in turn encourages information flow and that an open, generative culture leads to the increased uptake of innovation.708 It is therefore imperative that organisations as well as encouraging a culture of safety appoint leaders to drive improvement in medicines reconciliation and communication at transitions. Strategies to improve patient safety culture work best when several interventions are combined that focus on leadership, team work, and behaviour change. The most common types of interventions include:

318 team training, team communication tools, executive walk rounds and interdisciplinary rounding. Weaver suggests the organisation should consider incorporating these elements in order to improve safety culture.632 Tools exist which allow organisations to assess their safety culture.709,710

Recently, an interest in HROs has been explored. High reliability organisations (HROs) are:

‘Organisations that work in situations that have the potential for large- scale risk and harm, but which manage to balance effectiveness, efficiency and safety. They also minimise errors through teamwork, awareness of potential risk and constant improvement.’711

A small number of studies have considered HROs in healthcare, whilst there is little empirical evidence exploring the links between HROs and safety culture, it is thought that HROs have a positive safety culture.711 The AHRQ in the US has issued guidance for hospital leaders on becoming a HRO which recommends strong leadership, ownership, appropriate policies and procedures a safety culture and just culture, similar attributes to those suggested to promote a culture of patient safety.712 These principles in turn should be applied to medicines reconciliation processes to improve safety and outcomes.

7.7 Recommendations for Practice, Policy and Theory

The findings within this study have a series of implications for practice, policy and theory. The following recommendations are suggested based on these findings, supported through existing literature. First and foremost, a common definition of medicines reconciliation must be accepted nationally to ensure a shared understanding across organisations and between HCPs.

Each organisation within this study and within the UK appear to be at varying stages in improving medicines reconciliation. Ideally, improvement should also incorporate cross-organisational benchmarking. However, this is likely to prove difficult due to the lack of replicable quality indicators for medicines reconciliation; the lack of 'Gold-standard' methods for eliciting a medication history and transferring information across the interface; and the processing of this information within primary care. This lack of standardisation has the 319 potential to lead to error. Some national organisations have attempted to progress standardisation through the publication of various documents e.g. RPS which address aspects of medication safety at transitions.276,648,650,713 These, however, have predominantly used consensus of expert opinion rather than drawing on the extensive evidence base and or employing a theory-based approach. Much of the previous focus has been targeted at secondary care. Whilst this still remains important, this study has shown that an increased focus in primary care is also required to improve standard operating procedures and the way in which discharge summaries are processed.

This study has demonstrated that accurate medicines reconciliation requires a combination of appropriate questioning, together with effective, patient-centred communication skills plus correct interpretation of the information gathered. Considering this, together with the contributory factors and defences identified, a framework of standards for medicines reconciliation should be developed using the evidence-informed observation schedules as a template and integrating the consultation frameworks used for analysis of communication. The next stage would be to evaluate this overarching medicines reconciliation framework using a multidisciplinary group from primary and secondary care organisations potentially employing either a Delphi approach or focus groups. Ultimately testing these in practice and determining their applicability to other specialities and settings, for example, surgical wards, hospices and psychiatry. Fernandes et al suggest an evidence based bronze to platinum ranking which looks at the varying intensity of medicines reconciliation activities conducted and could be used to ascertain different levels of attainment in implementing good practice, ultimately providing a way of ranking improvement within an organisation (see table 7.3).289

The NHS medication safety thermometer is seeking to improve this through collating data nationally on safety measures including medicines reconciliation to allow organisations to determine their position.714 The NHS safety thermometer, nevertheless, is only quantitative, and does not look at the quality aspects of the process (i.e. just because an organisation reconciles medication for a high number of patients does not mean that they are accurate or that the best possible process is being used).714 The NHS Safety Thermometer is

320 currently under review as it has been criticised for measuring point prevalence rather than trends over time (Gerry Armitage. Professor, Health Services Research, University of Bradford. Personal Communication. 24th February 2014) The suggested framework would provide a way of measuring the quality aspects of the process.

Organisations must also clearly define the roles and responsibilities of their staff, determining the most appropriate skill mix for their organisation, whilst still recognising the limitations of the individual professions. Most significantly, the feasibility of pharmacy staff undertaking the initial medication history in terms of both accuracy and cost effectiveness should be fully evaluated given the wide body of literature and findings within this study. The ability of nursing staff to counsel effectively at discharge should also be considered. The nursing profession must also recognise the role they play within medicines reconciliation and the importance of information provided by patients during administration rounds. Community pharmacist could also have a greater role at both admission, providing information and at discharge, ensuring patients understand their medication.360,439

This thesis developed two evidence-informed observational schedules which were tested in practice then used as a framework to analyse the qualitative observations. Communication models including consultation models were also explored and successfully used to analyse patient-HCP interaction. Similarly, the YCFF was used as an analytical framework to categorise both the contributory factors and defences that affect the medicines reconciliation process. There is the opportunity that it could be applied to other observational work within patient safety to gain a deeper understanding of factors which could lead to error or patient harm, provided it is allied with other quantitative and qualitative methods that triangulate findings e.g. focus groups, questionnaires, interviews.

321

Table 7.3 Medicines reconciliation in varying levels of intensity as seen in published studies 289

Level Key components Published Examples Bronze BPMH with admission reconciliation 50,124 Silver Bronze level + reconciliation at discharge 183,463 by prescriber only ± electronically generated discharge prescription Gold Silver level + discharge reconciliation is 71,463 interprofessional e.g. prescribing physician and pharmacist collaboration + electronically generated discharge prescription Platinum Gold level + attention to broader 71,100,105,122 medication management issues e.g. appropriateness of agents, safety and effectiveness assessment + medication counselling prior to discharge (including discussion of medication changes) + provision of patient friendly reconciled medication schedules at discharge Diamond Platinum level + additional elements such 123,312,403,405 as post discharge follow up phone call by HCP

Communication of medication changes with rationale directly to community pharmacy and primary care physician

7.8 Further Research

The findings within this study have identified further areas associated with medication error at transitions which should be prioritised for research. These have been outlined throughout the thesis with the key suggestions summarised in this section.

Studies exploring the rate of discrepancies and errors which go on to cause tangible or measurable harm are needed to be able to more accurately look at 322 potential cost and cost savings due to intervention. These are difficult to design as it would be unethical to identify discrepancies in IMCs prospectively and not intervene to determine whether or not they would go on to harm. For this reason it must be done retrospectively, perhaps identifying the number of harm events that occur within a given timeframe, possibly identified through incident reporting or potentially prospective data collection, followed by the retrospective analysis of the number of discrepancies which occurred over the same timeframe. Further studies should also consider the identification of measures of severity which recognise that not all discrepancies will result in patient harm.

The accuracy of pharmacists taking the initial medication history and prescribing at admission, in comparison to doctors in the UK, should be explored through redesign of medication clerking processes to involve pharmacists using quality improvement techniques. This should be followed by an evaluation post implementation using a controlled before and after design. Alternatively a randomised control trial could be used to determine whether the frequency and severity of discrepancies decrease when pharmacists take the initial medication history.

Frameworks of standards for the medicines reconciliation process, plus competence and educational frameworks should be developed and tested. This could be conducted using the literature review, findings from this study, and the evidence-informed observation schedules as a template, integrating the consultation frameworks used for analysis of communication. The development of competence and educational frameworks for HCP based on existing literature and findings from this study in both patient and medication safety supported by regulatory and professional bodies, similar to the NPSEF in Australia, could assist the standardisation of knowledge and skills throughout undergraduate curriculum. A multidisciplinary group from primary and secondary care organisations potentially employing either a Delphi approach or focus groups could be used to develop these, before testing these in practice. Their applicability to other specialities and settings, for example, surgical wards, hospices and psychiatry could also be tested to try to improve the standard of communication at transitions and the reconciliation of medicines.

323

7.9 Strengths and Limitations of the Overall Study

The strengths and limitations of this study have been discussed in detail in previous chapters, consequently this section will summarise those judged to be the most notable, and those which have the broadest relevance to the entire study. The overall study design uses a pragmatic mixed methods approach which allows triangulation of quantitative with qualitative data. It has also allowed the quantitative findings to be contextualised through observation of primary and secondary care. At the time of writing, this approach does not appear to be evident in any previous medicines reconciliation studies.

An extensive literature review was conducted which informed the development of evidence informed observation schedules that were subsequently tested in practice. Epidemiological work demonstrated that the scale of the problems identified within the literature review existed within the research sites and consequently justified further exploration of the problem and undeniably the use of observation. The observation went on to identify contributory factors and defences from which specific interventions could be developed, or the process redesigned.

Despite these strengths, the study is only conducted in one health economy, nevertheless, the multi-centre design and purposive sampling allowed sites to be chosen to maximise generalisability and, in qualitative terms, transferability. Although some multi-centre studies have been published before in abstracts in the UK, none have previously been completed in the same depth. A key strength of the research design is the focus on care transitions, although the study only looks at the interface between hospital and general medical practice rather than other interfaces such as care homes and community pharmacy.

This work also included medical, elderly and renal patients. Whilst this is limited, and not a fully representative patient population, there is a significant body of evidence which demonstrates that similar reconciliation errors and discrepancies occur in areas such errors paediatrics, surgery and psychiatry. The gap in evidence concerning the nature of contributory factors in these specialties, however, remains. Further work looking at a broader applicability of the findings would be of benefit. There may be other potential challenges with

324 medicines reconciliation in these specialties, for example surgical patients may be away from their bedside in theatre, paediatric patients may rely more heavily on a carer to provide information and mental health processes differ from acute hospitals in the way they conduct ward rounds.

The researcher’s role as a pharmacist is a strength as it facilitated access to research sites and allowed the researchers professional knowledge to be applied to the study design and data analysis, however, it could potentially also lead to bias. This was minimised through reflexivity by the researcher. It is unclear the advantage of the researchers employment in the NHS or as a pharmacist. Whilst the employment of the researcher within one of the NHS hospitals being studied may have made the researcher appear more bona fide to the participants, it may also have created bias within observation. Insider knowledge can cause blindness and lack of objectivity.

Despite the limitations above, the principal strength of this study is that it is the first multi-centre, mixed-methods study to review the entire process of medicines reconciliation across primary and secondary care within the UK and internationally.

7.9 Conclusion Research illustrating issues with communication about medication at transition points in care dates back over 25 years within the UK, with medication, communication and care coordination continuing to be a problem. Previously studies of medicines reconciliation have looked at the scale of error and discrepancy and suggested interventions to improve inaccuracy at care transitions, without exploring the reasons why such phenomena occur. This mixed methods study has appraised routine medicines reconciliation practice across in four health economies in West Yorkshire, examining the entire process from hospital admission through to discharge and into primary care. Following a large scale literature review, the roles of different HCPs involved in the process were examined, alongside that of the patient and their interaction with the healthcare team. The various systems that support reconciliation have also been appraised including IT; the analysis was enhanced by a human factors perspective, so as to determine the safety of the process.

325

Firstly, the study determined the type and severity of discrepancies occurring at admission to hospital so as to demonstrate that medicines reconciliation discrepancies were a problem in the research sites studied and to justify further exploration of the problem. This was followed by a qualitative observational study of current medicines reconciliation practice in both primary and secondary care. Two evidence-informed, observational schedules were developed, tested, and used as a framework for data collection and analysis. The YCFF was also applied to identify potential factors which contribute to medicines reconciliation error/discrepancy, and also to identify and categorise potential defences which could lead to the reduction of such errors or discrepancies.608 Finally consultation models were used to examine HCP-patient interaction.

Both quantitative and qualitative findings have shown wide variation in practice both within and between organisations, some of which was detrimental to the accuracy of medicines reconciliation. The quantitative findings demonstrated higher rates of unintended medication discrepancies per patient than in previous studies; however, the majority were omissions, consistent with other research. And, although the literature and results in Chapter 4 strongly suggest that the majority of discrepancies related to medicines reconciliation do not relate to patient harm the study has demonstrated that a small but clinically significant majority were judged to cause harm. The observation placed the quantitative findings in context, inadequate communication, specific individual factors such as variation in approach and assumptions made about patients, and patient factors such as lack of capacity or understanding of English were commonplace. The quantitative findings also showed that the patient and patient’s own medication were used in less than half of all medicines reconciliations with much time being spent waiting for information from general practice. This was supported by the qualitative findings which determined that patients were not routinely involved in discussions about their medication, despite their capacity to do so, and were also unlikely to be given medicines information during admission, including changes to their prescription. Areas of good practice which acted as defence against errors and/or discrepancies were identified. These were mainly attributable to individuals and teams gained from their experience e.g. looking for certain ‘red flags’ within information sources rather than the organisations themselves. However, some organisations had

326 standardised the process e.g. documentation and implemented electronic discharge systems. The better use of IT would potentially reduce some of the potential risks, mindful that it could also introduce new types of error as novel ways of working are established.

These findings suggest potential reasons why there is such a high rate of error associated with medicines reconciliation. This work can also inform the design of interventions which target the cause of mis-communication about medication at interfaces in care. There is an increased need for consistency and greater standardisation of medicines reconciliation-related policy, procedures and documentation. This could be achieved through adoption of established good practice, better utilisation of IT and improvement of safety culture. Also, the development of a framework of standards which could be used as an assessment tool for organisations to assess their practice and benchmark their service quality. A review of skill mix is also needed to determine the most suitable staff to undertake relevant tasks associated with medicines reconciliation.

This study looked specifically at medical and renal patients; the generalisability and transferability of these findings to other specialties is unclear. Although the interface between secondary care and general practice, was studies. Other interfaces such as secondary care and intermediate care or care homes were excluded.

327

References

1 World Health Organisation (WHO). Assuring medication accuracy at transitions in care. Patient safety solutions. World Health Organisation 2007;1.

2 Cambridge M. 5 million lives campaign. Getting started kit: prevent adverse drug events (Medication Reconciliation). How-to guide. 2008. Accessed 23rd September 2010. www.ihi.org.

3 Australian Commision on Safety and Quality in Healthcare. Medication Reconciliation. 2013. Accessed 23rd March 2014. http://www.safetyandquality.gov.au/our-work/medication- safety/medication-reconciliation/.

4 Institute for Safe Medication Practices. National Medication Reconciliation Strategy; Identifying Practice Leaders for Medication Reconciliaiton in Canada. 2012. Accessed 25th May 2014. www.ismp-canada.org.

5 Institute of Safe Medication Practice. Assuring Medication Accuracy at Transitions in Care: Medication Reconciliation. High 5s: Action on Patient Safety Getting Started Kit. 2009. Accessed 25th April 2014. https://www.high5s.org/pub/Manual/TrainingMaterials/MR_Getting_Started _Kit.pdf.

6 National Institute for Health and Care Excellence & National Patient Safety Agency (NICE & NPSA). Technical patient safety solutions for medicines reconciliation on admission of adults to hospital. 2007. Accessed 10th August 2010. www.nice.org.uk/.

7 Agency for Healthcare Research and Quality (AHRQ). AHRQ Quality Indicators. 2013. Accessed 4th May 2014. http://qualityindicators.ahrq.gov/.

8 Brennan TA, Leape L, Laird N, Hebert L, Localio A, Lawthers A, Newhouse J, Weiler P. Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I. N Engl J Med 1991;324:370-76.

9 Leape LL, Ta B, Laird N, Lawthers AG, Localio AR, Barnes BA, Hebert L, Newhouse JP, Weiler PC, Hiatt H. The nature of adverse events in hospitalized patients. Results of the Harvard Medical Practice Study II. N Engl J Med 1991;324:377-84.

10 Paton C, McIntyre S, Bhatti S, Shingleton-Smith A, Gray R, Gerrett D, Barnes T. Medicines reconciliation on admission to inpatietn psychiatric care: findings from a UK quality improvement programme. Ther Adv Psychopharmacol 2011;1:101-10.

11 Kohn L, Corrigan J, Donaldson M. To Err Is Human: Building a Safer Health System. Washington, D.C: National Academy Press, 1999.

328

12 Department of Health. The New NHS: Modern, Dependable. 1997. London.

13 Department of Health. An Organisation with a Memory., 2000. Accessed 17th January 2014. http://www.dh.gov.uk/prod_consum_dh/groups/dh_digitalassets/@dh/@en /documents/digitalasset/dh_4065086.pdf.

14 Vincent C. Patient Safety. London: Elsevier, 2006.

15 National Patient Safety Foundation (NPSF). Patient Safety Dictionary A-Z. 2014. Accessed 24th May 2014. http://www.npsf.org/for-healthcare- professionals/resource-center/definitions-and-hot-topics/patient-safety- dictionary-n-z/.

16 Reason J. Human error: models and management. Brit Med J 2000;320:768 - 70.

17 Department of Health. Safety First: a report for patients, clinicians and health care managers. 2006. London.

18 National Patient Safety Agency (NPSA). Patient Safety First 2008 to 2010. The campaign review. 2011. London.

19 National Patient Safety Agency (NPSA). About Patient Safety First. 2008. Accessed http://www.patientsafetyfirst.nhs.uk/Content.aspx?path=/About- the-campaign/.

20 Association IE. Definition and Domains of Ergonomics 2014. Accessed 25th April 2014. http://www.iea.cc/whats/index.html

21 Russ AL, Fairbanks RJ, Karsh B-T, Militello LG, Saleem JJ, Wears RL. The science of human factors: separating fact from fiction. 2013. April 16, 2013. http://qualitysafety.bmj.com/content/early/2013/04/15/bmjqs-2012- 001450.abstract.

22 National Patient Safety Agency (NPSA). Human factors and patient safety culture. NHS. Accessed 3rd May 2014. http://www.nrls.npsa.nhs.uk/resources/patient-safety-topics/human- factors-patient-safety-culture/.

23 Aronsen J. Medication errors: definitions and classification. Brit J Clin Pharmaco 2009;67:599–604.

24 Morimoto T, Gandhi TK, Seger AC, Hsieh TC, Bates DW. Adverse drug events and medication errors: detection and classification methods. Qual Saf Health Care 2004;13:306-14.

25 Reason J. Safety in the operating theatre – Part 2: Human error and organisational failure. Qual Saf Health Care 2005;14:56-60.

26 Armitage G, Newell R, J W. Journal of Evaluation in Clinical Practice Improving the quality of drug error reporting 2010;16:1189–97.

329

27 Department of Health. Building a safer NHS for patients: implementing an organisation with a memory. 2001. London.

28 Department of Health. Building a safer NHS for patients: Improving Medication Safety. 2004. London.

29 National Patient Safety Agency (NPSA). Safety in doses: improving the use of medicines in the NHS. 2007. London.

30 Department of Health. Standards for better health 2006 London.

31 Boling PA. Care transitions and home health care. Clinics in Geriatric Medicine 2009;25:135-48.

32 Department of Health. Achieving timely ‘simple’discharge from hospital: a toolkit for the multi-disciplinary team. 2004. London.

33 Department of Health. Management of Medicines. A resource to support implementation of the wider aspects of medicines management for the National Service frameworks for Diabetes, Renal Services and Long-Term Conditions. 2004. London.

34 Department of Health. Medicines and Older People: Implementing medicines-related aspects of the NSF for Older People. 2001. London.

35 National Audit Office (NAO). Controller and Auditor General. Ensuring the effective discharge of older patients from NHS acute hospitals. London: HMSO 2003.

36 Shawahna R, Rahman NU, Ahmad M, Debray M, Yliperttula M, Decleves X. Impact of prescriber's handwriting style and nurse's duty duration on the prevalence of transcription errors in public hospitals. J Clin Nurs 2013;22:550-58.

37 Shekelle P. Nurse-patient ratios as a patient safety strategy. Ann Intern Med 2013;158:404-09.

38 Shah A. Discrepancies in transition of care for high alert medication warfarin. Pharmacotherapy 2013;33:E77-E78.

39 Foreshew G, Culshaw M, Milligan M. Medication history taking on the medical admissions ward and in the Emergency Department. Pharm World Sci 2008;30:964-65.

40 Argyle M, Newman C. An assessment of pharmacy discharge procedures and hospital communications with general practitioners. Pharm J 1996;256:903-05.

41 Brackenborough S. Views of patients, general practitioners and community pharmacists on medication related discharge information. Pharm J 1997;259:1020-23.

42 Harding J. Study of discharge communications from hospital doctors to an inner London general practice. J R Coll Gen Pract 1987;37:494-95. 330

43 Joint Commission on Accreditation of Healthcare Organizations USA. Using medication reconciliation to prevent errors. Sentinel Event Alert 2006:1-4.

44 Joint Commission on Accreditation of Healthcare Organizations USA. Hospital Accreditation Program. National Patient Safety Goals. Joint Commission on Accreditation of Healthcare Organizations, U. S. A. 2010:14-18.

45 Royal Pharmaceutical Society (RPS). Medicines Optimisation: Helping patients to make the most of medicines. Good practice guidance for healthcare professionals in England. 2013. Accessed 21st October 2013. http://www.rpharms.com/promoting-pharmacy-pdfs/helping-patients-make- the-most-of-their-medicines.pdf

46 Joint Commission. Approved: will not score medication reconciliation in 2009. Plans to review, refine NPSG 8 for 2010. Jt Comm Perspect 2009;29:1,3.

47 NPSA. NICE/NPSA issues its first patient safety solution guidance to improve medicines reconciliation at hospital admission 2007. Accessed 17th October 2010 2010. http://www.npsa.nhs.uk/corporate/news/guidance-to-improve-medicines- reconciliation/.

48 Campbell F, Karnon J, Czoski-Murray C, Jones R. A systematic review of the effectiveness and costeffectiveness of interventions aimed at preventing medication error (medicines reconciliation) at hospital admission. The University of Sheffield, School of Health and Related Research (ScHARR) 2007.

49 Gleason K, Groszek J, Sullivan C, Rooney D, Barnard C, Noskin G. Reconciliation of discrepancies in medication histories and admission orders of newly hospitalized patients. Am J Health-Syst Ph 2004;61:1689- 95.

50 Cornish PL, Knowles SR, Marchesano R, Tam V, Shadowitz S, Juurlink DN, Etchells EE. Unintended medication discrepancies at the time of hospital admission. Arch Intern Med 2005;165:424-29.

51 Ferner R, Aronson J. Clarification of terminology in medication errors: definitions and classification. Drug Safety 2006;29:1011-22.

52 Dean B, Schachter M, Vincent C, Barber N. Prescribing errors in hospital inpatients: their incidence and clinical significance. Qual Saf Health Care 2002;11:340-44.

53 Bates DW, Cullen DJ, Laird N, Petersen LA, Small SD, Servi D, Laffel G, Sweitzer BJ, Shea BF, Hallisey R. Incidence of adverse drug events and potential adverse drug events. Implications for prevention. ADE Prevention Study Group. JAMA-J Am Med Assoc 1995;274:29-34.

331

54 Forster A, Murff H, Peterson J, Gandhi T, Bates D. The incidence and severity of adverse events affecting patients after discharge from the hospital. Ann Intern Med 2003;138:161-7.

55 Boockvar K, Fishman E, Kyriacou C, Monias A, Gavi S, Cortes T. Adverse events due to discontinuations in drug use and dose changes in patients transferred between acute and long-term care facilities. Arch Intern Med;164:545-50.

56 Coleman E, Min S, Chomiak S, Kramer A. Posthospital care transitions: patterns, complications, and risk identification. Health Serv Res 2004;39:1449–66.

57 Collins. English Dictionary. 2014. Accessed 3rd March 2014. http://www.collinsdictionary.com/.

58 World Health Organisation (WHO). WHO Draft Guidelines for Adverse Event Reporting and Learning Systems. 2005. Accessed 23rd February 2014. http://www.who.int/patientsafety/events/05/Reporting_Guidelines.pdf

59 Mueller SK, Sponsler KC, Kripalani S, Schnipper JL. Hospital-based medication reconciliation practices: a systematic review. Arch Intern Med 2012;172:1057-69.

60 Chhabra PT, Rattinger GB, Dutcher SK, Hare ME, Parsons KL, Zuckerman IH. Medication reconciliation during the transition to and from long-term care settings: A systematic review. Res Social Adm Pharm 2012;8:60-75.

61 Steeb D, Webster L. Improving care transitions: optimizing medication reconciliation. J Am Pharm Assoc 2012;52:e43-e52.

62 Closs SJ. Discharge communications between hospital and community health care staff: a selective review. Health Soc Care Comm 1997;5:181- 97.

63 Gupta R, Marshall J, Munoz JC, Kottoor R, Jamal MM, Vega KJ. Decreased acid suppression therapy overuse after education and medication reconciliation. Int J Clin Pract 2013;67:60-65.

64 Varkey P, Cunningham J, O'Meara J, Bonacci R, Desai N, Sheeler R. Multidisciplinary approach to inpatient medication reconciliation in an academic setting. Am J Health-Syst Ph 2007;64:850-54.

65 Varkey P, Cunningham J, Bisping DS. Improving medication reconciliation in the outpatient setting. Jt Comm J Qual Patient Saf 2007;33:286-92.

66 Mills PR, McGuffie AC. Formal medicine reconciliation within the emergency department reduces the medication error rates for emergency admissions. Emerg Med J 2010;27:911-15.

67 Legrain S, Tubach F, Bonnet-Zamponi D, Lemaire A, Aquino J-P, Paillaud E, Taillandier-Heriche E, Thomas C, Verny M, Pasquet B, Moutet AL, Lieberherr D, Lacaille S. A new multimodal geriatric discharge-planning 332

intervention to prevent emergency visits and rehospitalizations of older adults: The Optimization of Medication in AGEd Multicenter Randomized Controlled Trial. J Am Geriatr Soc 2011;59:2017-28.

68 Gardner B, Graner K. Pharmacists' medication reconciliation-related clinical interventions in a children's hospital. Jt Comm J Qual Patient Saf 2009;35:278-82.

69 Gardella JE, Cardwell TB, Nnadi M. Improving medication safety with accurate preadmission medication lists and postdischarge education. Jt Comm J Qual Patient Saf 2012;38:452-58.

70 Feldman LS, Costa LL, Feroli ER, Jr., Nelson T, Poe SS, Frick KD, Efird LE, Miller RG. Nurse-pharmacist collaboration on medication reconciliation prevents potential harm. J Hosp Med 2012;7:396-401.

71 Dedhia P, Kravet S, Bulger J, Hinson T, Sridharan A, Kolodner K, Wright S, E H. A quality improvement intervention to facilitate the transition of older adults from three hospitals back to their homes. J Am Geriatr Soc 2009;57:1540-6.

72 Bergqvist M, Ulfvarson J, Karlsson EA. Nurse-led medication reviews and the quality of drug treatment of elderly hospitalized patients. European Journal of Clinical Pharmacology 2009;65:1089-96.

73 Bergkvist A, Midlöv P, Höglund P, Larsson L, Bondesson A, Eriksson T. Improved quality in the hospital discharge summary reduces medication errors--LIMM: Landskrona Integrated Medicines Management. Eur J Clin Pharmacol 2009;65:1037-46.

74 Bracey G, Miller G, Franklin BD, Jacklin A, Gaskin G. The contribution of a pharmacy admissions service to patient care. Clin Med 2008;8:53-57.

75 Brookes K, Scott MG, McConnell JB. The benefits of a hospital based community services liaison pharmacist. Pharm World Sci 2000;22:33-8.

76 Collins D, Nickless G, Green CF. Medication histories: Does anyone know what a patient should be taking? Int J Pharm Pract 2004;12:173-78.

77 Dodds L. Unintended discrepancies between pre-admission and admission prescriptions identified by pharmacy-led medicines reconciliation: Results of a collaborative service evaluation across East and South East England. Int J Pharm Pract 2010;18:9-10.

78 Dutton K, Hedger N, Wills S, Brown D, Davies P. Prevent medication errors on admission. Clin Govern Int J 2003;8:128-37.

79 McFadzean E, Isles C, Moffat J, Norrie J, Stewart D. Is there a role for a prescribing pharmacist in preventing prescribing errors in a medical admission unit? Pharm J 2003;270:896-99.

80 National Prescribing Centre (NPC). Medicines reconciliation: a guide to implementation. National Prescribing Centre, 2008. Accessed 27th October 2010. 333

http://www.npci.org.uk/medicines_management/safety/reconcil/resources/r econciliation_guide-a5ordered.pdf.

81 Vogelsmeier A, Pepper GA, Oderda L, Weir C. Medication reconciliation: A qualitative analysis of clinicians' perceptions. Res Social Adm Pharm 2013;9:419-30.

82 Pickrell L, Duggan C, Dhillon S. From hospital admission to discharge: an exploratory study to evaluate seamless care. Pharm J 2001;267:650-53.

83 Royal Pharmaceutical Society (RPS). Professional Standards for Hospital Pharmacy Services: Optimising patient outcomes from medicines. 2012. Accessed 25th March 2014. http://www.rpharms.com/support-pdfs/rps--- professional-standards-for-hospital-pharmacy.pdf.

84 Dahlkvist J, Pettersson H, Schwan A, Toss H. [Half of the patients older than 80 years in the Uppsala county with more than ten different medications are treated with furosemide]. Läkartidningen 2011;108:612- 12.

85 Zhang Y, Holford TR, Leaderer B, Zahm SH, Boyle P, Morton LM, Zhang B, Zou K, Flynn S, Tallini G, Owens PH, Zheng T. Prior medical conditions and medication use and risk of non-Hodgkin lymphoma in Connecticut United States women. Cancer Cause Control 2004;15:419-28.

86 Department of Health. NHS Plan: A plan for investment a plan for reform. 2000. Accessed 24th January 2014. http://webarchive.nationalarchives.gov.uk/+/www.dh.gov.uk/en/publication sandstatistics/publications/publicationspolicyandguidance/dh_4002960.

87 Spinewine A, Claeys C, Foulon V, Chevalier P. Approaches for improving continuity of care in medication management: a systematic review. Int J Qual Health Care 2013;25:403-17.

88 Morrison P, Abu-Arafeh I, Davison J, Chapman S. Optimum prescribing of discharge medicines: roles of hospital and community pharmacists Pharm J 2004;272:224-27.

89 Couch G, Quah SM, Ng YC, McKellar A. Continuity of medication supply and provision of patient information on discharge: development of a survey tool. J Pharm Pract Res 2007;37:210-13.

90 Lesselroth BJ, Holahan PJ, Adams K, Sullivan ZZ, Church VL, Woods S, Felder R, Adams S, Dorr DA. Primary care provider perceptions and use of a novel medication reconciliation technology. Inform Prim Care 2011;19:105-18.

91 Tully M, Kettis A, Hoglund A, Morlin C, Schwan A, Ljungberg C. Transfer of data or re-creation of knowledge - Experiences of a shared electronic patient medical records system. Res Social Adm Pharm 2013;9:965-74.

334

92 Smith J, Coleman E, Min S. A new tool for identifying discrepancies in postacute medications for community-dwelling older adults. Am J Geriatr Pharmacother 2004;2:141-7.

93 Bray-Hall S, Schmidt K, Aagaard E. Toward safe hospital discharge: a transitions in care curriculum for medical students. J Gen Intern Med 2010;25:878-81.

94 Bray-Hall ST, Abraham A, Aagaard EM. Development of a transitions in care curriculum for medical students. J Gen Intern Med 2008;23:198-99.

95 Sando KR, Elliott J, Stanton ML, Doty R. An educational tool for teaching medication history taking to pharmacy students. Am J Pharm Educ 2013;77:105.

96 van Zuilen MH, Kaiser RM, Mintzer MJ. A competency-based medical student curriculum: taking the medication history in older adults. J Am Geriatr Soc 2012;60:781-85.

97 Walsh KE, Ettinger WH, Klugman RA. Physician quality officer: a new model for engaging physicians in quality improvement. Am J Med Qual 2009;24:295-301.

98 Day SJ, Altman DG. Blinding in clinical trials and other studies. BMJ 2000;321:504.

99 Campbell NC, Murray E, Darbyshire J, Emery J, Farmer A, Griffiths F, Guthrie B, Lester H, Wilson P, Kinmonth AL. Designing and evaluating complex interventions to improve health care. Brit Med J 2007;334:455-59.

100 Al-Rashed S, Wright D, Roebuck N, Sunter W, Chrystyn H. The value of in-patient counselling to elderly patients prior to discharge. Brit J Clin Pharmaco 2002;54:657-64.

101 Bolas H, Brookes K, Scott M, McElnay J. Evaluation of a hospital-based community liaison pharmacy service in Northern Ireland. Pharm World Sci 2004;26:114-20.

102 Boorman S, Cairns C. Another way forward for pharmaceutical care: a team-based clinical pharmacy service. Pharm J 2000;264:343-46.

103 McRobbie D, Bednall R, West T. Assessing the impact of re-engineering of pharmacy services to general medical wards Pharm J 2003;270:342-45.

104 Nassaralla CL, Naessens JM, Chaudhry R, Hansen MA, Scheitel SM. Implementation of a medication reconciliation process in an ambulatory internal medicine clinic. Qual Saf Health Care 2007;16:90-94.

105 Nazareth I, Burton A, Shulman S, Smith P, Haines A, Timberall H. A pharmacy discharge plan for hospitalized elderly patients - a randomized controlled trial. Age Ageing 2001;30:33-40.

335

106 Shaw H, Mackie C, Sharkie I. Evaluation of effect of pharmacy discharge planning on medication problems experienced by discharged accute admission mental health patients. Int J Pharm Pract 2000;8:144-53.

107 Wilcock M, Lawrence J. Is there a role for community pharmacists in identifying discrepancies in medication histories for patients admitted to hospital. Pharm J 2004;272:253-56.

108 Midlöv P, Deierborg E, Holmdahl L, Höglund P, Eriksson T. Clinical outcomes from the use of medication report when elderly patients are discharged from hospital. Pharm World Sci 2008;30:840-45.

109 Cornu P, Steurbaut S, Leysen T, De Baere E, Ligneel C, Mets T, Dupont AG. Discrepancies in medication information for the primary care physician and the geriatric patient at discharge. Ann Pharmacother 2012;46:983-90.

110 Stitt D, Elliott DP, Thompson S. Medication discrepancies identified at time of hospital discharge in a geriatric population. Am J Geriatr Pharmac 2011;9:234-40.

111 Hu S, Capezuti E, Foust J, Boltz M, Kim H. Medication discrepancy and potentially inappropriate medication in older chinese-american home-care patients after hospital discharge. Am J Geriatr Pharmac 2012;10:284-95.

112 Alibhai SM, Han RK, Naglie G. Medication education of acutely hospitalized older patients. J Gen Intern Med 1999;14:610-16.

113 Villanyi D, Fok M, Wong R. Medication reconciliation: identifying medication discrepancies in acutely ill hospitalized older adults. Am J Geriatr Pharmac 2011;9:339-44.

114 Midlöv P, Holmdahl L, Eriksson T, Bergkvist A, Ljungberg B, Widner H, Nerbrand C, Höglund P. Medication report reduces number of medication errors when elderly patients are discharged from hospital. Pharm World Sci 2008;30:92-98.

115 Widagdo IS, Nyfort-Hansen K, Kowalski SR. Prevalence of potentially inappropriate medication use in elderly hospitalised patients. J Pharm Pract Res 2011;41:122-25.

116 Manning D, O'Meara J, Williams A, Rahman A, Tammel K, Myhre D, Carter L. 3D: a tool for medication discharge education. Qual Saf Health Care 2007;16:71-76.

117 Hellström LM, Bondesson Å, Höglund P, Eriksson T. Errors in medication history at hospital admission: prevalence and predicting factors. BMC Clinical Pharmacology 2012;12:9-9.

118 Peyton L, Ramser K, Hamann G, Patel D, Kuhl D, Sprabery L, Steinhauer B. Evaluation of medication reconciliation in an ambulatory setting before and after pharmacist intervention. J Am Pharm Assoc 2010;50:490-95.

119 Tulloch J, Evans B. Evaluation of the accuracy of the Saskatchewan Health Pharmaceutical Information Program for determining a patient's 336

medication use immediately before admission. Can J Hosp Pharm 2009;62:21-7.

120 Herrero-Herrero J, García-Aparicio J. Medication discrepancies at discharge from an internal medicine service. Eur J Intern Med 2011;22:43- 48.

121 Andersen SE, Pedersen AB, Bach KF. Medication history on internal medicine wards: assessment of extra information collected from second drug interviews and GP lists. Pharmacoepidem Dr S 2003;12:491-98.

122 Murphy E, Oxencis C, Klauck J, Meyer D, Zimmerman J. Medication reconciliation at an academic medical center: Implementation of a comprehensive program from admission to discharge. Am J Health-Syst Ph 2009;66:2126-31.

123 Schnipper JL, Kirwin JL, Cotugno MC, Wahlstrom SA, Brown BA, Tarvin E, Kachalia A, Horng M, Roy CL, McKean SC, Bates DW. Role of pharmacist counseling in preventing adverse drug events after hospitalization. Arch Intern Med 2006;166:565-71.

124 Kwan Y, Fernandes OA, Nagge JJ, Wong GG, Huh JH, Hurn DA, Pond GR, Bajcar JM. Pharmacist medication assessments in a surgical preadmission clinic. Arch Intern Med 2007;167:1034-40.

125 Unroe K, Pfeiffenberger T, Riegelhaupt S, Jastrzembski J, Lokhnygina Y, Colón-Emeric C. Inpatient medication reconciliation at admission and discharge: A retrospective cohort study of age and other risk factors for medication discrepancies. Am J Geriatr Pharmac 2010;8:115-26.

126 Nelson LA, Graham MR, Schaefer MG. Characterization of medication discrepancies occurring at the time of discharge from an adult state psychiatric inpatient facility. Hosp Pharm 2011;46:254-61.

127 Brown EL, Raue PJ, Mlodzianowski AE, Meyers BS, Greenberg RL, Bruce ML. Transition to home care: quality of mental health, pharmacy, and medical history information. Int J Psychiat Med 2006;36:339-49.

128 Terry DRP, Solanki GA, Sinclair AG, Marriott JF, Wilson KA. Clinical significance of medication reconciliation in children admitted to a UK pediatric hospital: observational study of neurosurgical patients. Paediatr Drugs 2010;12:331-37.

129 Huynh C, Wong ICK, Tomlin S, Terry D, Sinclair A, Wilson K, Jani Y. Medication discrepancies at transitions in pediatrics: a review of the literature. Paediatr Drugs 2013;15:203-15.

130 Huynh C, Jani Y, Wong ICK, Dewchand B, Tomlin S. What is the current practice of medicines reconciliation in children nationally in the UK? The Int J Pharm Pract 2013;21:62-5.

131 Cranmer PS. Pharmacist interventions on a medical admissions unit: drug history taking. Pharm World Sci 2008;30:484-85.

337

132 Badham HJ, Dorward BJ, Thomas N. Audit of drug history taking, prescribing and documentation of changes to anti-convulsant drugs within Royal Hallamshire Hospital, Sheffield. Pharm World Sci 2007;29:A269- A69.

133 Virani N, Onatade R, Mehta R. Frequency of drug history taking by pharmacists at King's College Hospital. Pharm World Sci 2008;30:408-09.

134 Burridge N. SHPA standards of practice for the provision of medication reconciliation. J Pharm Pract Res 2007;37:231-33.

135 Joint Commission on Accreditation of Healthcare Organizations. Medication reconciliation: complying with NPSG.03.06.01. Joint Commission Perspectives on Patient Safety 2011;11:1.

136 Hassali MAA, Al-Haddad M, Shafie AA, Tangiisuran B, Saleem F, Atif M, Al-Qazaz H. Perceptions among general medical practitioners toward implementation of medication reconciliation program for patients discharged from hospitals in Penang, Malaysia. J Patient Saf 2012;8:76- 80.

137 Clay BJ, Halasyamani L, Stucky ER, Greenwald JL, Williams MV. Results of a medication reconciliation survey from the 2006 Society of Hospital Medicine national meeting. J Hosp Med 2008;3:465-72.

138 Johnston R, Saulnier L, Gould O. Best possible medication history in the emergency department: comparing pharmacy technicians and pharmacists. Can J Hosp Pharm 2010;63:359-65.

139 Dersch-Mills D, Hugel K, Nystrom M. Completeness of information sources used to prepare best possible medication histories for pediatric patients. Can J Hosp Pharm 2011;64:10-5.

140 Fitzsimons M, Grimes T, Galvin M. An evaluation of the availability and accuracy of sources of pre-admission medication history data. Pharmacoepidem Dr S 2010;19:648-49.

141 Peterson JF, Shi Y, Denny JC, Matheny ME, Schildcrout JS, Waitman LR, Miller RA. Prevalence and clinical significance of discrepancies within three computerized pre-admission medication lists. Proc AMIA Symp 2010;2010:642-6.

142 Turchin A, Hamann C, Schnipper JL, Graydon-Baker E, Millar SG, McCarthy PC, Coley CM, Gandhi TK, Broverman CA. Evaluation of an inpatient computerized medication reconciliation system. J Am Med Inform Assn 2008;15:449-52.

143 Salanitro A, Kripalani S, Resnic J, Mueller S, Wetterneck T, Haynes K, Stein J, Kaboli P, Labonville S, Etchells E, DJ C, Hanson D, Greenwald J, Williams M, Schnipper J. Rationale and design of the Multicenter Medication Reconciliation Quality Improvement Study (MARQUIS). 2013. 2nd July 2013. http://www.biomedcentral.com/1472-6963/13/230.

338

144 Galvin M, Jago-Byrne M-C, Fitzsimons M, Grimes T. Clinical pharmacist's contribution to medication reconciliation on admission to hospital in Ireland. Int J Clin Pharm 2013;35:14-21.

145 Pippins JR, Gandhi TK, Hamann C, Ndumele CD, Labonville SA, Diedrichsen EK, Carty MG, Karson AS, Bhan I, Coley CM, Liang CL, Turchin A, McCarthy PC, Schnipper JL. Classifying and predicting errors of inpatient medication reconciliation. J Gen Intern Med 2008;23:1414-22.

146 Agrawal A, Aronson J, Britten N, Ferner R, de Smet P, Fialová D, Fitzgerald R, Likić R, Maxwell S, Meyboom R, Minuz P, Onder G, Schachter M, Velo G. Medication errors: problems and recommendations from a consensus meeting. Brit J Clin Pharmaco 2009;67:592-98.

147 Warholak TL, McCulloch M, Baumgart A, Smith M, Fink W, Fritz W. An exploratory comparison of medication lists at hospital admission with administrative database records. J Manag Care Pharm 2009;15:751-58.

148 Gracious B, Abe N, Sundberg J. The importance of taking a history of over-the-counter medication use: a brief review and case illustration of "PRN" antihistamine dependence in a hospitalized adolescent. J Child Adolesc Psychopharmacol 2010;20:521-4.

149 Oborne CA, Luzac ML. Over-the-counter medicine use prior to and during hospitalization. Ann Pharmacother 2005;39:268-73.

150 Klieman P, Fredman B, Goldberg A, Confino-Cohen R. Drug allergy in hospitalized perioperative patients: the contribution of a structured questionnaire for accurate diagnosis. Immunotherapy 2013;5:837-41.

151 Bhalla N, Duggan C, Dhillon S. The incidence and nature of drug-related admissions to hospital Pharm J 2003;270:583-86.

152 Cohen D. The Psychiatric Medication History: context, purpose, and method. Social Work in Mental Health 2003;1:5-28.

153 Moore C, Wisnivesky J, Williams S, McGinn T. Medical errors related to discontinuity of care from an inpatient to an outpatient setting. J Gen Intern Med 2003;18:646-51.

154 Pronovost P, Weast B, Schwarz M, Wyskiel RM, Prow D, Milanovich SN, Berenholtz S, Dorman T, Lipsett P. Medication reconciliation: A practical tool to reduce the risk of medication errors. J Crit Care 2003;18:201-05.

155 World Health Organisation (WHO). High 5s: Action on Patient Safety. Standard Operating Protocol Fact Sheet: Medication Reconciliation. 2011. Accessed February 7th 2014. http://www.who.int/patientsafety/implementation/solutions/high5s/ps_med_ rec_fs_Mar_2011.pdf.

156 Agency for Healthcare Research and Quality. Medications at Transitions and Clinical Handoffs (MATCH) Toolkit for Medication Reconciliation 2012. Accessed 23rd March 2014. www.ahrq.gov.

339

157 Chen D, Burns A. ASHP – APhA Medication Reconciliation Initiative Workgroup Meeting February 12, 2007. Summary and Recommendations. 2007. 4th October 2013. http://www.ashp.org/s_ashp/docs/files/MedRec_ASHP_APhA_Wkgrp_Mtg Summary.pdf.

158 Allende Bandres M, Arenere Mendoza M, Gutierrez Nicolas F, Calleja Hernandez M, Ruiz La Iglesia F. Pharmacist-led medication reconciliation to reduce discrepancies in transitions of care in Spain. Int J Clin Pharm 2013;35:1083-90.

159 Boockvar KS, Liu S, Goldstein N, Nebeker J, Siu A, Fried T. Prescribing discrepancies likely to cause adverse drug events after patient transfer. Qual Saf Health Care 2009;18:32-36.

160 Cua YM, Kripalani S. Medication use in the transition from hospital to home. Ann Acad Med Singap 2008;37:136-41.

161 Schwarz M, Wyskiel R. Medication reconciliation: developing and implementing a program. Crit Care Nurs Clin North Am 2006;18:503-07.

162 Mergenhagen K, Blum S, Kugler A, Livote E, Nebeker J, Ott M, Signor D, Sung S, Yeh J, Boockvar K. Pharmacist-versus physician-initiated admission medication reconciliation: impact on adverse drug events. Am J Geriatr Pharmac 2012;10:242-50.

163 Meyer C, Stern M, Woolley W, Jeanmonod R, Jeanmonod D. How reliable are patient-completed medication reconciliation forms compared with pharmacy lists? Am J Emerg Med 2012;30:1048-54.

164 Miller SL, Miller S, Balon J, Helling TS. Medication reconciliation in a rural trauma population. Ann Emerg Med 2008;52:483-91.

165 Rogers G, Alper E, Brunelle D, Federico F, Fenn CA, Leape LL, Kirle L, Ridley N, Clarridge BR, Bolcic-Jankovic D, Griswold P, Hanna D, Annas CL. Reconciling medications at admission: safe practice recommendations and implementation strategies. Jt Comm J Qual Patient Saf 2006;32:37- 50.

166 Vira T, Colquhoun M, Etchells E. Reconcilable differences: correcting medication errors at hospital admission and discharge. Qual Saf Health Care 2006;15:122-26.

167 Yi SB, Shan JCP, Hong GL. Medication reconciliation service in Tan Tock Seng hospital. Int J Health Care Qual Assur 2013;26:31-36.

168 Zeigler AJ, McAllen KJ, Slot MG, Barletta JF. Medication reconciliation effect on prolonged inpatient stress ulcer prophylaxis. Ann Pharmacother 2008;42:940-6.

169 Society of Hospital Pharmacists of Australia Committee of Speciality Practice in Clinical Pharmacy. SHPA standards of practice for the provision of medication reconciliation. J Pharm Pract Res 2007;37:231-3.

340

170 Coleman E. Commissioned Paper: Transitional Care Performance Measurement. Performance Measurement Report, Institute of Medicine, 2006. Appendix I. 2006:250-76.

171 Patel K, Wright DJ, Desborough J. Effectiveness of electronic discharge handover system at Basildon and Thurrock University Hospitals NHS Foundation Trust: a service evaluation. Int J Pharm Pract 2011;19:34.

172 Richards M, Ashiru-Oredope D, Chee N. What errors can be identified by pharmacy-led medicines reconciliation? A prospective study. Acute medicine 2011;10:18-21.

173 Linsky A, Simon SR. Medication discrepancies in integrated electronic health records. Qual Saf Health Care 2013;22:103-09.

174 McMillan A, Trompeter J, Havrda D, Fox J. Continuity of care between family practice physicians and hospitalist services. J Healthc Qual 2013;35:41-9.

175 Quelennec B, Beretz L, Paya D, Ciobanu E, Blickle JF, Gourieux B, Andres E, Michel B. Abstract medication reconciliation at hospital admission: what about potential clinical impacts of unintentional discrepancies? A prospective study in an iInternal medicine department. Int J Clin Pharm 2012;34:202-02.

176 AbuYassin B, Aljadhey H, Al-Sultan M, Al-Rashed S, Adam M, Bates D. Accuracy of the medication history at admission to hospital in Saudi Arabia. Saudi Pharm J 2011;19:263-67.

177 Alassaad A, Gillespie U, Bertilsson M, Melhus H, Hammarlund-Udenaes M. Prescription and transcription errors in multidose-dispensed medications on discharge from hospital: an observational and interventional study. J Eval Clin Pract 2013;19:185-91.

178 Kalb K, Shalansky S, Legal M, Khan N, Ma I, Hunte G. Unintended medication discrepancies associated with reliance on prescription databases for medication reconciliation on admission to a general medical ward. Can J Hosp Pharm 2009;62:284-89.

179 Lau HS, Florax C, Porsius AJ, De Boer A. The completeness of medication histories in hospital medical records of patients admitted to general internal medicine wards. Brit J Clin Pharmaco 2000;49:597-603.

180 Lee YY, Kuo LN, Chiang YC, Hou JY, Wu TY, Hsu MH, Chen HY. Pharmacist-conducted medication reconciliation at hospital admission using information technology in Taiwan. Int J Med Inform 2013;82:522-27.

181 Pourrat X, Corneau H, Floch S, Kuzzay MP, Favard L, Rosset P, Hay N, Grassin J. Communication between community and hospital pharmacists: impact on medication reconciliation at admission. Int J Clin Pharm 2013;35:656-63.

341

182 Shepherd G, Schwartz R. Frequency of incomplete medication histories obtained at triage. Am J Health-Syst Ph 2009;66:65-69.

183 Wong JD, Bajcar JM, Wong GG, Alibhai SMH, Huh JH, Cesta A, Pond GR, Fernandes OA. Medication reconciliation at hospital discharge: Evaluating discrepancies. Ann Pharmacother 2008;42:1373-79.

184 Miller DF, Garcia D, Kreys T-J, Phan SV, VandenBerg A, Garrison KL. Evaluating the quality of performance of medication reconciliation on hospital admission. Hosp Pharm 2012;47:526-31.

185 Slee A, Farrar K, Hughes D, Constable S. Optimising medical treatment - How pharmacist-acquired medication histories have a positive impact on patient care. Pharm J 2006;277:737-39.

186 Boso-Ribelles V, Montero-Hernandez M, Font-Noguera I, Hernandez- Martin J, San Martin-Ciges E, Luis Poveda-Andres J. Evaluation of a plan for cardiology medication reconciliation on admission, and patient information at discharge, in a teaching hospital. Ejhp Pract 2011;17:26-30.

187 Bonnet-Zamponi D, d'Arailh L, Konrat C, Delpierre S, Lieberherr D, Lemaire A, Tubach F, Lacaille S, Legrain S. Drug-related readmissions to medical units of older adults discharged from acute geriatric units: results of the Optimization of Medication in AGEd multicenter randomized controlled trial. J Am Geriatr Soc 2013;61:113-21.

188 Fertleman M, Barnett N, Patel T. Improving medication management for patients: the effect of a pharmacist on post-admission ward rounds. Qual Saf Health Care 2005;14:207-11.

189 Climente-Martí M, García-Mañón ER, Artero-Mora A, Jiménez-Torres NV. Potential risk of medication discrepancies and reconciliation errors at admission and discharge from an inpatient medical service. Ann Pharmacother 2010;44:1747-54.

190 Tjia J, Bonner A, Briesacher BA, McGee S, Terrill E, Miller K. Medication discrepancies upon hospital to skilled nursing facility transitions. J Gen Intern Med 2009;24:630-35.

191 Quelennec B, Beretz L, Paya D, Ciobanu E, Blickle JF, Gourieux B, Andres E, Michel B. Medication Reconciliation at hospital admission: what about potential clinical impacts of unintentional discrepancies? A prospective study in an Internal Medicine Department. Int J Clin Pharm 2012;34:202-02.

192 Owen MC, Chang NM, Chong DH, Vawdrey DK. Evaluation of medication list completeness, safety, and annotations. Proc AMIA Symp 2011;2011:1055-61.

193 Mazer M, DeRoos F, Hollander J, McCusker C, Peacock N, Perrone J. Medication history taking in emergency department triage is inaccurate and incomplete. Acad Emerg Med 2011;18:102-04.

342

194 Lessard S, DeYoung K, Vazzana N. Medication discrepancies affecting senior patients at hospital admission. Am J Health-Syst Ph 2006;63:740- 43.

195 Grant MC, Larsen GY. Clinical information transfer and medication reconciliation in patients transferred from the pediatric intensive care unit. J Patient Saf 2007;3:195-99.

196 Forjuoh SN, Reis MD, Couchman GR, Symm B, Mason S, O'Banon R. Physician response to written feedback on a medication discrepancy found with their elderly ambulatory patients. J Am Geriatr Soc 2005;53:2173-77.

197 Coffey M, Mack L, Streitenberger K, Bishara T, De Faveri L, Matlow A. Prevalence and clinical significance of medication discrepancies at pediatric hospital admission. Academic Pediatrics 2009;9:360-65.

198 Sharma AN, Dvorkin R, Tucker V, Margulies J, Yens D, Rosalia A, Jr. Medical reconciliation in patients discharged from the Emergency department. J Emerg Med 2012;43:366-73.

199 Perren A, Previsdomini M, Cerutti B, Soldini D, Donghi D, Marone C. Omitted and unjustified medications in the discharge summary. Qual Saf Health Care 2009;18:205-08.

200 Geurts MME, van der Flier M, de Vries-Bots AMB, Brink-van der Wal TIC, de Gier JJ. Medication reconciliation to solve discrepancies in discharge documents after discharge from the hospital. Int J Clin Pharm 2013;35:600-07.

201 Larsen MD, Nielsen LP, Jeffery L, Staehr ME. [Medication errors on hospital admission]. Ugeskrift For Laeger 2006;168:2887-90.

202 Hellström LM, Bondesson Å, Höglund P, Eriksson T. Errors in medication history at hospital admission: prevalence and predicting factors. BMC Clinical Pharmacology 2012;12:9.

203 Terry D, Solanki G, Sinclair A, J M, K W. Medicines reconciliation on admission – don’t forget the children. Br J Clin Pharm 2009;1:22.

204 Irvine F, Almond A, Robertson S, Donaldson K, Flynn R, Isles C. Getting the right medicine: A survey of haemodialysis patients in SW Scotland. Pharm J 2009;283:575-76.

205 Campbell AJ, Bloomfield R, Noble DW. An observational study of changes to long-term medication after admission to an intensive care unit. Anaesthesia 2006;61:1087-92.

206 Vargas BR, Silveira ED, Caro TG, Vicedo TB. Medication discrepancies at emergency department. Pharm World Sci 2010;32:261-61.

207 Bridgeman P, Rynn K. Medication reconciliation in the emergency department. Am J Health-Syst Ph 2008;65:2325-26.

343

208 Carter M, Allin D, Scott L, Grauer D. Pharmacist-acquired medication histories in a university hospital emergency department. Am J Health-Syst Ph 2006;63:2500-03.

209 Walker-Cillo G, Fowler SB. Medication reconciliation practices and perceptions in the emergency department. Advanced Emergency Nursing Journal 2008;30:333-38.

210 Vasileff HM, Whitten LE, Pink JA, Goldsworthy SJ, Angley MT. The effect on medication errors of pharmacists charting medication in an emergency department. Pharm World Sci 2009;31:373-79.

211 Rothschild JM, Churchill W, Erickson A, Munz K, Schuur JD, Salzberg CA, Lewinski D, Shane R, Aazami R, Patka J, Jaggers R, Steffenhagen A, Rough S, Bates DW. Medication errors recovered by emergency department pharmacists. Ann Emerg Med 2010;55:513-21.

212 Manias E, Gerdtz MF, Weiland TJ, Collins M. Medication use across transition points from the emergency department: identifying factors associated with medication discrepancies. Ann Pharmacother 2009;43:1755-64.

213 Chan TC, Killeen JP, Castillo EM, Vilke GM, Guss DA. Impact of electronic medication reconciliation on triage times for patients seen in the emergency department. Ann Emerg Med 2007;50:225.

214 Cumbler E, Carter J, Kutner J. Failure at the transition of care: Challenges in the discharge of the vulnerable elderly patient. J Hosp Med 2008;3:349- 52.

215 Burda SA, Hobson D, Pronovost PJ. What is the patient really taking? Discrepancies between surgery and anesthesiology preoperative medication histories. Qual Saf Health Care 2005;14:414-16.

216 Elliott RA, Tim T, Taylor SE, Harvey PA, Belfrage MK, Jennings RJ, Marriott JL. Gaps in continuity of medication management during the transition from hospital to residential care: an observational study (MedGap Study). Australas J Ageing 2012;31:247-54.

217 Arora V, Kao J, Lovinger D, Seiden SC, Meltzer D. Medication discrepancies in resident sign-outs and their potential to harm. J Gen Intern Med 2007;22:1751-55.

218 Karkov LL, Schytte-Hansen S, Haugbølle LS. Discrepancies between sources providing the medication histories of acutely hospitalised patients. Pharm World Sci 2010;32:449-54.

219 Nickerson A, MacKinnon NJ, Roberts N, Saulnier L. Drug-therapy problems, inconsistencies and omissions identified during a medication reconciliation and seamless care service. Healthcare Quarterly (Toronto, Ont.) 2005;8:65-72.

344

220 Petty D, Zermansky A, Alldred D. The scale of repeat prescribing – time for an update. BMC Health Services Research 2014;14:76.

221 Turple J, MacKinnon NJ, Davis B. Frequency and type of medication discrepancies in one tertiary care hospital. Healthcare Quarterly (Toronto, Ont.) 2006;9:119-23.

222 Porter SC, Kohane IS, Goldmann DA. Improving data quality: A parent- completed computer interview to capture medication history. Pediatr Res 2004;55:120A-20A.

223 Eijsbroek H, Howell DCJ, Smith F, Shulman R. Medication issues experienced by patients and carers after discharge from the intensive care unit. J Crit Care 2013;28:46-50.

224 Barrett NA, Jones A, Whiteley C, Yassin S, McKenzie CA. Management of long-term hypothyroidism: a potential marker of quality of medicines reconciliation in the intensive care unit. Int J Pharm Pract 2012;20:303-06.

225 Bell CM, Bajcar J, Bierman AS, Li P, Mamdani MM, Urbach DR. Potentially unintended discontinuation of long-term medication use after elective surgical procedures. Arch Intern Med 2006;166:2525-31.

226 Ong SW, Fernandes OA, Cesta A, Bajcar JM. Drug-related problems on hospital admission: Relationship to medication information transfer. Ann Pharmacother 2006;40:408-13.

227 Terry D, Sinclair A. Prescribing for children at the interfaces of care. Arch Dis Childhood-E 2012;97:152-56.

228 Wong ICK, Basra N, Yeung VW, Cope J. Supply problems of unlicensed and off-label medicines after discharge. Arch Dis Child 2006;91:686-88.

229 Aljamal M, Ashcroft D, Tully M. How could indicators be used to assess the medicine reconciliation process? Int J Clin Pharm 2012;34:172-73.

230 Candlish K, Young G, Warholak T. Challenges in obtaining accurate home medication lists in emergency triage at an academic medical center: perceptions from nurses. Advanced Emergency Nursing Journal 2012;34:333-40.

231 Chevalier BAM, Parker DS, MacKinnon NJ, Sketris I. Nurses' perceptions of medication safety and medication reconciliation practices. Nurs Leadersh (Tor Ont) 2006;19:61-72.

232 Leape L, Bates D, Cullen D, Cooper J, Demonaco H, Gallivan T, Hallisey R, Ives J, Laird N, Laffel G. Systems analysis of adverse drug events. ADE Prevention Study Group. JAMA-J Am Med Assoc 1995;274:35-43.

233 FitzGerald RJ. Medication errors: the importance of an accurate drug history. Brit J Clin Pharmaco 2009;67:671-75.

345

234 Tam V, Knowles S, Cornish P, Fine N, Marchesano R, Etchells E. Frequency, type and clinical importance of medication history errors at admission to hospital: a systematic review. CMAJ 2005;173:510–5.

235 Moore P, Armitage G, Dobrzanski S. Medicines reconciliation using a shared electronic healthcare record. J Patient Saf 2011;7:147-53.

236 British National Formulary (BNF). British National Formulary. 66. London: BMJ Group & Pharmaceutical Press, September 2013.

237 Kent AJ, Harrington L, Skinner J. Medication reconciliation by a pharmacist in the emergency department: a pilot project. Can J Hosp Pharm 2009;62:238-42.

238 Quelennec B, Beretz L, Paya D, Blickle JF, Gourieux B, Andres E, Michel B. Potential clinical impact of medication discrepancies at hospital admission. Eur J Intern Med 2013;24:530-35.

239 Remtulla S, Brown G, Frighetto L. Best possible medication history by a pharmacy technician at a tertiary care hospital. Can J Hosp Pharm 2009;62:402-5.

240 Orrico KB. Sources and types of discrepancies between electronic medical records and actual outpatient medication use. J Manag Care Pharm 2008;14:626-31.

241 Schmiemann G, Bahr M, Gurjanov A, Hummers-Pradier E. Differences between patient medication records held by general practitioners and the drugs actually consumed by the patients. Int J Clin Pharm Th 2012;50:614-17.

242 Staroselsky M, Volk LA, Tsurikova R, Newmark LP, Lippincott M, Litvak I, Kittler A, Wang T, Wald J, Bates DW. An effort to improve electronic health record medication list accuracy between visits: patients' and physicians' response. Int J Med Inform 2008;77:153-60.

243 Tulner L, Kuper I, Frankfort S, van Campen J, Koks C, Brandjes D, Beijnen J. Discrepancies in reported drug use in geriatric outpatients: Relevance to adverse events and drug-drug interactions. Am J Geriatr Pharmac 2009;7:93-104.

244 Green CF, Burgul K, Armstrong DJ. A study of the use of medicine lists in medicines reconciliation: please remember this, a list is just a list. Int J Pharm Pract 2010;18:116-21.

245 Foss S, Schmidt JR, Andersen T, Rasmussen JJ, Damsgaard J, Schaefer K, Munck LK. Congruence on medication between patients and physicians involved in patient course. Eur J Clin Pharmacol 2004;59:841-47.

246 Turchin A, Gandhi TK, Coley CM, Shubina M, Broverman C. The use of electronic medication reconciliation to establish the predictors of validity of computerized medication records. Stud Health Technol Inform 2007;129:1022-26.

346

247 Price D, Cooke J, Singleton S, Feely M. Doctors' unawareness of the drugs their patients are taking: a major cause of overprescribing? Brit Med J 1986;292:99-100.

248 Department of Health. Pharmacy in the Future: implementing the NHS Plan., 2000. Accessed 234th April 2014. http://webarchive.nationalarchives.gov.uk/20130107105354/http:/www.dh. gov.uk/prod_consum_dh/groups/dh_digitalassets/@dh/@en/documents/di gitalasset/dh_4068204.pdf.

249 East & South East England Specialist Pharmacy Services. Moving Medicines Safely: Implementing a 'Green Bag' Scheme. 2010. Accessed 17th March 2014. http://www.medicinesresources.nhs.uk/en/Communities/NHS/SPS-E-and- SE-England/Meds-use-and-safety/QIPP/Transfer/Moving-Medicines- Safely-Implementing-and-sustaining-a-Green-Bag-Scheme--V2-/.

250 Ekedahl A, Brosius H, Jonsson J, Karlsson H, Yngvesson M. Discrepancies between the electronic medical record, the prescriptions in the Swedish national prescription repository and the current medication reported by patients. Pharmacoepidem Dr S 2011;20:1177-83.

251 Fung KW, Kayaalp M, Callaghan F, McDonald CJ. Comparison of electronic pharmacy prescription records with manually collected medication histories in an emergency department. Ann Emerg Med 2013;62:205-11.

252 Glintborg B, Poulsen HE, Dalhoff KP. The use of nationwide on-line prescription records improves the drug history in hospitalized patients. Brit J Clin Pharmaco 2008;65:265-69.

253 Glintborg B, Andersen S, Poulsen H. Prescription data improve the medication history in primary care. Qual Saf Health Care. 2010;19:164-8.

254 Glintborg B, Andersen SK, Poulsen HE. Prescription data improve the medication history in primary care. Qual Saf Health Care 2010;19:164-68.

255 Schalekamp T, Smit C, van Geest-Daalderop JHH, de Vries- Goldschmeding H, de Boer A. Discrepancies between medication records of anticoagulation clinics and pharmacy records. Pharmacoepidem Dr S 2006;15:823-28.

256 Balon J, Thomas SA. Comparison of hospital admission medication lists with primary care physician and outpatient pharmacy lists. J Nurs Scholarship 2011;43:292-300.

257 Olson MD, Tong GL, Steiner BD, Viera AJ, Ashkin E, Newton WP. Medication documentation in a primary care network serving North Carolina medicaid patients: results of a cross-sectional chart review. BMC Family Practice 2012;13.

347

258 Cockayne N, Duguid M, Shenfield G. Health professionals rarely record history of complementary and alternative medicines. Brit J Clin Pharmaco 2004;59:254–58.

259 Holmes HM, Kaiser K, Jackson S, McPherson ML. Soliciting an herbal medicine and supplement use history at hospice admission. J Palliat Med 2010;13:685-94.

260 Kinney MO. Mildly abnormal liver tests. Don't forget coeliac disease and drug history. Brit Med J 2010;341.

261 Hansen LO, Young RS, Hinami K, Leung A, Williams MV. Interventions to reduce 30-day rehospitalization: a systematic review. Ann Intern Med 2011;155:520-28.

262 Bayley B, Savitz L, Maddalone T, Stoner S, Hunt J, Wells R. Evaluation of patient care interventions and recommendations by a transitional care pharmacist. Ther Clin Risk Manag 2007;3:695-703.

263 Karnon J, Campbell F, Czoski-Murray C. Model-based cost-effectiveness analysis of interventions aimed at preventing medication error at hospital admission (medicines reconciliation). J Eval Clin Pract 2009;15:299-306.

264 Department of Health. The Operating Framework for the NHS in England 2011/12. 2010.

265 Hansen LO, Strater A, Smith L, Lee J, Press R, Ward N, Weigelt JA, Boling P, Williams MV. Hospital discharge documentation and risk of rehospitalisation. Qual Saf Health Care 2011;20:773-78.

266 Hellström LM, Höglund P, Bondesson Å, Petersson G, Eriksson T. Clinical implementation of systematic medication reconciliation and review as part of the Lund Integrated Medicines Management model - impact on all- cause emergency department revisits. J Clin Pharm Ther 2012;37:686-92.

267 de Blok C, Koster , Schilp J, Wagner C. Implementation of the Dutch National Patient Safety Programme. 2013. Accessed 3rd March 2014. http://www.nivel.nl/sites/default/files/bestanden/Samenvatting-engels- implementatie-VMS.pdf.

268 World Health Organisation (WHO). WHO High 5's project. 2009. Accessed 29th October 2013. https://www.high5s.org/bin/view/Main/WebHome.

269 Scotland Patient Safety Initiative. Accessed 4th May 2014. http://www.scottishpatientsafetyprogramme.scot.nhs.uk/.

270 Aljamal M, Ashcroft DM, Tully MP. Medicines reconciliation: developing quality indicators. Pharm World Sci 2010;32:233-34.

271 Cheng R, Yoo L, Ho C, Kadija M. Identification of medication safety indicators in acute care settings for public reporting in Ontario. Healthcare Quarterly (Toronto, Ont.) 2010;13:26-34.

348

272 Lesselroth BJ, Adams K, Tallett S, Wood S, Keeling A, Cheng K, Church V, Felder R, Tran H. Design of admission medication reconciliation technology: a human factors approach to requirements and prototyping. HERD 2013;6:30-48.

273 Monkman H, Borycki E, Kushniruk A, Kuo MH. Exploring the contextual and human factors of electronic medication reconciliation research: a scoping review. Stud Health Technol Inform 2013;194:166-72.

274 Brady A-M, Malone A-M, Fleming S. A literature review of the individual and systems factors that contribute to medication errors in nursing practice. J Nurs Manag 2009;17:679-97.

275 McLeod SE, Lum E, Mitchell C. Value of medication reconciliation in reducing medication errors on admission to hospital. J Pharm Pract Res 2008;38:196-99.

276 Royal Pharmaceutical Society of Great Britain (RPSGB). Moving Patients Medicines Safely: Guidance on Discharge and Transfer Planning. 2006. Accessed 21st March 2014. http://psnc.org.uk/wp- content/uploads/2013/07/Moving20Medicines20new1.pdf.

277 Ruder AD, Smith DL, Madsen MT, Kass FH. Is there a benefit to having a clinical oncology pharmacist on staff at a community oncology clinic? J Oncol Pharm Pract 2011;17:425-32.

278 Packard K, Herink M, Kuhlman P. Pharmacist's role in an Interdisciplinary Cardiac Rehabilitation Team. J Allied Health 2012;41:113-17.

279 Poon L, Lee A, Chiao T, Kang G, Heath S, Glass G. Pharmacist's role in a Parkinson's disease and movement disorders clinic. Am J Health-Syst Ph 2012;69:518-20.

280 Stock R, Scott J, Gurtel S. Using an electronic prescribing system to ensure accurate medication lists in a large multidisciplinary medical group. Jt Comm J Qual Patient Saf 2009;35:271-77.

281 Lees M, Chapman P, Dickson S. A process improvement strategy for patient safety. Healthcare Management Forum 2011;24:S25-33.

282 White CM, Schoettker PJ, Conway PH, Geiser M, Olivea J, Pruett R, Kotagal UR. Utilising improvement science methods to optimise medication reconciliation. Qual Saf Health Care 2011;20:372-80.

283 Lu Y, Clifford P, Bjorneby A, Thompson B, VanNorman S, Won K, Larsen K. Quality improvement through implementation of discharge order reconciliation. Am J Health-Syst Ph 2013;70:815-20.

284 Hummel J, Evans PC, Lee H. Medication reconciliation in the emergency department: opportunities for workflow redesign. Qual Saf Health Care 2010;19:531-35.

285 ECRI Institute. Health Care Improvement Foundation. Failure Mode and Effects Analysis: Medication Reconciliation. 2007. Accessed 13th April 349

2014. https://www.ecri.org/Documents/Patient_Safety_Center/PPC_Medication_ Reconciliation.pdf.

286 Allen T, Tseng S, Swanson K, M M. Improving the hospital discharge process with Six Sigma methods. Quality Engineering 2010;22:13-20.

287 Marvin V, Vaughan L, Kuo S, Bovill I, Clahrc. Improving medicines reconciliation and communication at transition of care to and from an acute medical unit. Pharmacoepidem Dr S 2012;21:795-95.

288 Garfield S, Barber N, Walley P, Willson A, Eliasson L. Quality of medication use in primary care - mapping the problem, working to a solution: a systematic review of the literature. BMC Medicine 2009;7:50.

289 Fernandes O. Medication reconciliation in the hospital: what, why, where, when, who and how? Healthcare Quarterly (Toronto, Ont.) 2012;15:42-49.

290 National Health Service (NHS). Main procedures and interventions: summary 2009-10 2010. Accessed November 2010. http://www.hesonline.nhs.uk/Ease/servlet/ContentServer?siteID=1937&cat egoryID=204.

291 Sagripanti M, Dean B, Barber N. An evaluation of the process-related medication risks for elective surgery patients from pre-operative assessment to discharge. Int J Pharm Pract 2002;10:161-70.

292 Bedard P, Tardif L, Ferland A, Bussieres JF, Lebel D, Bailey B, Girard M, Lachaine J. A medication reconciliation form and its impact on the medical record in a paediatric hospital. J Eval Clin Pract 2010;17:222-27.

293 Poole DL, Chainakul JN, Pearson M, Graham L. Medication reconciliation: a necessity in promoting a safe hospital discharge. J Healthc Qual 2006;28:12-19.

294 Tattersall C, Hartry N, Ram AR. Improving the accuracy of medication history taking in a diabetic eye clinic. British Journal of Nursing (BJN) 2008;17:598-600.

295 De Winter S, Vanbrabant P, Spriet I, Desruelles D, Indevuyst C, Knockaert D, Gillet JB, Willems L. A simple tool to improve medication reconciliation at the emergency department. Eur J Intern Med 2011;22:382-85.

296 Brady D, BD. F. An evaluation of the contribution of the medical admissions pharmacist at a London teaching hospital. Int J Pharm Pract 2004;12:1-6.

297 Neufeld NJ, Fernandez MG, Christo PJ, Williams KA. Positive recognition program increases compliance with medication reconciliation by resident physicians in an outpatient clinic. Am J Med Qual 2013;28:40-45.

298 Nassaralla CL, Naessens JM, Hunt VL, Bhagra A, Chaudhry R, Hansen MA, Tulledge-Scheitel SM. Medication reconciliation in ambulatory care: attempts at improvement. Qual Saf Health Care 2009;18:402-07. 350

299 Maxson PM, Derby KM, Wrobleski DM, Foss DM. Bedside nurse-to-nurse handoff promotes patient safety. Medsurg Nurs 2012;21:140-45.

300 Beckett RD, Crank CW, Wehmeyer A. Effectiveness and feasibility of pharmacist-led admission medication reconciliation for geriatric patients. J Pharm Pract 2012;25:136-41.

301 Lizer M, Brackbill M. Medication history reconciliation by pharmacists in an inpatient behavioral health unit. Am J Health-Syst Ph 2007;64:1087-91.

302 De Winter S, Spriet I, Indevuyst C, Vanbrabant P, Desruelles D, Sabbe M, Gillet JB, Wilmer A, Willems L. Pharmacist- versus physician-acquired medication history: a prospective study at the emergency department. Qual Saf Health Care 2010;19:371-75.

303 Crook M, Ajdukovic M, Angley C, Soulsby N, Doecke C, Stupans I, Angley M. Eliciting comprehensive medication histories in the emergency department: the role of the pharmacist. Pharmacy Practice 2007;5:78-84.

304 Cawthon C, Walia S, Osborn CY, Niesner KJ, Schnipper JL, Kripalani S. Improving care transitions: The patient perspective. J Health Commun 2012;17:312-24.

305 Steurbaut S, Leemans L, Leysen T, De Baere E, Cornu P, Mets T, Dupont AG. Medication history reconciliation by clinical pharmacists in elderly inpatients admitted from home or a nursing home. Ann Pharmacother 2010;44:1596-603.

306 Hayes B, Donovan J, Smith B, Hartman C. Pharmacist-conducted medication reconciliation in an emergency department. Am J Health-Syst Ph 2007;64:1720-23.

307 Nester T, Hale L. Effectiveness of a pharmacist-acquired medication history in promoting patient safety. Am J Health-Syst Ph 2002;59:2221-25.

308 Ajdukovic M, Crook M, Angley C, Stupans I, Soulsby N, Doecke C, Anderson B, Angley M. Pharmacist elicited medication histories in the Emergency Department: identifying patient groups at risk of medication misadventure. Pharmacy Practice 2007;5:162-68.

309 Reeder T, Mutnick A. Pharmacist-versus physician-obtained medication histories. Am J Health-Syst Ph 2008;65:857-60.

310 Chantelois E, Suzuki N. A pilot program comparing physician- and pharmacist-ordered discharge medications at a Veterans Affairs medical center. Am J Health-Syst Ph 2003;60:1652-56.

311 Yusuff KB, Tayo F, Aina BA. Pharmacists' participation in the documentation of medication history in a developing setting: an exploratory assessment with new criteria. Pharmacy Practice 2010;8:139- 45.

351

312 Walker PC, Bernstein SJ, Jones JNT, Piersma J, Kim H-W, Regal RE, Kuhn L, Flanders SA. Impact of a pharmacist-facilitated hospital discharge program: a quasi-experimental study. Arch Intern Med 2009;169:2003-10.

313 Stewart AL, Lynch KJ. Identifying discrepancies in electronic medical records through pharmacist medication reconciliation. J Am Pharm Assoc 2012;52:59-66.

314 Becerra-Camargo J, Martinez-Martinez F, Garcia-Jimenez E. A multicentre, double-blind, randomised, controlled, parallel-group study of the effectiveness of a pharmacist-acquired medication history in an emergency department. BMC Health Services Research 2013;13.

315 van den Bemt P, van der Schrieck-de Loos EM, van der Linden C, Theeuwes A, Pol AG. Effect of medication reconciliation on unintentional medication discrepancies in acute hospital admissions of elderly adults: A multicenter study. J Am Geriatr Soc 2013;61:1262-68.

316 Musgrave CR, Pilch NA, Taber DJ, Meadows HB, McGillicuddy JW, Chavin KD, Baliga PK. Improving transplant patient safety through pharmacist discharge medication reconciliation. Am J Transplant 2013;13:796-801.

317 Maldonado A, Weeks D, Bitterman A, McCleary J, Seiger T, Carson R, Tabon P, Goroski A, Ojogho O. Changing transplant recipient education and inpatient transplant pharmacy practices: A single-center perspective. Am J Health-Syst Ph 2013;70:900-04.

318 Mortimer C, Emmerton L, Lum E. The impact of an aged care pharmacist in a department of emergency medicine. J Eval Clin Pract 2011;17:478-85.

319 Ledger S, Choma G. Medication reconciliation in hemodialysis patients. Cannt J 2008;18:41-3.

320 Stone BL, Boehme S, Mundorff MB, Maloney CG, Srivastava R. Hospital admission medication reconciliation in medically complex children: an observational study. Arch Dis Child 2010;95:250-55.

321 George LJW, Senturk-Raif R, Hodgkinson MR, Emmerton M, Larmour I. Impact of a surgical preadmission clinic pharmacist on the quality of medication management from preadmission to discharge: a randomised controlled study. J Pharm Pract Res 2011;41:212-16.

322 van den Bemt P, van den Broek S, van Nunen AK, Harbers JBM, Lenderink AW. Medication reconciliation performed by pharmacy technicians at the time of preoperative screening. Ann Pharmacother 2009;43:868-74.

323 Fit KE. Documentation of student pharmacists' interventions from medication histories in a primary care setting. J Pharm Tech 2008;24:330- 35.

352

324 Mersfelder T, Bickel R. Inpatient medication history verification by pharmacy students. Am J Health-Syst Ph 2008;65:2273-75.

325 Lubowski TJ, Cronin LM, Pavelka RW, Briscoe-Dwyer LA, Briceland LL, Hamilton RA. Effectiveness of a medication reconciliation project conducted by PharmD students. Am J Pharm Educ 2007;71.

326 Andrus MR. Student pharmacist initiated medication reconciliation in the outpatient setting. Pharmacy Practice 2012;10:78-82.

327 Tompson AJ, Peterson GM, Jackson SL, Hughes JD, Raymond K. Utilizing community pharmacy dispensing records to disclose errors in hospital admission drug charts. Int J Clin Pharm Th 2012;50:639-46.

328 Lindquist LA, Gleason KM, McDaniel MR, Doeksen A, Liss D. Teaching medication reconciliation through simulation: a patient safety initiative for second year medical students. J Gen Intern Med 2008;23:998-1001.

329 Chan AHY, Garratt E, Lawrence B, Turnbull N, Pratapsingh P, Black PN. Effect of education on the recording of medicines on admission to hospital. J Gen Intern Med 2010;25:537-42.

330 Roulet L, Asseray N, Foucher N, Potel G, Lapeyre-Mestre M, Ballereau F. A questionnaire to document self-medication history in adult patients visiting emergency departments. Pharmacoepidem Dr S 2013;22:151-9.

331 Hershey C, Grant B. Controlled trial of a patient completed history questionnaire: Effects on quality of documentation and patient and physician satisfaction. Am J Med Qual 2002;17:126-35.

332 Lingaratnam S, Aranda S, Pearce T, Kirsa S. A controlled before and after study to evaluate a patient and health professional partnership model towards effective medication reconciliation. J Oncol Pharm Pract 2013;19:48-56.

333 Chan EW, Taylor SE, Marriott J, Barger B. An intervention to encourage ambulance paramedics to bring patients' own medications to the ED: Impact on medications brought in and prescribing errors. Emerg Med Australas 2010;22:151-58.

334 Chan EW, Taylor SE, Marriott JL, Barger B. Bringing patients' own medications into an emergency department by ambulance: effect on prescribing accuracy when these patients are admitted to hospital. Med J Aust 2009;191:374-77.

335 Weingart SN, Cleary A, Seger A, Eng TK, Saadeh M, Gross A, Shulman LN. Medication reconciliation in ambulatory oncology. Jt Comm J Qual Patient Saf 2007;33:750-7.

336 Ziaeian B, Araujo KLB, Van Ness PH, Horwitz LI. Medication reconciliation accuracy and patient understanding of intended medication changes on hospital discharge. J Gen Intern Med 2012;27:1513-20.

353

337 Yusuff KB, Tayo F, Aina BA. What probable factors underlie inadequacy of medication history information recorded in patients' case notes by physicians in a Nigerian teaching hospital? The Int J Pharm Pract 2009;17:317-21.

338 Ross S, Loke Y. Development of learning outcomes for an undergraduate prescribing curriculum. Brit J Clin Pharmaco 2010;70:604-08.

339 American Medical Association (AMA). The physicians role in medication reconciliation: issues, strategies and safety principles. 2007. Accessed 24th October 2012. http://bcpsqc.ca/documents/2012/09/AMA-The- physician%E2%80%99s-role-in-Medication-Reconciliation.pdf.

340 ASHP Council Pharmacy Practice. ASHP statement on the pharmacist's role in medication reconciliation. Am J Health-Syst Ph 2013;70:453-6.

341 Frei P, Huber LC, Simon RW, Bonani M, Lüscher TF. Insufficient medication documentation at hospital admission of cardiac patients: a challenge for medication reconciliation. J Cardiovasc Pharm 2009;54:497- 501.

342 Grimes T, Duggan C, Delaney T. Pharmacy services at admission and discharge in adult, acute, public hospitals in Ireland. The Int J Pharm Pract 2010;18:346-52.

343 Pedersen C, Schneider P, Scheckelhoff D. ASHP national survey of pharmacy practice in hospital settings: Prescribing and transcribing - 2007. Am J Health-Syst Ph 2008;65:827-43.

344 deClifford J-M, Blewitt P, Lam SS, Leung BK. How do clinical pharmacists spend their working day? A time-and-motion study. J Pharm Pract Res 2012;42:134-39.

345 Higham C. Drug history taking - a role for the ward pharmacist. Pharm J 1982:302-05.

346 Galvin M, Grimes T, Fitzsimons M, Jago-Byrne MC. Clinical pharmacy input into the admission medication reconciliation process in Ireland. Pharmacoepidem Dr S 2010;19:07.

347 Miranda TMM, Petriccione S, Ferracini FT, Borges Filho WM. Interventions performed by the clinical pharmacist in the emergency department. Einstein (São Paulo, Brazil) 2012;10:74-78.

348 Nana B, Sun L-S, Allen G. Initiation of an emergency department pharmacy program during economically challenging times. Am J Health- Syst Ph 2012;69:1682-86.

349 Hatch J, Becker T, Fish JT. Difference between pharmacist-obtained and physician-obtained medication histories in the intensive care unit. Hosp Pharm 2011;46:262-68.

354

350 Steurbaut S, Leysen T, De Baere E, Lambert M, Leemans L, Mets T, Dupont AG. Medication reconciliation by clinical pharmacists at a university hospital geriatric ward. Pharm World Sci 2009;31:260-60.

351 Aldridge V, Park H, Bounthavong M, Morreale A. Implementing a comprehensive, 24-hour emergency department pharmacy program. Am J Health-Syst Ph 2009;66:1943-7.

352 Randolph T. Expansion of pharmacists' responsibilities in an emergency department. Am J Health-Syst Ph 2009;66:1484-7.

353 Collignon U, Oborne CA, Kostrzewski A. Pharmacy services to UK emergency departments: a descriptive study. Pharm World Sci 2010;32:90-96.

354 Latter S, Smith A, Blenkinsopp A, Nicholls P, Little P, Chapman S. Are nurse and pharmacist independent prescribers making clinically appropriate prescribing decisions? An analysis of consultations. J Health Serv Res Policy 2012;17:149-56.

355 Haggerty K, Brechin W, Elfellah M, Haxton J. Audit to evaluate the Role of a Clinical Pharmacy Technician in providing Pharmaceutical Care at Ward Level at Aberdeen Royal Infirmary. 2004. Accessed 27th January 2014. http://www.hi- netgrampian.org/hinet/file/527/Pharmacy%20Technician_Final%20Report %20June%202004.pdf.

356 Knight H, Edgerton L, Foster R. Pharmacy technicians obtaining medication histories within the emergency department. Am J Health-Syst Ph 2010;67:512-14.

357 Turner SJ, Lam SS, Leung BK, Toh CT. Establishing the role of a clinical pharmacy technician. J Pharm Pract Res 2005;35:119-21.

358 Daniel M, Banerjee AR. Is a doctor needed in the adult ENT pre-admission clinic? J Laryngol Otol 2004;118:796-98.

359 MacLaren A. Medicines reconciliation in hospital. Good Practice Guidance. NHS Greater Glasgow and Clyde, 2012. Accessed 25th April 2014.

360 Urban R, Paloumpi E, Rana N, Morgan J. Communicating medication changes to community pharmacy post-discharge: the good, the bad, and the improvements. Int J Clin Pharm 2013;35:813-20.

361 Mistry R. How a green bag can keep people safe. Pharm J 2011;287:163.

362 Hobson RJ, Sewell GJ. A national survey of pharmacist transcribing of discharge prescriptions. Int J Pharm Pract 2003;11:89-95.

363 Callen J, McIntosh J, Li J. Accuracy of medication documentation in hospital discharge summaries: A retrospective analysis of medication transcription errors in manual and electronic discharge summaries. Int J Med Inform 2010;79:58-64. 355

364 Nursing and Midwifery Council (NMC). Standards for medicines management. 2010. Accessed August 2011. http://www.nmc- uk.org/Documents/Standards/nmcStandardsForMedicinesManagementBo oklet.pdf.

365 Patel R, Butler K, Garrett D, Badger N, Cheoun D, Hallman L. Impact of pharmacist's participation on hospitalists' rounds. Hosp Pharm 2010;45:129-34.

366 Sexton J, Ho YJ, Green CF, Caldwell NA. Ensuring seamless care at hospital discharge: a national survey. J Clin Pharm Ther 2000;25:385-93.

367 Medicines Healthcare Regulatory Agency (MHRA). Nurse and pharmacist independent prescribing. 2006. Accessed 23rd September 2011 2011. http://www.mhra.gov.uk/Howweregulate/Medicines/Availabilityprescribings ellingandsupplyingofmedicines/ExemptionsfromMedicinesActrestrictions/N urseandpharmacistindependentprescribing/index.htm.

368 Parkin DM, Henney CR, Quirk J, Crooks J. Deviation from prescribed drug treatment after discharge from hospital. Brit Med J 1976;2:686-88.

369 Mottram D, Slater S, West P. Hospital discharge correspondence - how effective is it? Int J Pharm Pract 1994;3:24-26.

370 Kendrick AR, Hindmarsh DJ. Which type of hospital discharge report reaches general practitioners most quickly? Brit Med J 1989;298:362-63.

371 Munday A, Kelly B, Forrester JW, Timoney A, McGovern E. Do general practitioners and community pharmacists want information on the reasons for drug therapy changes implemented by secondary care? Brit J Gen Pract 1997;47:563-66.

372 Kripalani S, LeFevre F, Phillips CO, Williams MV, Basaviah P, Baker DW. Deficits in communication and information transfer between hospital-based and primary care physicians - Implications for patient safety and continuity of care. JAMA-J Am Med Assoc 2007;297:831-41.

373 Coleman A, Ricketts L, Teal S, Mitchell E, Silcock J, Wright DJ. An audit of hospital discharge and outpatient information from a primary care perspective. Journal of Social and Administrative Pharmacy 2001;18:226- 31.

374 Committee AGSHCS. Improving the quality of transitional care for persons with complex care needs. J Am Geriatr Soc 2003;51:556-7.

375 Corbett CF, Setter SM, Daratha KB, Neumiller JJ, Wood LD. Nurse identified hospital to home medication discrepancies: implications for improving transitional care. Geriatr Nurs 2010;31:188-96.

376 Coleman E, Parry C, Chalmers S, Min S. The care transitions intervention: results of a randomized controlled trial. Arch Intern Med 2006 166:1822-8.

377 Laugaland K, Aase K, Barach P. Interventions to improve patient safety in transitional care - a review of the evidence. Work 2012;41:2915-24. 356

378 Hesselink G, Schoonhoven L, Barach P, Spijker A, Gademan P, Kalkman C, Liefers J, Vernooij-Dassen M, Wollersheim H. Improving patient handovers from hospital to primary care a systematic review. Ann Intern Med 2012;157:417-28.

379 Cochrane RA, Mandal AR, Ledger-Scott M, Walker R. Changes in drug treatment after discharge from hospital in geriatric patients. Brit Med J 1992;305:694-96.

380 Akram U, Urban R. An evaluation in to the efficiency of the discharge process from secondary to primary care. Int J Pharm Pract 2012;20 (Suppl. 2):31–101.

381 Paquette-Lamontagne N, McLean WM, Besse L, Cusson J. Evaluation of a new integrated discharge prescription form. Ann Pharmacother 2001;35:953-58.

382 Chua D, Chu E, Lo A, Lo M, Pataky F, Tang L, Bains A. Effect of misalignment between hospital and provincial formularies on medication discrepancies at discharge: PPITS (proton pump inhibitor therapeutic substitution) study. Can J Hosp Pharm 2012;65:98-102.

383 Louis-Simonet M, Kossovsky MP, Sarasin FP, Chopard P, Gabriel V, Perneger TV, Gaspoz J-M. Effects of a structured patient-centered discharge interview on patients' knowledge about their medications. Am J Med 2004;117:563-68.

384 McLeod M, Ahmed Z, Barber N, Franklin B. A national survey of inpatient medication systems in English NHS hospitals. BMC Health Services Research 2014;14:93.

385 Ahmed Z, McLeod M, Barber N, Jacklin A, Franklin B. The use and functionality of electronic prescribing systems in English Acute NHS Trusts: A cross-sectional survey. PLoS One 2013;8.

386 Delate T, Chester EA, Stubbings TW, Barnes CA. Clinical outcomes of a home-based medication reconciliation program after discharge from a skilled nursing facility. Pharmacotherapy 2008;28:444-52.

387 Kripalani S, Roumie CL, Dalal AK, Cawthon C, Businger A, Eden SK, Shintani A, Sponsler KC, Harris LJ, Theobald C, Huang RL, Scheurer D, Hunt S, Jacobson TA, Rask KJ, Vaccarino V, Gandhi TK, Bates DW, Williams MV, Schnipper JL, Low P-CPI. Effect of a pharmacist intervention on clinically important medication errors after hospital discharge a randomized trial. Ann Intern Med 2012;157:1-10.

388 Eggink R, Lenderink A, Widdershoven J, van den Bemt P. The effect of a clinical pharmacist discharge service on medication discrepancies in patients with heart failure. Pharm World Sci 2010;32:759-66.

389 Boockvar K, Carlson LaCorte H, Giambanco V, Fridman B, Siu A. Medication reconciliation for reducing drug-discrepancy adverse events. Am J Geriatr Pharmac 2006;4:236-43.

357

390 Voirol P, Kayser SR, Chang CY, Chang QL, Youmans SL. Impact of pharmacists' interventions on the pediatric discharge medication process. Ann Pharmacother 2004;38:1597-602.

391 Costa LL, Poe SS, Lee MC. Challenges in posthospital care: nurses as coaches for medication management. J Nurs Care Qual 2011;26:243-51.

392 Cromarty E, Downie G, Wilkinson S, J C. Communication regarding the discharge medicines of elderly patients: a controlled trial. Pharm J 1998;260:62-64.

393 Cattell R, Conroy C, Sheikh A. Pharmacist integration into the discharge process: a qualitative and quantitative impact assessment. Int J Pharm Pract 2001;9:59-64.

394 Bowden J. Reissuing patients' medicines - a step to seamless care. Pharm J 1993;251:356-57.

395 Smith L, McGowan L, Moss-Barclay C, Wheater J, Knass D, Chrystyn H. An investigation of hospital generated pharmaceutical care when patients are discharged home from hospital. Brit J Clin Pharmaco 1997;44:163-65.

396 Milliken B, Rea Y. A clinical pharmacist's contribution to discharge planning and patient counselling. Pharm J 1997;259:777-78.

397 Binyon D. Pharmaceutical Care: its impact on patient care and the hospital-community interface. Pharm J 1994;253:344-49.

398 Lalonde L, Lampron A, Vanier M, Levasseur P, Khaddag R, Chaar N. Effectiveness of a medication discharge plan for transitions of care from hospital to outpatient settings. Am J Health-Syst Ph 2008;65:1451-57.

399 Eadon H. Use of pharmacy discharge information for transplant patients Pharm J 1994;253:314-16.

400 Sandler DA, Mitchell JR, Fellows A, Garner ST. Is an information booklet for patients leaving hospital helpful and useful? Brit Med J 1989;298:870- 74.

401 Sandler DA, Heaton C, Garner ST, Mitchell JR. Patients' and general practitioners' satisfaction with information given on discharge from hospital: audit of a new information card. Brit Med J 1989;299:1511-13.

402 Chae SY, Chae MH, Isaacson N, James TS. The patient medication list: Can we get patients more involved in their medical care? J Am Board Fam Med 2009;22:677-85.

403 Jack BW, Chetty VK, Anthony D, Greenwald JL, Sanchez GM, Johnson AE, Forsythe SR, O'Donnell JK, Paasche-Orlow MK, Manasseh C, Martin S, Culpepper L. A reengineered hospital discharge program to decrease rehospitalization: a randomized trial. Ann Intern Med 2009;150:178-87.

404 Kornburger C, Gibson C, Sadowski S, Maletta K, Klingbeil C. Using "Teach-Back" to promote a safe transition from hospital to home: An 358

evidence-based approach to improving the discharge process. J Pediatr Nurs 2013;28:282-91.

405 Karapinar-Carkit F, Borgsteede SD, Zoer J, Smit HJ, Egberts ACG, van den Bemt P. Effect of medication reconciliation with and without patient counseling on the number of pharmaceutical interventions among patients discharged from the hospital. Ann Pharmacother 2009;43:1001-10.

406 Wilkinson ST, Pal A, Couldry RJ. Impacting readmission rates and patient satisfaction: results of a discharge pharmacist pilot program. Hosp Pharm 2011;46:876-83.

407 Chinthammit C, Armstrong EP, Warholak TL. A cost-effectiveness evaluation of hospital discharge counseling by pharmacists. J Pharm Pract 2012;25:201-08.

408 deClifford J, Lam SS, Leung BK. Evaluation of a pharmacist-initiated e- script transcription service for discharged patients. J Pharm Pract Res 2009;39:39-42.

409 DeClifford J, Lam S, Leung B. Implementation and acceptance of a pharmacist-initiated e-script transcription service for discharge prescriptions. J Pharm Pract Res 2009;39:286-89.

410 Beauchesne M-F, Nenciu LM, Dinh T-H, Tasse M, Fillion A, Labrecque M, Blais L. Active communication of a pharmacy discharge plan for patients with respiratory diseases: A pilot study. J Pharm Tech 2007;23:67-74.

411 Tran T, Elliott RA, Taylor SE, Garrett K. Development and evaluation of a hospital pharmacy generated interim residential care medication administration chart. J Pharm Pract Res 2012;42:100-05.

412 Elliott RA, Tran T, Taylor SE, Harvey PA, Belfrage MK, Jennings RJ, Marriott JL. Impact of a pharmacist-prepared interim residential care medication administration chart on gaps in continuity of medication management after discharge from hospital to residential care: a prospective pre- and post-intervention study (MedGap Study). BMJ Open 2012;2.

413 Setter S, Corbett C, Neumiller J, Gates B, Sclar D, Sonnett T. Effectiveness of a pharmacist-nurse intervention on resolving medication discrepancies for patients transitioning from hospital to home health care. Am J Health-Syst Ph 2009;66:2027-31.

414 Crotty M, Rowett D, Spurling L, Giles L, Phillips P. Does the addition of a pharmacist transition coordinator improve evidence-based medication management and health outcomes in older adults moving from the hospital to a long-term care facility? Results of a randomized, controlled trial. Am J Geriatr Pharmac 2004;2:257-64.

415 Walker PC, Tucker Jones JN, Mason NA. An advanced pharmacy practice experience in transitional care. Am J Pharm Educ 2010;74:20-20.

359

416 Hugtenburg JG, Borgsteede SD, Beckeringh JJ. Medication review and patient counselling at discharge from the hospital by community pharmacists. Pharm World Sci 2009;31:630-37.

417 Paulino EI, Bouvy ML, Gastelurrutia MA, Guerreiro M, Buurma H. Drug related problems identified by European community pharmacists in patients discharged from hospital. Pharm World Sci 2004;26:353-60.

418 Johnson CM, Marcy TR, Harrison DL, Young RE, Stevens EL, Shadid J. Medication reconciliation in a community pharmacy setting. J Am Pharm Assoc 2010;50:523-26.

419 Kilcup M, Schultz D, Carlson J, Wilson B. Postdischarge pharmacist medication reconciliation: Impact on readmission rates and financial savings. J Am Pharm Assoc 2013;53:78-84.

420 Wilson B. Are patients provided with sufficient information about their medicines on discharge from hospital? Pharm World Sci 2008;30:956- 1013.

421 Grimes T, Duggan C, Gallagher P, Strawbridge J. Care of the stroke patient-communication between the community pharmacist and prescribers in the Republic of Ireland. Pharm World Sci 2009;31:648-55.

422 Brown RL. The home health model: reducing hospitalizations by improving medication reconciliation and communication. J Ark Med Soc 2009;105:204-05.

423 Al-Rashid A, Wright D, Reeves J, Chrystyn H. Opinions about hospital discharge information in the UK Part 1: General Practitioners. J Soc Admin Pharm 2001;18:91-96.

424 Pegrum S. Seamless care: The need for communication between hospital and comunity pharmacists. Pharm J 1995;254:445-46.

425 Cairns C. Pharmacist admission and discharge checklists: how should they be used? Pharm J 1994;253:770-73.

426 Karaksha AA, Hattingh HL, Hall T. Quality of discharge summaries sent by a regional hospital to general practitioners. J Pharm Pract Res 2010;40:208-10.

427 Karapinar F, van den Bemt PMLA, Zoer J, Nijpels G, Borgsteede SD. Informational needs of general practitioners regarding discharge medication: content, timing and pharmacotherapeutic advices. Pharm World Sci 2010;32:172-78.

428 Roughead EE, Kalisch LM, Ramsay EN, Ryan P, Gilbert AL. Continuity of care: when do patients visit community healthcare providers after leaving hospital? Intern Med J 2011;41:662-67.

429 Crowe S, Tully M, Cantrill J. Information in general medical practices: the information processing model. Fam Pract 2009;27:230-36.

360

430 Burns J, Sneddon I, Lovell M, McLean A, Martin B. Elderly patients and their medication: A post-discharge follow-up study. Age Ageing 1992;21:178-81.

431 Rahmner PB, Gustafsson LL, Holmstrom I, Rosenqvist U, Tomson G. Whose job is it anyway? Swedish general practitioners' perception of their responsibility for the patient's drug List. Ann Fam Med 2010;8:40-46.

432 Duggan C, Bates I, Hough J. Discrepancies in prescribing - where do they occur? Pharm J 1996;256:65-67.

433 Duffin J, Norwood J, Blenkinsopp A. An investigation into medication changes initiated in general practice after patients are discharged from hospital. Pharm J 1998;261:R32.

434 Bayoumi I, Howard M, Holbrook AM, Schabort I. Interventions to improve medication reconciliation in primary care. Ann Pharmacother 2009;43:1667-75.

435 Livingstone C. Onwards and upwards with target MURs. Pharm J 2010;284:57-60.

436 PSNC. National target groups for MURs. 2014. Accessed 3rd May 2014. http://psnc.org.uk/services-commissioning/advanced- services/murs/national-target-groups-for-murs/.

437 Royal Pharmaceutical Society (RPS). Help get the right medicines when you move care providers. Royal Pharmaceutical Society, 2011. Accessed 17th March 2011. http://www.rpharms.com/medicines-safety/getting-the- medicines-right.asp.

438 Dicomidis J. Discharge medicines review service in Wales: Enhancing the role of community pharmacies. Pharm Manage 2012;28:17-21.

439 Khideja N. Domicilliary Medicines Use Review (MUR) bridging the interface between secondary and primary care. Pharmacoepidem Dr S 2012;21:794-95.

440 Carter SR, Chen TF, White L. Home medicines reviews: a quantitative study of the views of recipients and eligible non-recipients. Int J Pharm Pract 2012;20:209-17.

441 Ponniah A, Shakib S, Doecke CJ, Boyce M, Angley M. Post-discharge medication reviews for patients with heart failure: a pilot study. Pharm World Sci 2008;30:810-15.

442 Akwagyriam I, Goodyer L, Harding L, Khakoo S, Millington H. Drug history taking and the identification of drug related problems in an accident and emergency department. J Accid Emerg Med 1996;13:166-8.

443 Royal Pharmaceutical Society of Great Britain (RPSGB). Checklists for admission and discharge. Pharm J 1993;250:HS6.

361

444 Nicholls I, Wilcock M. A primary care initiative to support medicines reconciliation. Prescriber 2010;21:35-38.

445 Royal Pharmaceutical Society (RPS). Keeping patients safe when they transfer between care providers – getting the medicines right. A guide for all providers and commissioners of NHS services. 2011. Accessed 23rd May 2014. http://www.rpharms.com/getting-the-medicines- right/organisational-guidance.asp.

446 Gauthier J. Audit of discharge summaries in primary care. Int J Pharm Pract 2010;18:11-12.

447 Bell D, Morgan R. Repairing careless seams: preventing the fall-out between hospital and intermediate care. Pharmacoepidem Dr S 2012;21:788-98.

448 Foster DS, Paterson C, Fairfield G. Evaluation of immediate discharge documents--room for improvement? Scott Med J 2002;47:77-79.

449 Grimes T, Delaney T, Duggan C, Kelly JG, Graham IM. Survey of medication documentation at hospital discharge: implications for patient safety and continuity of care. Ir J Med Sci 2008;177:93-97.

450 Kind A, Anderson P, Hind J, Robbins J, Smith M. Omission of dysphagia therapies in hospital discharge communications. Dysphagia 2011;26:49- 61.

451 Legault K, Ostro J, Khalid Z, Wasi P, You JJ. Quality of discharge summaries prepared by first year internal medicine residents. BMC Medical Education 2012;12.

452 Patel K, Wright DJ, Desborough J. Effectiveness of electronic discharge and handover system at Basildon and Thurrock University Hospitals NHS Foundation Trust: A service evaluation. Int J Pharm Pract 2011;19.

453 Zamora Z, McCall B, Patel L, Biese K, LaMantia M, Platts-Mills T, Naus N, Jerkewitz HP, Cairns CB, Busby-Whitehead J, Kizer JS. Implementation of a web-based system to improve the transitional care of older adults. J Nurs Care Qual 2012;27:182-89.

454 Moen A, Andersen SK, Aarts J, Hurlen P, Heleen, Rootjes I. The shift in workarounds upon implementation of computerized physician order entry. Stud Health Technol Inform 2011;169:290-94.

455 Kuperman GJ, Teich JM, Gandhi TK, Bates DW. Patient safety and computerized medication ordering at Brigham and Women's Hospital. Jt Comm J Qual Improv 2001;27:509-21.

456 Lee JY, Leblanc K, Fernandes OA, Huh J-H, Wong GG, Hamandi B, Lazar NM, Morra D, Bajcar JM, Harrison J. Medication reconciliation during internal hospital transfer and impact of computerized prescriber order entry. Ann Pharmacother 2010;44:1887-95.

362

457 Lesselroth BJ, Felder RS, Adams SM, Cauthers PD, Dorr DA, Wong GJ, Douglas DM. Design and implementation of a medication reconciliation kiosk: the Automated Patient History Intake Device (APHID). J Am Med Inform Assn 2009;16:300-04.

458 Lesselroth B, Adams S, Felder R, Dorr DA, Cauthers P, Church V, Douglas D. Using consumer-based kiosk technology to improve and standardize medication reconciliation in a specialty care setting. Jt Comm J Qual Patient Saf 2009;35:264-70.

459 Salemi CS, Singleton N. Decreasing medication discrepancies between outpatient and inpatient care through the use of computerized pharmacy data. Perm J 2007;11:31-4.

460 Grimes T, Fitzsimons M, Galvin M, Delaney T. Relative accuracy and availability of an Irish National Database of dispensed medication as a source of medication history information: observational study and retrospective record analysis. J Clin Pharm Ther 2013;38:219-24.

461 Boockvar KS, Santos SL, Kushniruk A, Johnson C, Nebeker JR. Medication reconciliation: Barriers and facilitators from the perspectives of resident physicians and pharmacists. J Hosp Med 2011;6:329-37.

462 Rappaport DI, Collins B, Koster A, Mercado A, Greenspan J, Lawless S, Hossain J, Sharif I. Implementing medication reconciliation in outpatient pediatrics. Pediatrics 2011;128:e1600-e07.

463 Schnipper JL, Hamann C, Ndumele CD, Liang CL, Carty MG, Karson AS, Bhan I, Coley CM, Poon E, Turchin A, Labonville SA, Diedrichsen EK, Lipsitz S, Broverman CA, McCarthy P, Gandhi TK. Effect of an electronic medication reconciliation application and process redesign on potential adverse drug events: a cluster-randomized trial. Arch Intern Med 2009;169:771-80.

464 Schnipper JL, Liang CL, Hamann C, Karson AS, Palchuk MB, McCarthy PC, Sherlock M, Turchin A, Bates DW. Development of a tool within the electronic medical record to facilitate medication reconciliation after hospital discharge. J Am Med Inform Assn 2011;18:309-13.

465 Kramer J, Hopkins P, Rosendale J, Garrelts J, Hale L, Nester T, Cochran P, Eidem L, Haneke R. Implementation of an electronic system for medication reconciliation. Am J Health-Syst Ph 2007;64:404-22.

466 Agrawal A, Wu W. Reducing medication errors and improving systems reliability using an electronic medication reconciliation system. Jt Comm J Qual Patient Saf 2009;35:106-14.

467 Zoni AC, Durán García ME, Jiménez Muñoz AB, Salomón Pérez R, Martin P, Herranz Alonso A. The impact of medication reconciliation program at admission in an internal medicine department. Eur J Intern Med 2012;23:696-700.

363

468 Vawdrey DK, Chang N, Compton A, Tiase V, Hripcsak G. Impact of electronic medication reconciliation at hospital admission on clinician workflow. Proc AMIA Symp 2010;2010:822-6.

469 Wang T, Biederman S. Enhance the accuracy of medication histories for the elderly by using an electronic medication checklist. Perspect Health Inf Manag 2012;9:1-15.

470 Giménez Manzorro Á, Zoni AC, Rodríguez Rieiro C, Durán-García E, Trovato López AN, Pérez Sanz C, Bodas Gutiérrez P, Jiménez Muñoz AB. Developing a programme for medication reconciliation at the time of admission into hospital. Int J Clin Pharm 2011;33:603-09.

471 van der Kam WJ, Meyboom de Jong B, Tromp TF, Moorman PW, van der Lei J. Effects of electronic communication between the GP and the pharmacist. The quality of medication data on admission and after discharge. Fam Pract 2001;18:605-09.

472 Shawahna R, Rahman N-U, Ahmad M, Debray M, Yliperttula M, Declèves X. Electronic prescribing reduces prescribing error in public hospitals. J Clin Nurs 2011;20:3233-45.

473 Maslove D, Leiter R, Griesman J, Arnott C, Mourad O, Chow C, Bell C. Electronic versus dictated hospital discharge summaries: a randomized controlled trial. J Gen Intern Med 2009;24:995–1001.

474 Midlöv P, Bahrani L, Seyfali M, Höglund P, Rickhag E, Eriksson T. The effect of medication reconciliation in elderly patients at hospital discharge. Int J Clin Pharm 2012;34:113-19.

475 Huckvale C, Car J, Akiyama M, Jaafar S, Khoja T, Bin Khalid A, Sheikh A, Majeed A. Information technology for patient safety. Qual Saf Health Care 2010;19:i25-i33

476 Cortelyou-Ward K, Swain A, Yeung T. Mitigating error vulnerability at the transition of care through the use of health IT applications. J Med Syst 2012;36:3825-31.

477 Bassi J, Lau F, Bardal S. Use of information technology in medication reconciliation: A scoping review. Ann Pharmacother 2010;44:885-97.

478 Bates D, Gawande A. Improving safety with information technology. New Engl J Med 2003;348:2526-34.

479 Nemeth C, Nunnally M, O’Connor M, Cook R. Creating resilient IT: How the sign-out sheet shows clinicians make healthcare work. Proc AMIA Symp 2006: 584–88.

480 Barber N. Designing information technology to support prescribing decision making. Qual Saf Health Care 2004;13:450-54.

481 Poon EG, Blumenfeld B, Hamann C, Graydon-Baker E, Turchin A, Baker G, McCarthy P, Poikonen J, Schnipper JL, Hallisey RK, Smith S, McCormack C, Paterno M, Coley CM, Karson A, Chueh HC, Van Putten 364

C, Millar SG, Clapp M, Bhan I, Meyer GS, Gandhi TK, Broverman C. Design and implementation of an electronic application to support multidisciplinary medication reconciliation efforts at two academic medical centers. J Gen Intern Med 2006;21:192-92.

482 Samaranayake N, Cheung S, Chui W, Cheung B. Technology-related medication errors in a tertiary hospital: a 5-year analysis of reported medication incidents. Int J Med Inform 2012;81:828-33.

483 Zhu X, Gold S, Lai A, Hripcsak G, Cimino JJ. Using Timeline Displays to Improve Medication Reconciliation. International Conference on Ehealth, Telemedicine, and Social Medicine: Etelemed , Proceedings; 2009.

484 Hasan S, Duncan GT, Neill DB, Padman R. Towards a collaborative filtering approach to medication reconciliation. Proc AMIA Symp 2008:288- 92.

485 Hasan S, Duncan GT, Neill DB, Padman R. Automatic detection of omissions in medication lists. J Am Med Inform Assn 2011;18:449-58.

486 Agrawal A, Wu W, Khachewatsky I. Evaluation of an Electronic Medication Reconciliation System in Inpatient Setting in an Acute Care Hospital. Medinfo 2007: Proceedings of the 12th World Congress on Health (Medical) Informatics, Pts 1 and 2 2007;129:1027-31.

487 Schnipper JL, Hamann C, Ndumele C, Liang C, Carty M, Karson AS, Bhan I, Coley M, Poon EG, Turchin A, Labonville S, Diedrichsen E, Lipsitz S, Broverman C, McCarthy P, Gandhi T. The effects of a medication reconciliation application and process redesign on potential adverse drug events: Results of a controlled trial. J Gen Intern Med 2008;23:413-13.

488 Weeks DL, Corbett CF, Stream G. Beliefs of ambulatory care physicians about accuracy of patient medication records and technology-enhanced solutions to improve accuracy. J Healthc Qual 2010;32:12-21.

489 Campbell EM, Sittig DF, Ash JS, Guappone KP, Dykstra RH. Types of unintended consequences related to computerized provider order entry. J Am Med Inform Assn 2006;13:547-56.

490 Lesselroth B, Eisenhauer W, Adams S, Dorr DA, Randall C, Channon P, Adams K, Church V, Felder R, Douglas D. Simulation modeling of a check-in and medication reconciliation ambulatory clinic kiosk. J Healthc Eng 2011;2:197-221.

491 Greenhalgh T, Stramer K, Bratan T, Byrne E, Russell J, Hinder S, Potts H. The Devil's in the Details: Final report of the independent evaluation of the Summary Care Record and HealthSpace programmes. University College London, 2010. Accessed 20th August 2011. https://www.ucl.ac.uk/news/scriefullreport.pdf.

492 Traer E, Madhavan K. Electronic prescribing and transfer/discharge summaries: a survey of UK cancer networks. Pharm J 2010;284:453-54.

365

493 Callen JL, Alderton M, McIntosh J. Evaluation of electronic discharge summaries: A comparison of documentation in electronic and handwritten discharge summaries. Int J Med Inform 2008;77:613-20.

494 Greene RJ, Cavell GF, Jackson SHD. Interprofessional clinical education of medical and pharmacy students. Med Educ 1996;30:129-33.

495 Simoens S, Spinewine A, Foulon V, Paulus D. Review of the cost- effectiveness of interventions to improve seamless care focusing on medication. Int J Clin Pharm 2011;33:909-17.

496 Backes AC, Kuo GM. The association between functional health literacy and patient-reported recall of medications at outpatient pharmacies. Res Social Adm Pharm 2012;8:349-54.

497 Kuo GM, Steinbauer JR, Spann SJ. Conducting medication safety research projects in a primary care physician practice-based research network. Journal Of The American Pharmacists Association: Japha 2008;48:163-70.

498 Lindquist LA, Go L, Fleisher J, Jain N, Friesema E, Baker DW. Relationship of health literacy to intentional and unintentional non- adherence of hospital discharge medications. J Gen Intern Med 2012;27:173-78.

499 Persell SD, Osborn CY, Richard R, Skripkauskas S, Wolf MS. Limited health literacy is a barrier to medication reconciliation in ambulatory care. J Gen Intern Med 2007;22:1523-26.

500 Marvanova M, Roumie CL, Eden SK, Cawthon C, Schnipper JL, Kripalani S. Health literacy and medication understanding among hospitalized adults. J Hosp Med 2011;6:488-93.

501 Greenwald JL, Halasyamani L, Greene J, LaCivita C, Stucky E, Benjamin B, Reid W, Griffin FA, Vaida AJ, Williams MV. Making inpatient medication reconciliation patient centered, clinically relevant and implementable: a consensus statement on key principles and necessary first steps. Journal Of Hospital Medicine: An Official Publication Of The Society Of Hospital Medicine 2010;5:477-85.

502 Schwappach D. Engaging patients as vigilant partners in safety: a systematic review. Med Care Res Rev. 2010 67:119-48.

503 Schwappach D, Frank O, Buschmann U, Babst R. Effects of an educational patient safety campaign on patients' safety behaviours and adverse events. J Eval Clin Pract 2013;19:285-91.

504 van Sluisveld N, Zegers M, Natsch S, Wollersheim H. Medication reconciliation at hospital admission and discharge: insufficient knowledge, unclear task reallocation and lack of collaboration as major barriers to medication safety. BMC Health Services Research 2012;12:170.

366

505 Davis C. Emergency department visits: we are not prepared. Am J Emerg Med 2012;30:1364-70.

506 Knight DA, Thompson D, Mathie E, Dickinson A. 'Seamless care? Just a list would have helped!' Older people and their carer's experiences of support with medication on discharge home from hospital. Health Expect 2013;16:277-91.

507 Glintborg B, Hillestrom PR, Olsen LH, Dalhoff KP. Are patients reliable when self-reporting medication use? Validation of structured drug interviews and home visits by drug analysis and prescription data in acutely hospitalized patients. J Clin Pharmacol 2007;47:1440-49.

508 Sjahid SI, van der Linden PD, Stricker BHC. Agreement between the pharmacy medication history and patient interview for cardiovascular drugs: the Rotterdam elderly study. Brit J Clin Pharmaco 1998;45:591-95.

509 Yasein NA, Barghouti FF, Irshaid YM, Suleiman AA. Discrepancies between elderly patient's self-reported and prescribed medications: a social investigation. Scand J Caring Sci 2013;27:131-8.

510 Rennke S, Nguyen O, Shoeb M, Magan Y, Wachter R. Hospital-initiated transitional care interventions as a patient safety strategy. Ann Intern Med 2013;158:433-40.

511 Bikowski RM, Ripsin CM, Lorraine VL. Physician-patient congruence regarding medication regimens. J Am Geriatr Soc 2001;49:1353-57.

512 Stephens M, Fox B, Kukulka G, Bellamy J. Medication, allergy, and adverse drug event discrepancies in ambulatory care. Fam Med 2008;40:107-10.

513 Lapane K, Dube C, Schneider K, Quilliam B. Misperceptions of patients vs providers regarding medication-related communication issues. Am J Manag Care 2007;13:613-18.

514 Dooley MJ, Tan E. Patient-completed medication histories versus those obtained by a pharmacist in a pre-admission clinic. J Pharm Pract Res 2008;38:216-18.

515 Al Mahdy H, Seymour D. How much can elderly patients tell us about their medications? Postgrad Med J 1990;66:116-21.

516 Riley-Lawless K. Family-identified barriers to medication reconciliation. J Spec Pediatr Nurs 2009;14:94-101.

517 Witting MD, Hayes BD, Schenkel SM, Drucker CB, DeWane MP, Lantry JH, Vashi SV. Emergency department medication history taking: current inefficiency and potential for a self-administered form. J Emerg Med 2013;45:105-09.

518 Montpetit LM, Roy MT. Evaluation of a patient-completed versus health professional-conducted medication history. Drug Intel Clin Phar 1988;22:964-69. 367

519 Mageean RJ. Study of 'discharge communications' from hospital. Brit Med J 1986;293:1283-4.

520 Foust J, Bixby M, Naylor M, Schwartz J. Post-hospital discharge medication discrepancies among elders with heart failure. Gerontologist 2004;44:169-69.

521 Moczygemba L, Gatewood S, Kennedy A, Osborn R, Alexander A, Matzke G, Goode J-V. Medication reconciliation campaign in a clinic for homeless patients. Am J Health-Syst Ph 2012;69:558.

522 Cumbler E, Wald H, Kutner J. Lack of patient knowledge regarding hospital medications. J Hosp Med 2010;5:83-86.

523 Ellitt GR, Engblom E, Aslani P, Westerlund T, Chen TF. Drug related problems after discharge from an Australian teaching hospital. Pharm World Sci 2010;32:622-30.

524 Ostini R, Kairuz T. Investigating the association between health literacy and non-adherance. Int J Clin Pharm 2014;36:36-44.

525 Hain DJ, Tappen R, Diaz S, Ouslander JG. Cognitive impairment and medication self-management errors in older adults discharged home from a community hospital. Home Healthc Nurse 2012;30:246-54.

526 Cain CH, Neuwirth E, Bellows J, Zuber C, Green J. Patient experiences of transitioning from hospital to home: An ethnographic quality improvement project. J Hosp Med 2012;7:382-87.

527 Waterworth S, Luker K. Reluctant collaborators: do patients want to be involved in decisions concerning care? J Adv Nurs 1990;15:971-76.

528 Borgsteede SD, Karapinar-Çarkit F, Hoffmann E, Zoer J, Pm. Information needs about medication according to patients discharged from a general hospital. Patient Educ Couns 2011;83:22-28.

529 Porter SC, Kohane IS, Goldmann DA. Parents as partners in obtaining the medication history. J Am Med Inform Assn 2005;12:299-305.

530 Kimland E, Bottiger Y, Lindemalm S. Recent drug history in children visiting a pediatric emergency room and documentation in medical records. Eur J Clin Pharmacol 2011;67:1085-89.

531 Daji S, Urban R. Why don't patients bring their own drugs into hospital? Int J Pharm Pract 2012;20:63.

532 Mansur N, Weiss A, Hoffman A, Gruenewald T, Beloosesky Y. Continuity and adherence to long-term drug treatment by geriatric patients after hospital discharge: a prospective cohort study. Drugs Aging. 2008;25:861- 70.

533 Stange D, Kriston L, von Wolff A, Baehr M, Dartsch DC. Medication complexity, prescription behaviour and patient adherence at the interface

368

between ambulatory and stationary medical care. Eur J Clin Pharmacol 2013;69:573-80.

534 Tsilimingras D, Bates DW. Addressing postdischarge adverse events: a neglected area. Jt Comm J Qual Patient Saf 2008;34:85-97.

535 Tsilimingras D, Brummel-Smith K, Brooks RG. Postdischarge adverse events in the elderly. J Patient Saf 2009;5:201-04.

536 Olson DP, Windish DM. Communication discrepancies between physicians and hospitalized patients. Arch Intern Med 2010;170:1302-07.

537 MacDonald ET, MacDonald JB, Phoenix M. Improving drug compliance after hospital discharge. Brit Med J 1977;2:618-21.

538 American Society of Health-System Pharmacists. Executive summary of the continuity of care in medication use summit. Am J Health-Syst Ph 2008;65:e3-9.

539 Webb J, Feinglass J, Makoul G, Wilkes CL, Dunham DP, Baker DW, Wolf MS. Can electronic health records help improve patients' understanding of medications? Am J Manag Care 2010;16:919-22.

540 Schnipper JL, Gandhi TK, Wald JS, Grant RW, Poon EG, Volk LA, Businger A, Williams DH, Siteman E, Buckel L, Middleton B. Effects of an online personal health record on medication accuracy and safety: a cluster-randomized trial. J Am Med Inform Assn 2012;19:728-34.

541 Fricton JR, Davies D. Personal Health Records to Improve Health Information Exchange and Patient Safety Technology and Medication Safety). In: Henriksen K, Battles J, Keyes M, editors. Advances in Patient Safety: New Directions and Alternative Approaches (Vol. 4: Technology and Medication Safety). Rockville: Agency for Healthcare Research and Quality (US), 2008.

542 Creswell J. Research Design. Qualitative, Quantitative and Mixed Method Approaches. 3rd ed. USA: Sage Publications, 2009.

543 Crotty M. The Foundations of Social Research. London Sage, 1998.

544 Gray D. Doing Research in the Real World. 2nd ed. London: Sage, 2009.

545 Barbour RS. The Case for Combining Qualitative and Quantitative Approaches in Health Services Research. Journal of Health Services Research & Policy 1999;4:39-43.

546 Burgess R. Field Research: A Sourcebook and Field Manual. London: George Allen and Unwin, 1982.

547 Creswell J, Plano Clark V, Guttmann M, Hanson W. Advances in Mixed Method Design. In: Tashakkorie A, Teddlie C, editors. Handbook of Mixed Methods in the Social and Behavioural Sciences. Thousand Oaks , CA: Sage 2003.

369

548 Newman I, Benz R. Qualitative-Quantitative Research Methodology: Exploring the Interactive Continuum. Carbondale, Illinois: Southern Illinois: University Press, 1998.

549 Tashakkori A, Teddlie C. Handbook of Mixed Methods in Social and Behavioral Research. . Thousand Oaks, CA: Sage, 2003.

550 Brannen J. Mixing methods: The entry of qualitative and quantitative approaches into the research process. Int J Soc Res Methodol 2005;8:173-84.

551 Johnson R, Onwuegbuzie A, Turner L. Toward a definition of mixed methods research. . Journal of Mixed Methods Research 2007;1:112-33.

552 Green J, Thorogood. Qualitative Methods for Health Research. London: Sage Publications, 2004.

553 Bryman A. Integrating quantitative and qualitative research: how is it done? Qualitative Research 2006;6:97-113.

554 Campbell D, Fiske D. Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin 1959;56:81-105.

555 Denzin N. The Research Act: A Theoretical Introduction to Sociological Methods. New York: McGraw-Hill, 1978.

556 Sieber S. The integration of fieldwork and survey methods. Am J Sociol 1973;78:1335-39.

557 Jick T. Mixing Qualitative and Quantitative Methods: Triangulation in Action. In: Mannen JV, editor. Qualitative methodology Beverly Hills, CA: Sage, 1983.

558 Greene J, Caracelli V, Graham W. Toward a conceptual framework for mixed-method evaluation designs. Educ Eval Policy An 1989;11:255-74.

559 Murphy E, Dingwall R, Greatbatch D, Parker S, Watson P. Qualitative research methods in health technology assessment: a review of the literature. Health Technology Assessment 1998;2:iii-ix.

560 Miles M, Huberman A. Qualitative Data Analysis : An Expanded Sourcebook 2nd ed. London: Sage, 1994.

561 Gantley M, Harding G, Kumar S, Tissier J. An Introduction to Qualitative Methods for Health Professionals. In: Carter Y, Shaw S, Thomas C, editors. Master Classes in Primary Care Research Series. London: Royal College of General Practitioners, 1999.

562 Thorne S. Data analysis in qualitative research. Evidence Based Nursing 2000;3:68-70.

563 Bryman A. The debate about quantitative and qualitative research. In: Bryman A, Burgess R, editors. Qualitative Research. London: Sage, 1999:35-69. 370

564 Barbour R. The role of qualitative research in broadening the 'evidence base' for clinical practice. J Eval Clin Pract 2000;6:155-63.

565 Hammersley M, Atkinson P. Ethnography: Principles in Practice 3rd ed. London: Taylor & Francis, 2007.

566 Mason J. Qualitative Researching. 2nd ed. London: Sage, 2002.

567 Atkinson P, Hammersley M. Ethnography and Participant Observation. In: Denzin N, Lincoln Y, editors. Strategies of Qualitative Enquiry. 2nd ed. Thousand Oaks, California: Sage, 2004.

568 Bryman A. Quantity and Quality in Social Research. London: Unwin Hyman, 1988.

569 Trochim W. Research Methods. Knowledge Base. Deduction and Induction. 2006. Accessed 21st January 2014. http://www.socialresearchmethods.net/kb/dedind.php.

570 Jorgensen D. Participant Observation : A Methodology for Human Studies USA: Sage, 1989.

571 Hall T. At home with the young homeless. Int J Soc Res Methodol 2000;3:121-33.

572 Gillham B. Observation Techniques: Structured to Unstructured London: Bloomsbury Academic, 2008.

573 Emerson RM, Fretz RI, Shaw LL. Participant observation and fieldnotes. In: Atkinson P, Coffey A, Delamont S, Lofland J, Lofland L, editors. Handbook of Ethnography. London: Sage, 2001:352-68.

574 Emerson R, Fretz R, Shaw L. Writing Ethnographic Fieldnotes. US: University of Chicago Press, 1995.

575 Bailey C. A Guide to Field Research. Thousand Oaks, Ca: Pine Forge Press, 1996.

576 Kumar R. Research Methodology : A Step by Step Guide for Beginners. London: Sage, 1999.

577 Maxwell J. Understanding and validity in qualitative research. Harvard Educational Review 1992;62:279-301.

578 Smith H. Strategies of Social Research US: Harcourt School, 1991.

579 Lincoln Y, Guba E. Naturalistic inquiry. Beverly Hills, CA: Sage, 1985.

580 Johnson P, Harris D. Qualitative and quantitative issues in research design. In: Partington D, editor. Essential Skills for Management Research. London: Sage, 2002.

371

581 Campbell D, Stanley J. Experimental and quasi-experimental designs for research on teaching. In: Gage N, editor. Handbook of Research on Teaching. Chicago, IL: Rand McNally, 1963:171–246.

582 Strong P. The Ceremonial Order of the Clinic: Parents, Doctors and Medical Bureaucracies. London: Routledge & Kegan Paul, 1979.

583 Kompier MA. The "Hawthorne effect" is a myth, but what keeps the story going?

584 Rasmussen K. Halo Effect. In: Salkind N, Rasmussen K, editors. Encyclopedia of Educational Psychology. Thousand Oaks, CA: Sage Publications, Inc., 2008.

585 Glaser B, Strauss A. The Discovery of Grounded Theory: Strategies for Qualitative Research. . Chicago: Aldine, 1967.

586 Denscombe M. The Good Research Guide: for Small-scale Social Research Projects London: Open University Press, 2003.

587 Weber M. Politics as a Vocation. In: Gerth H, Wright Mills C, editors. Essays in Sociology. Oxford: University Press, 1946.

588 Warren C. Gender Issues in Field Research. London: Sage, 1988.

589 Barbour R. Checklists for improving rigour in qualitative research: a case of the tail wagging the dog. Brit Med J 2001;322:1115-7.

590 Curran-Everett D. Explorations in statistics: standard deviations and standard errors. Adv Physiol Educ 2008 32:203-8.

591 Barbour R. The case for combining qualitative and quantitative approaches in health services research. J Health Serv Res Policy 1999; 4:39-43

592 Patton M. Qualitative Research and Evaluation Methods. 3rd ed. Newbury Park, CA: Sage, 2002.

593 Elliott Ra, Woodward MC. Medication-related problems in patients referred to aged care and memory clinics at a tertiary care hospital. Australas J Ageing 2011;30:124-9.

594 St. Peter W. Improving medication safety in chronic kidney disease patients on dialysis through medication reconciliation. Adv Chronic Kidney D 2010;17:413-19.

595 Broadhead R, Rist R. Gatekeepers and the social control of social research. Social Problems 1976;23:325-36.

596 Berg B. Qualitative Research Methods for the Social Sciences. 6th ed. Needham Heights, MA: Allyn and Bacon, 2006.

372

597 Morse J. “Emerging from the data”: the cognitive processes of analysis in qualitative inquiry. In: Morse J, editor. Critical Issues in Qualitative Research Methods. Thousand Oaks, CA: Sage, 1994:23–43.

598 Smith J, Osborn M. Interpretative phenomenological analysis. In: Smith J, editor. Qualitative Psychology. London: Sage, 2003.

599 Dey I. Qualitative Data Analysis. London: Routledge Taylor Francis, 1993.

600 Braun V, Clarke V. Using thematic analysis in psychology. Qualitative Research in Psychology, 2006.

601 Silverman D. Interpreting Qualitative data: Methods for Analysing Talk, Text and Interaction 2nd Edition ed. London: Sage, 2005.

602 Strauss A. Qualitative Analysis for Social Scientists. New York: Cambridge University Press, 1987.

603 Tam VC, Knowles SR, Cornish PL, Fine N, Marchesano R, Etchells EE. Frequency, type and clinical importance of medication history errors at admission to hospital: a systematic review. Can Med Assoc J 2005;173:510-15.

604 Gleason KM, McDaniel MR, Feinglass J, Baker DW, Lindquist L, Liss D, Noskin GA. Results of the Medications at Transitions and Clinical Handoffs (MATCH) study: an analysis of medication reconciliation errors and risk factors at hospital admission. J Gen Intern Med 2010;25:441-47.

605 Kuo CN, Lin YM, Wu MT, Kuo LN, Lee LW, Chen HY. Pharmacist-directed reconciliation for reducing medication discrepancies: A pilot study in a nursing home setting in Taiwan. J Food Drug Anal 2013;21:160-64.

606 General Pharmaceutical Council (GPhC). Standards of conduct, ethics and performance. 2012. Accessed 23rd March 2014. http://www.pharmacyregulation.org/sites/default/files/Standards%20of%20 conduct%20ethics%20and%20performance%20July%202012.pdf.

607 NHS Information Centre. Hospital Episode Statistics: Accident and emergency - Patient journey: further analysis of the published 2007-08 A&E HES data (experimental statistics). The Health and Social Care Information Centre, 2009. Accessed September 2011 2011.

608 Lawton R, McEachan RRC, Giles SJ, Sirriyeh R, Watt IS, Wright J. Development of an evidence-based framework of factors contributing to patient safety incidents in hospital settings: a systematic review. 2012. March 15, 2012. http://qualitysafety.bmj.com/content/early/2012/03/14/bmjqs-2011- 000443.abstract.

609 Lingard L, Espin S, Whyte S, Regehr G, Baker GR, Reznick R, Bohnen J, Orser B, Doran D, Grober E. Communication failures in the operating room: an observational classification of recurrent types and effects. Qual Saf Health Care 2004;13:330-34.

373

610 Simon C. The consultation. InnovAiT 2009;2:113-21.

611 Ellington AM, Barnett CW, Johnson DR, Nykamp D. Current methods used to teach the medication history interview to Doctor of Pharmacy students. Am J Pharm Educ 2002;66:103-07.

612 Thistlethwaite J. Communication theory models—Outside the face-to-face consultation. InnovAiT 2008;1:488-93.

613 Shannon CE, Weaver W. The Mathematical Theory of Communication. Urbana, Illinois: University of Illinois Press, 1949.

614 Schramm W. How communication works. In: Schramm W, editor. The process and effects of communication. Urbana, Illinois: University of Illinois Press, 1954:3-26.

615 Management Study Guide. Berlo’s Model of Communication. 2013. Accessed 4th February 2014. http://www.managementstudyguide.com/berlo-model-of- communication.htm.

616 Communcation Theory.org. Berlo’s SMCR Model of Communication. 2014. Accessed 14th February 2014 http://communicationtheory.org/berlos- smcr-model-of-communication/.

617 Tate P. Ideas, concerns and expectations. Medicine 2005;33:26-27.

618 National Audit Office (NAO). Improving emergency care in England. 2004. London.

619 National Patient Safety Agency (NPSA). Rapid Response Report. Reducing harm from omitted and delayed medicines in hospital. Gateway ref: 13614. 2010.

620 Zhang Y, Peck K, Spalding M, Jones BG, Cook RL. Discrepancy between patients' use of and health providers' familiarity with CAM. Patient Educ Couns 2012;89:399-404.

621 BusinessBalls. Conscious Competence Learning Model. 2013. Accessed 22nd January 2014. http://www.businessballs.com/consciouscompetencelearningmodel.htm#c onscious-competence-theory-origins.

622 Abadel FT, Hattab AS. How does the medical graduates' self-assessment of their clinical competency differ from experts' assessment? BMC Medical Education 2013;13.

623 Dubin R. Human Relations in Administration. 4th ed ed. Englewood Cliffs, NJ: Prentice-Hall, 1974.

624 Barrick MR, Mount MK, Judge TA. Personality and Performance at the Beginning of the New Millennium: What Do We Know and Where Do We Go Next? International Journal of Selection and Assessment 2001;9:9-30.

374

625 Tett RP, Jackson DN, Rothstein M. Personality measures as predictors of job performance: A meta-analytic review. Personnel Psychology 1991;44:703-42.

626 Gettman DA, Ranelli PL, Ried LD. Influence of gender on outcomes of medication-history interviewing. Patient Education And Counseling 1996;27:147-60.

627 Lively BT. Evaluating risk taking propensity as a predictor of the outcome dimensions of medication history taking. American Journal of Pharmaceutical Education 1983;47:129-35.

628 Bacon F. The New Organon or True Directions Concerning the Interpretation of Nature. 1620. Accessed 23rd May 2014. http://www.constitution.org/bacon/nov_org.htm.

629 Health Foundation. The measurement and monitoring of safety. Health Foundation, 2013. Accessed 24th April 2014. http://patientsafety.health.org.uk/sites/default/files/resources/the_measure ment_and_monitoring_of_safety.pdf.

630 Hall P. Interprofessional teamwork: professional cultures as barriers. J Interprof Care 2005;Suppl 1:188-96.

631 Axelsson SB, Axelsson R. From territoriality to altruism in interprofessional collaboration and leadership. J Interprof Care 2009;23:320-30.

632 Weaver SJ, Lubomksi LH, Wilson RF, Pfoh ER, Martinez KA, Dy SM. Promoting a culture of safety as a patient safety strategy: A systematic review. Ann Intern Med 2013;158:369-74.

633 Huckvale C, Car J, Akiyama M, Jaafar S, Khoja T, Bin Khalid A, Sheikh A, Majeed A. Information technology for patient safety. Quality and Safety in Health Care 2010;19:i25-i33.

634 Confino-Cohen R, Leader A, Klein N, Pereg D, Khoury S, Perl L, Goldberg A. Drug allergy in hospitalized patients: the contribution of allergy consultation and a structured questionnaire. Int Arch Allergy Imm 2012;158:307-12.

635 Porter S, Manzi S, Volpe D, Stack A. Getting the data right: information accuracy in pediatric emergency medicine. Qual Saf Health Care 2006;15:296-301.

636 Goulding L. The quality and safety of healthcare provided to hospital inpatients who are placed on clinically inappropriate wards. University of York, 2011.

637 Duggan C, Feldman R, Hough J, Bates I. Reducing adverse prescribing discrepancies following hospital discharge. Int J Pharm Pract 1998;6:77- 82.

638 National Audit Office (NAO). Emergency admissions to hospital: managing the demand. 2013. London. 375

639 Doran T, Kontopantelis E, Valderas JM, Campbell S, Roland M, Salisbury C, Reeves D. Effect of financial incentives on incentivised and non- incentivised clinical activities: longitudinal analysis of data from the UK Quality and Outcomes Framework. Brit Med J 2011;342.

640 Burns EM, Rigby E, Mamidanna R, Bottle A, Aylin P, Ziprin P, Faiz OD. Systematic review of discharge coding accuracy. J Public Health 2012;34:138-48.

641 Nouraei SA, O'Hanlon S, Butler C, Hadovsky A, Donald E, Benjamin E, Sandhu GS. A multidisciplinary audit of clinical coding accuracy in otolaryngology: financial, managerial and clinical governance considerations under payment-by-results. Clin Otolaryngol 2009;34:43-51.

642 Maso I. Necessary subjectivity: exploring researchers' motives, passions and prejudices in pursuit of answering 'true' questions. In: Finlay L, Gough B, editors. Reflexivity: A Practical Guide for Researchers in Health and Social Sciences. Oxford: Blackwell Publishing, 2003.

643 Miles A, Bentley P, Polychronis A, Grey J, Price N. Recent progress in health services research: on the need for evidence-based debate. Journal of Evaluation in Clinical Practice 1998;4:257-65.

644 Zadik Y, Abu-Tair J, Yarom N, Zaharia B, Elad S. The importance of a thorough medical and pharmacological history before dental implant placement. Aust Dent J 2012;57:388-92.

645 Shojania KG. Conventional evaluations of improvement interventions: more trials or just more tribulations? BMJ Qual Saf 2013;22:881-84.

646 (NPSA) NPSA. Never Events Framework. Update for 2010/2011. Policies and action for Primary Care Trusts. 2010. NPSA. 1st June 2014. http://www.nrls.npsa.nhs.uk/resources/collections/never- events/?entryid45=68518.

647 Kitson N, Price M, Lau F, Showler G. Developing a medication communication framework across continuums of care using the Circle of Care Modeling approach. BMC Health Services Research 2013;13:418.

648 Royal Pharmaceutical Society (RPS). Keeping patients safe when they transfer between care providers – getting the medicines right. Prescribing and Primary Care Services, November 2011. Picton C & Wright H 2012.

649 Care Quality Commission (CQC). Managing patients' medicines after discharge from hospital. Care Quality Commission 2009.

650 Healthcare Improvement Scotland. SIGN 128. The SIGN discharge document. 2012. Edinburgh.

651 Carayon P, Wetterneck TB, Rivera-Rodriguez AJ, Hundt AS, Hoonakker P, Holden R, Gurses AP. Human factors systems approach to healthcare quality and patient safety. Applied Ergonomics 2014;45:14-25.

376

652 Carayon P, Schoofs Hundt A, Karsh B-T, Gurses AP, Alvarado CJ, Smith M, Flatley Brennan P. Work system design for patient safety: the SEIPS model. Qual Saf Health Care 2006;15:i50-i58.

653 Donabedian A. The quality of care: How can it be assessed? JAMA 1988;260:1743-48.

654 Holden RJ, Carayon P, Gurses AP, Hoonakker P, Hundt A, Ozok A, Rivera-Rodriguez AJ. SEIPS 2.0: a human factors framework for studying and improving the work of healthcare professionals and patients. Ann Intern Med 2012;157:417-28.

655 Cohen M. Medication Errors. Washington, US: American Pharmacist Association (APhA), 2007.

656 Van Cott H. Human Errors: Their Causes and Reduction. In: Bogner M, editor. Human Error in Medicine. Hillsdale, New Jersey: Lawrence Erbaum Associates, 1994.

657 Business Dictionary. Process. 2014. Accessed 25th February 2014. http://www.businessdictionary.com/definition/process.html#ixzz2a5VTivUD .

658 Nancarrow SA, Roots A, Grace S, Moran A, Vanniekerk-Lyons K. Implementing large-scale workforce change: learning from 55 pilot sites of allied health workforce redesign in Queensland, Australia. Hum Resour Health 2013;11:66.

659 Smith R, Duffy J. Developing a competent and flexible workforce using the Calderdale Framework Int J Ther Rehabil 2010;17:254 - 62.

660 Shekelle PG, Pronovost PJ, Wachter RM, McDonald KM, Schoelles K, Dy SM, Shojania K, Reston JT, Adams AS, Angood PB, Bates DW, Bickman L, Carayon P, Donaldson L, Duan N, Farley DO, Greenhalgh T, Haughom JL, Lake E, Lilford R, Lohr KN, Meyer GS, Miller MR, Neuhauser DV, Ryan G, Saint S, Shortell SM, Stevens DP, Walshe K. The top patient safety strategies that can be encouraged for adoption now. Ann Intern Med 2013;158:365-68.

661 Walton M, Elliott S. Improving safety and quality: how can education help? Med J Aust 2006;184:S60-64.

662 Ghazanfar M, Honore P, Nielsen T, Andersen S, Rasmussen M. Hospital admission interviews are time-consuming with several interruptions. Dan Med J 2012;59:A4534.

663 Vashitz G, Nunnally ME, Parmet Y, Bitan Y, O'Connor MF, Cook RI. How do clinicians reconcile conditions and medications? The cognitive context of medication reconciliation. Cognition Technology & Work 2013;15:109- 16.

377

664 Vashitz G, Nunnally M, Bitan Y, Parmet Y, O’Connor M, Cook R. Making sense of diseases in medication reconciliation. Cogn Tech Work 2010; 13:151–58.

665 Kwan J, Lo L, Sampson M, Shojania K. Medication reconciliation during transition of care as a patient safety strategy. Ann Intern Med 2013;158:397-403.

666 Tarn DM, Heritage J, Paterniti DA, Hays RD, Kravitz RL, Wenger NS. Physician communication when prescribing new medications. Arch Intern Med 2006;166:1855-62.

667 Lundgrén-Laine H, Salanterä S. Think-Aloud Technique and Protocol Analysis in Clinical Decision-Making Research. Qualitative Health Research 2010;20:565-75.

668 Davis RE, Sevdalis N, Vincent CA. Patient involvement in patient safety: How willing are patients to participate? Qual Saf Health Care 2011;20:108- 14.

669 Ward J, Giles S, Hrisos S, Rhodes P, Din I, Walsh P. Involving patients in their own safety. In: Lawton R, G A, editors. Innovating for Patient Safety in Medicine. London: Sage, 2012.

670 Entwistle V, Mello M, Brennan T. Advising patients about patient safety:currentinitiatives risk shifting responsibility. Jt Comm J Qual Patient Saf 2005;31:483-94.

671 Department of Health. The NHS Constitution for England. 2010. London.

672 Ward JK, Armitage G. Can patients report patient safety incidents in a hospital setting? A systematic review. BMJ Qual Saf 2012;21:685-99.

673 National Patient Safety Agency (NPSA). NHS-NPSA. 2012. Accessed 18th May 2014. http://www.npsa.nhs.uk/.

674 Johnstone M, Kanistaki O. Engaging patients as safety partners: some considerations for ensuring a culturally and linguistically appropriate approach. Health Policy 2009;90:1-7.

675 Entwistle V, McCaughan D, Watt I, Birks Y, Hall J, Peat M, Williams B. Speaking up about safety concerns: multi-setting qualitative study of patients' views and experiences. Qual Saf Health Care 2010;19:e33.

676 Davis R, Koutantji M, Vincent C. How willing are patients to question healthcare staff on issues related to the quality and safety of their healthcare? An exploratory study. Qual Saf Health Care 2008;17:90-6.

677 Henderson S. Power imbalance between nurses and patients: a potential inhibitor of partnership in care. J Clin Nurs 2003;12:501-8.

678 Bate P, Robert G. Experience-based design: from redesigning the system around the patient to co-designing services with the patient. Qual Saf Health Care 2006;15:307-10. 378

679 Royal College of Physicians & Health and Social Care Information Centre. Standards for the clinical structure and content of patient records. 2013. 4th March 2014. http://www.rcplondon.ac.uk/sites/default/files/standards- for-the-clinical-structure-and-content-of-patient-records.pdf.

680 Burgin A, O’Rourke R, Tully M. Learning to work with computerisation of medical notes and prescription charts. Int J Pharm Pract 2011;20:14.

681 Samaranayake N, Cheung S, Chui W, Cheung B. Technology-related medication errors in a tertiary hospital: a 5-year analysis of reported medication incidents. 2012.

682 Bates DW. Using information technology to reduce rates of medication errors in hospitals. Brit Med J 2000;320:788-91.

683 NHS Institute for Innovation and Improvement. SBAR - Suituation, Background, Assessment, Recommendation. 2008. Accessed 2nd June 2014. http://www.institute.nhs.uk/quality_and_service_improvement_tools/quality _and_service_improvement_tools/sbar_-_situation_-_background_- _assessment_-_recommendation.html.

684 Manias T, Tomlinson J. Implementation and evaluation of the SBAR tool in the communication between medical staff in obstetrics. Archives of Disease in Childhood - Fetal and Neonatal Edition 2011;96:131.

685 Rozich JD, Howard RJ, Justeson JM, Macken PD, Lindsay ME, Resar RK. Standardization as a mechanism to improve safety in health care. Jt Comm J Qual Saf 2004;30:5-14.

686 West E. Sociological contributions to patient safety. In: Walshe K, Boaden R, editors. Patient Safety Research into Practice. New York, USA: Open University Press, 2006:19-30

687 NHS Institute for Innovation and Improvement. Protocol Based Care. 2008. Accessed 2nd June 2014. http://www.institute.nhs.uk/quality_and_service_improvement_tools/quality _and_service_improvement_tools/protocol_based_care.html.

688 Lingard L, Regehr G, Orser B, et al. Evaluation of a preoperative checklist and team briefing among surgeons, nurses, and anesthesiologists to reduce failures in communication. Archives of Surgery 2008;143:12-17.

689 Leonard M, Graham S, Bonacum D. The human factor: the critical importance of effective teamwork and communication in providing safe care. Qual Saf Health Care 2004;13:i85-i90.

690 National Patient Safety Agency (NPSA). The adult patient’s passport to safer use of insulin. NHS, 2011. Accessed 23rd May 2014. http://www.nrls.npsa.nhs.uk/alerts/?entryid45=130397.

691 National Patient Safety Agency (NPSA). Rapid Response Report NPSA/2010/RRR009: Reducing harm from omitted and delayed medicines

379

in hospital. Supporting Information. NHS, 2010. Accessed 3rd April 2014. http://www.nrls.npsa.nhs.uk/alerts/?entryid45=66720.

692 National Patient Safety Agency (NPSA). Patient Safety Alert NPSA 2009/PSA005: Safer lithium therapy. Supporting Information. NHS, 2009. Accessed 2nd June 2014. http://www.nrls.npsa.nhs.uk/alerts/?entryid45=65426&p=2.

693 National Patient Safety Agency (NPSA). Improving compliance with oral methotrexate guidelines. NHS, 2006. Accessed 25th May 2014. http://www.nrls.npsa.nhs.uk/alerts/?entryid45=59800&p=4.

694 Härgestam M, Lindkvist M, Brulin C, Jacobsson M, Hultin M. Communication in interdisciplinary teams: exploring closed-loop communication during in situ trauma team training. BMJ Open 2013;3.

695 Lo L. Teamwork and communication in healthcare. Canadian Patient Safety Institute, 2011. Accessed 25th April 2014. http://www.patientsafetyinstitute.ca/English/toolsResources/teamworkCom munication/Documents/Canadian%20Framework%20for%20Teamwork%2 0and%20Communications%20Lit%20Review.pdf.

696 Maund T. Closing the communication loop: Using readback/hearback to support patient safety. Jt Comm J Qual Saf 2004;30:460-64.

697 Patidar A. Communication and Nursing Education. India: Pearson Education, 2012.

698 Leonard M, Graham S, Bonacum D. The human factor: the critical importance of effective teamwork and communication in providing safe care. Qual Saf Health Care 2004;13:i85-90.

699 Gerbner G. Toward a general model of communication. Audiovisual communication review 1956;4:171-99.

700 Timmons S, Baxendale B, Buttery A, Miles G, Roe B, Browes S. Implementing human factors in clinical practice. Emergency Medicine Journal 2014.

701 Gallwey T. The Inner Game of Tennis. New York: Random House, 1974.

702 Stevick E. Working With Teaching Methods. Canada: Thomson Learning, 1998.

703 Evans AS, Lazar EJ, Tiase VL, Fleischut P, Bostwick SB, Hripcsak G, Liebowitz R, Forese LL, Kerr G. The role of housestaff in implementing medication reconciliation on admission at an academic medical center. Am J Med Qual 2011;26:39-42.

704 Wass V, Van der Vleuten C Fau - Shatzer J, Shatzer J Fau - Jones R, Jones R. Assessment of clinical competence. Lancet 2001;24:945-9.

380

705 Carraccio C, Wolfsthal Sd Fau - Englander R, Englander R Fau - Ferentz K, Ferentz K Fau - Martin C, Martin C. Shifting paradigms: from Flexner to competencies. Acad Med 2002;77:361-7.

706 Walton MM, Shaw T, Barnet S, Ross J. Developing a national patient safety education framework for Australia. Qual Saf Health Care 2006;15:437-42.

707 Health and Safety Commission (HSC). ACSNI Study Group on Human Factors. 3rd Report: Organising for Safety. London: HMSO 1993.

708 Westrum R. A typology of organisational cultures. Qual Saf Health Care 2004;13:ii22-ii27.

709 University M. Manchester Patient Safety Framework. (MaPSaF). 2006. Accessed 24th April 2014. http://www.nrls.npsa.nhs.uk/resources/patientsafety-topics/human-factors-patient-safety-culture/?entryid45=59796&P=1.

710 Nieva VF, Sorra J. Safety culture assessment: a tool for improving patient safety in healthcare organizations. Quality and Safety in Health Care 2003;12:ii17-ii23.

711 Health Foundation. High Reliability Organisations. 2011. Accessed 25th January 2014. http://www.health.org.uk/publications/high-reliability- organisations/.

712 Quality AfHRa. Becoming a High Reliability Organization: Operational Advice for Hospital Leaders 2008. Accessed 26th April 2013. http://www.ahrq.gov/professionals/quality-patient-safety/quality- resources/tools/hroadvice/hroadvice.pdf.

713 Royal Pharmaceutical Society (RPS). Keeping patients safe when they transfer between care providers – getting the medicines right. Good practice guidance for healthcare professions. 2011. Accessed 17th October 2011. http://www.rpharms.com/current-campaigns-pdfs/1303--- rps---transfer-of-care-10pp-professional-guidance---final-final.pdf.

714 National Health Service (NHS). Safety thermometer. 2014. Accessed 29th April 2014. http://www.safetythermometer.nhs.uk/index.php?option=com_dashboards &view=medication&Itemid=334.

381