BMJ Open BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from
Pharmacy provided medicines reconciliation within 24 hours of admission: A randomised controlled pilot study
ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2016-013647
Article Type: Research
Date Submitted by the Author: 01-Aug-2016
Complete List of Authors: Cadman, Brit; Cambridge University Hospitals NHS Foundation Trust, Pharmacy Wright, David; University of East Anglia, School of Pharmacy Bale, Amanda; Cambridge University Hospitals NHS Foundation Trust, Pharmacy Barton, Garry; University of East Anglia, Norwich Medical School Hammad, Eman; University of Jordan Desborough, James; University of East Anglia, School of Pharmacy Holland, Richard; University of East Anglia, Norwich Medical School Howe, Helen; Cambridge University Hospitals NHS Foundation Trust, Pharmacy Nunney, Ian; University of East Anglia, Norwich Medial school Irvine, Lisa; University of East Anglia, Norwich Medical School
Primary Subject http://bmjopen.bmj.com/ Health services research Heading:
Secondary Subject Heading: Evidence based practice, Health economics, Pharmacology and therapeutics
medicines reconciliation, pharmacy, cost-effectiveness, randomized Keywords: controlled trial, pilot study
on September 27, 2021 by guest. Protected copyright.
For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 20 BMJ Open
BMJ Open Version 1 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Pharmacy provided medicines reconciliation within 24 hours of admission: A 4 randomised controlled pilot study 5 1 2 1 3 4 2 6 Brit Cadman , David Wright , Amanda Bale , Garry Barton , James Desborough , Eman A 5 3 1 3 3 4 7 Hammad , Richard Holland , Helen Howe , Ian Nunney , Lisa Irvine 8 9 1 Pharmacy Department, Cambridge University Hospitals, Cambridge, UK 10 2 School of Pharmacy, University of East Anglia, UK 11
12 3 Norwich Medical School, University of East Anglia, UK 13 14 4 Norwich Clinical Trials Unit, University of East Anglia, UK 15 For peer review only 16 5 School of Pharmacy, University of Jordan, Jordan 17 18 6 Patient and Public Involvement Representative 19 20 21 Correspondence to: Professor David Wright, School of Pharmacy (Room 0.17), University of 22 East Anglia, Norwich Research Park, Norwich, NR4 7TJ [email protected] Telephone: 01603 592042Fax: 01603 592003 23
24 Ms Brit Cadman 25 Consultant Pharmacist and Principal Investigator 26 Pharmacy Department 27 Box 104 28 Cambridge University Hospitals 29 Cambridge 30 CB2 0QQ 31 [email protected] 32 01223 217980 33 http://bmjopen.bmj.com/ 34 Ms Amanda Bale 35 Senior Research Assistant Pharmacy Department 36 Box 104 37 Cambridge University Hospitals 38 Cambridge 39 CB2 0QQ 40 [email protected] 41 01223 217980 42 on September 27, 2021 by guest. Protected copyright. 43 Dr Garry Barton 44 Reader in Health Economics 45 Norwich Medical School 46 Faculty of Medicine and Health Sciences 47 University of East Anglia Norwich 48 NR4 7TJ 49 [email protected] 50 01603 591936 51 52 Dr James Desborough 53 Senior Lecturer in Pharmacy Practice 54 School of Pharmacy (Room 0.14) 55 University of East Anglia 56 Norwich Research Park 57 Norwich 58 NR4 7TJ 59 60 1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 20
BMJ Open Version 1 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 [email protected] 4 01603 593413 5 6 Dr Eman Hammad 7 Assistant Professor of Clinical Pharmacy and Health Economics Department of Biopharmaceutics and Clinical Pharmacy 8 Faculty of Pharmacy University of Jordan 9 Amman 11942 10 Jordan 11 [email protected] 12 +9626 5355000 (ext.23369) 13 14 Professor Richard Holland 15 ProfessorFor of Public Health peer Medicine review only 16 Norwich Medical School (Room 1.07A) 17 University of East Anglia 18 Norwich Research Park 19 Norwich 20 NR4 7TJ [email protected] 21 01603 593574 22 23 Professor Helen Howe 24 Professor of Pharmacy Practice 25 School of Pharmacy 26 University of East Anglia 27 Norwich 28 NR4 7TJ 29 [email protected] 30 31 Mr Ian Nunney 32 Lecturer Norwich Medical School (Room 0.30) 33 University of East Anglia http://bmjopen.bmj.com/ 34 Norwich Research Park 35 Norwich 36 NR4 7TJ 37 [email protected] 38 01603 597282 39 40 Miss Lisa Irvine 41 Senior Research Associate in Health Economics 42 Norwich Medical School on September 27, 2021 by guest. Protected copyright. 43 University of East Anglia 44 Norwich Research Park 45 Norwich NR4 7TJ 46 [email protected] 47 01603 591280 48 49 MeSH words: Medication reconciliation, cost effectiveness, pharmacy, randomised controlled 50 trial, pilot study. 51 52 Word count: 4165 excluding figures, tables and references 53 54 55 56 57 58 59 60 2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 20 BMJ Open
BMJ Open Version 1 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Abstract – Word count 279 4 5 Background 6 The UK government currently recommends that all patients receive medicines reconciliation 7 (MR) from a pharmacy within 24 hours of hospital admission. This is currently not undertaken 8 for every patient and the associated cost effectiveness is unknown. A pilot study to inform the 9 design of a future randomised controlled trial to determine effectiveness and cost 10 effectiveness was undertaken.
11 Method 12 Patients were recruited seven days a week from five adult medical wards in one hospital over 13 a 9 month period and randomised using an automated system to intervention (MR within 24 14 hours of admission and at patient discharge) or usual care which may include MR (control). 15 RecruitmentFor and retention peer rates were determined. review Length of stay only (LOS), quality of life (via 16 EQ 5D 3L), unintentional discrepancies (UDs) and emergency re admission (ER) within 3 17 months were all tested as outcome measures. The feasibility of identifying and measuring 18 intervention associated resources was determined. 19 20 Result 21 200 patients were randomised to either intervention or control. Groups were comparable at 22 baseline.The intervention resolved 250 of the 255 UDs identifed at admission. Only 2 UDs were identifed in the intervention group at discharge compared with 268 in the control. The 23 median LOS was 94 hours in the intervention arm and 118 hours in the control, with ER rates 24 of 17.9% and 26.5%,respectively. Assuming 5% loss to follow up 1120 patients (560 in each 25 arm) are required to detect a 6% reduction in 3 month ER rates. 26 27 Conclusions 28 The results suggest that changes in outcome measures resulting from MR within 24 hours 29 were in the appropriate direction and readmission within 3 months is the most appropriate 30 primary outcome measure. A future study to determine cost effectiveness of the intervention 31 is feasible and warranted. 32 33 http://bmjopen.bmj.com/ 34 35 Strengths and limitations 36 • Pilot randomised controlled trial • Intervention fidelity enhanced through competency assessment of medicine reconciliation 37 providers 38 • Robust process for identifying unintentional discrepancies at each stage developed to 39 prevent results contamination 40 • Pragmatic design and therefore elements of the intervention found within the control arm 41 • Limited response to request for patient data 3 months post discharge on September 27, 2021 by guest. Protected copyright. 42 43 44 Main points 45 • No robust evidence exists for the cost effectiveness of medicines reconciliation provided 46 at both admission and discharge to hospital by members of the pharmacy team 47 • This paper reports a pilot study to determine the design of a future RCT of medicines 48 reconciliation 49 • The anticipated number of patients were recruited and the data obtained was of sufficient 50 quality to enable translation into a future study 51 • The control arm, as part of ‘usual care’, received some level of medicines reconciliation 52 however this seemed to have minimal impact on unintentional discrepancies 53 • Rehospitalisation within 3 months is the most suitable primary outcome measure for a 54 future definitive study 55 56 57 58 59 60 3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 20
BMJ Open Version 1 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Background 4 Medicines Reconciliation (MR) is defined by the World Health organisation as ‘the formal 5 process in which healthcare professionals partner with patients to ensure accurate and 6 complete medication information transfer at the interfaces of care’.(1) Researchers have 7 identified unintentional error rates within hospital prescriptions on admission of between 30 8 70% (2 5) and these can include omission of usual medicines, prescription of incorrect 9 dosages and addition of medicines which have not been previously prescribed. The majority of these errors are believed to result from deficiencies in the MR process with such errors 10 known to contribute to patient morbidity and mortality and increase the length of hospital stay 11 (LOS).(6 10) 12 13 Consequently, within the UK, patient safety guidance issued by the National Patient Safety 14 Agency (NPSA) and the National Institute of Health and Care Excellence (NICE) 15 recommendedFor that policies peer for MR should review be implemented in hospitals only for all adult patient 16 admissions, and that pharmacy should be involved in the MR process and that this should be 17 within 24 hours of admission.(11) NICE guidance on pharmacy involvement was based 18 primarily on one randomised controlled trial (RCT) which demonstrated that the inclusion of 19 the pharmacist in MR reduced the error rate from 44% to 19%.(2) The threshold of a cost per 20 QALY of between £20000 and £30000 used by NICE when deciding which interventions the 21 NHS should adopt (12) was not used to inform this recommendation as robust data to support 22 this was not available.
23 Recent systematic reviews of hospital based medicines reconciliation practices consistently 24 demonstrate a reduction in medicine discrepancies, however this has not been shown to 25 result in reductions in post discharge healthcare utilisation.(13, 14) When considering only 26 pharmacist led medicines reconciliation interventions a 19% reduction in the rate of all cause 27 readmissions was seen but similarly pooled data on mortality and composite readmission and 28 emergency department visit did not find in favour of either pharmacist led intervention or 29 usual care.(15) 30 31 With the significant resources required for pharmacists to deliver medicines reconciliation 32 services it is not adequate to demonstrate reduction in errors (which may or may not result in 33 adverse drug events), it is also important in resource limited health systems to show that such http://bmjopen.bmj.com/ 34 investment constitutes value for money. There is currently a recognised lack of evidence supporting the cost effectiveness of medicines reconciliation interventions provided by 35 pharmacists.(16) A model built in 2008 to estimate the likely cost effectiveness of preventing 36 medication error at hospital admission using medicines reconciliation suggested that a 37 pharmacist based intervention is likely to be cost effective but was based solely upon USA 38 error data and made assumptions concerning the severity of errors prevented.(17) 39 40 Despite national guidance few hospitals are providing MR as envisioned by NICE for all 41 patient admissions.(18) Thus, it would appear that further high quality evidence which 42 demonstrates cost effectiveness is required to ensure that resources are appropriately on September 27, 2021 by guest. Protected copyright. 43 allocated to this service in order to meet national recommendations. The most suitable 44 approach to determine the cost effectiveness of a complex intervention such as medicines 45 reconciliation is to perform a RCT which collates data on cost from the appropriate 46 perspective and outcomes which are most proximal to the intervention. Within the UK it is 47 additionally necessary to collect data on quality of life and mortality to enable the cost per quality adjusted life year to be estimated. Recent guidance however suggests that before an 48 RCT of such nature is undertaken feasibility testing and piloting are required.(19) 49 50 Aims 51 The aim of this study was therefore to pilot a randomised controlled trial to inform the design 52 of a future definitive study to determine the clinical and cost effectiveness of pharmacy 53 providing a full MR service on admission and at discharge. The objectives of the pilot study 54 were to 55 • identify the most suitable outcome measure for a future definitive trial with respect to 56 proximity and response to the intervention and quality of data obtained 57 • determine potential recruitment and retention rates 58 59 60 4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 20 BMJ Open
BMJ Open Version 1 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 • develop and test the process for measuring resource utilisation associated with the 4 intervention and use of other NHS services 5 6 7 Method 8 9 Study design and location 10 The trial was randomised controlled pilot study undertaken at Cambridge University Hospitals 11 NHS Foundation Trust (CUHFT) on five adult medical wards from a range of medical 12 specialities where patients did not routinely receive MR from a pharmacist within 24 hours of 13 admission. One similar ward was identified as a “backup”, in the eventuality that one of the study wards was closed for any reason (e.g. norovirus outbreak) during the recruitment 14 period. 15 For peer review only 16 Intervention 17 A standard operating procedure (SOP), outlined in Figure 1, was used to deliver MR by a 18 trained medicines reconciliation pharmacist (MRP) within 24 hours of admission (including 19 weekends) and at the point of transfer of care out of hospital, or as soon as possible following 20 patient discharge from hospital to the next care provider. MRPs covered for each other’s 21 holidays, sick leave and absences wherever possible. 22 23 MRPs recorded all unintentional discrepancies they identified between the information they 24 collated and the inpatient medication chart on admission and again any differences between 25 the inpatient chart and discharge letter. MRPs followed up on all identified unintentional 26 discrepancies to ensure that they were addressed.
27 Control 28 Patients in the control arm received usual care which may or may not consist of medicines 29 reconciliation and where it was provided it may not have occurred within 24 hours and could 30 either be delivered by a pharmacist or pharmacy technician. The MRPs within the 31 intervention arm did not deliver MR to control patients and the SOP used for study 32 intervention purposes was not automatically followed within the control arm. For the purposes 33 of the study all MR details regarding interventions undertaken within the control arm were http://bmjopen.bmj.com/ 34 recorded and costed. 35 36 Intervention fidelity 37 To enhance intervention fidelity all five MRPs were observed by the Principal Investigator (PI) 38 on at least 3 occasions to confirm adherence to the SOP. All MRPs had provided MR to more than 30 patients in the year previous to delivering the intervention for the trial. 39
40 Recruitment 41 A recruitment target of 200 patients was set for the nine month pilot phase. Study wards were on September 27, 2021 by guest. Protected copyright. 42 visited every morning by the research assistant (RA) during the study period to identify 43 potential participants. The nurse in charge of the ward confirmed that it was appropriate for 44 the patient to be approached regarding the study. Patients were recruited based on the 45 following inclusion and exclusion criteria. 46 47 Inclusion criteria: 48 • Adult (≥18 years of age) 49 • Admitted with prescribed medicines to one of the five medical wards 50 • Patient had not already received MR from the pharmacy team as part of routine 51 pharmaceutical input at the time of recruitment 52 • Identified from hospital computer system as having been admitted straight from the 53 Emergency Department (ED) to one of the five participating wards within the previous 24 54 hours 55 56 With the intervention required to be delivered within 24 hours of admission, patients were 57 given a maximum of 2 hours to consider whether they wished to participate. Intervention 58 patients received MR in accordance with the SOP. Information obtained by the RA for the 59 60 5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 20
BMJ Open Version 1 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 purposes of the study was given to the MRP pharmacist prior to the visit to prevent 4 duplication of effort and to ensure the patient was not interviewed for the same information 5 twice. 6 7 Randomisation was performed utilising the Norwich Clinical Trials Unit automated service with patients stratified by ward. When wards were later closed for infection control reasons, 8 participants on the “backup” ward were randomised and stratified as if they had entered the 9 closed ward. 10 11 Patients randomised to the control group received usual care; this may have included 12 elements of MR by members of the pharmacy team and in some cases may have occurred 13 within 24 hours of admission but post randomisation. Whilst this may have affected patients in 14 the control arm an intention to treat analysis was performed and consequently for the purpose 15 of the analysisFor patients peerthey remained inreview their allocated arm. only 16 17 Outcome measures 18 Although undertaken as a pilot study with study aims to identify the most suitable outcome 19 measure, length of stay (LOS) was nominally selected as the primary outcome measure for 20 this pilot trial Secondary outcome measures were unplanned (emergency) re admission at 3 months, quality of life (EQ 5D 3L) and unintentional discrepancies. 21
22 Data collection 23 To enable comparison of intervention and control groups, age, gender, primary reason for 24 admission, all co morbidities and the admission ward was recorded. Additionally consented 25 patients or their third party consultees completed a quality of life score (EQ 5D 3L) (20) on 26 admission, including the related visual analogue scale (VAS). 27 28 LOS, reported in hours, was calculated as the difference in time from arrival at the hospital to 29 the time of discharge as recorded in the hospital information system (HISS). Unplanned re 30 admissions to the intervention hospital within the 3 months post discharge were also obtained 31 from HISS. EQ 5D 3L responses were also obtained via postal survey 3 months post 32 discharge, allowing the calculation of Quality Adjusted Life Years (QALYs)(21) in the subsequent economic evaluation. 33 http://bmjopen.bmj.com/ 34 Unintentional discrepancy identification 35 To enable the identification of unintentional discrepancies the following information was 36 photocopied by the research associate (RA) for all consented patients (both intervention and 37 control) and stored securely: 38 • All versions of the inpatient medication chart(s) and discharge letters, 39 • Medical notes during admission 40 • GP medication list on admission 41
• GP medication list at 3 month post discharge (when received from GP surgery) on September 27, 2021 by guest. Protected copyright. 42 • Any additional medicines related information brought in by the patient on admission e.g. 43 copies of labels from patient medicines, handwritten or typed medicine lists. 44 45 The RA was a trained nurse and consequently was experienced in medical data collation and 46 extraction. 47 48 3 months post discharge the stored information was used to develop an ‘accurate medication 49 list’ for the control arm patients by the research team on admission and at discharge. These 50 were then compared with the inpatient chart on admission and discharge letter to identify any 51 discrepancies. Medical notes were subsequently reviewed, unblinded to group allocation, to 52 enable differentiation between those which were unintentional i.e. discrepancies which could 53 not be explained from the information available and those which were intentional. 54 55 The GP medication list obtained 3 months post discharge (where available) was used to enable the identification of discrepancies which still remained at 3 months. Without access to 56 GP medical notes it was not possible to establish whether identified discrepancies were 57 unintentional. 58 59 60 6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 20 BMJ Open
BMJ Open Version 1 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 4 All potential discrepancies which were identified at the 3 month point as having translated into 5 the patient notes were reported to the patient’s GP. 6 7 Resource use The time taken by the pharmacy team to deliver the MR was monitored in both the 8 intervention and control group (if applicable). Additionally, the following items were requested 9 in both groups: 10 • Time in hospital (Length of stay) 11 • Medication (in patient medication and GP medication list at 3 months) 12 • Re hospitalisations 13 • Other health care contacts 14 15 With the exceptionFor of thepeer final item, which review included all health professional only contacts and was 16 requested from the participant at 3 months post discharge, these were extracted from medical 17 records. 18 19 Sample size calculation 20 As a pilot study a formal power calculation was not performed. The consequences for the 21 precision of the primary outcome variable (LOS) of the choice of sample size were however 22 estimated. Summary statistics on LOS taken from a study undertaken at St James Hospital, 23 Dublin (22) gave quartiles for two groups as (3, 7, 5) and (2, 5, 12) days. To derive an estimate of variability from this, an underlying log normal distribution was assumed, which is 24 consistent with the position of the medians, and its standard deviation estimated using the 25 geometric mean of the ratios between the upper and lower quartiles. This produced an 26 estimated SD of 1.26 which implied that a comparison between 2 groups of 100 patients, 27 would provide an expected half width for the 95% confidence interval of the difference 28 between group means (on the natural log scale) of 0.35. Translated back to the confidence 29 interval for the ratio of average length of stay between the intervention and control groups 30 would extend by a factor of about 1.4 either side of the point estimate. 31 32 Data analysis 33 As a pilot study descriptive statistics were utilised to determine the suitability of the different http://bmjopen.bmj.com/ 34 outcome measures and report the variation in the differences in order to determine a sample 35 size for a future randomised controlled trial. Similarly, completion rates are reported for each source of resource use data and the EQ 5D 3L. 36
37 The primary outcome variable, length of stay, was reported using median, arithmetic mean 38 and geometric mean. The rate of Trust re admissions, Trust emergency re admissions and 39 mortality are also reported for both arms, along with the mean change in the VAS from the 40 EQ 5D 3L. 41 42 Results on September 27, 2021 by guest. Protected copyright. 43 Nine hundred and nineteen patients were assessed for eligibility of which 224 did not meet 44 inclusion criteria. 200 out of 310 patients who were subsequently approached by the RA 45 consented to take part in the study (Figure 2). Of those patients identified as potentially 46 eligible but not approached this was primarily because either the nurse in charge of the ward 47 advised that the patient was not suitable to be approached or that the end of the 24 hour 48 window for intervention was due to expire. 49 Recruitment took place between July 2012 and April 2013 (9 months and 2 weeks), resulting 50 in a recruitment rate of 5.2 patients per seven days. 51 52 There was one post randomisation exclusion in each arm (they were subsequently found not 53 to meet the inclusion criteria), and were accordingly excluded from all analyses. Table 1 54 thereby summarises the characteristics of the remaining198 participants (96 intervention 102 55 control). The groups were broadly comparable. 56 57 One patient in the intervention arm did not receive the MR as he was not on the ward when 58 the pharmacist visited. Data was erroneously not collected for this patient post randomisation. 59 60 7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 20
BMJ Open Version 1 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 The time taken by the pharmacist to deliver the intervention was recorded for the remaining 4 95 intervention participants, where the mean total time was 48.6 minutes (range 2 to 195 5 minutes). In the control group, 31 participants received input from a pharmacist (mean 6 reported time 15 minutes) and 31 had input from a pharmacy technician (mean reported time 7 12 minutes), for 25 of these patients this occurred within the 24 hour window.
8 During the initial hospital admission three participants died, with a further 11 deaths during the 9 three months follow up period (see Figure 2). In total six intervention patients died and 8 10 control patients over the trial period (p=0.78, Fisher’s Exact). Additionally, in the control arm, 11 one participant was lost to follow up (address not known) and one participant withdrew. After 12 taking account of the 1 patient who did not receive the MR and a further loss to follow up in 13 the intervention arm, this left a total of 88 available cases in the intervention arm and 92 14 available cases in the control arm at final assessment point. In terms of outcomes, there was 15 complete Fordata available peer on LOS and re admission review data. only 16 17 Table 2 provides a summary of the unintentional discrepancies which were identified at each 18 stage. 16 (16%) intervention patients had no discrepancies at admission and discharge, 19 compared to 12 (12%) control patients. Overall, 2 were known to remain at discharge in the 20 intervention arm compared with 268 in the control group. Neither of the unintentional discrepancies in the intervention arm identified at discharge were present at 3 months. 21
22 One hundred and fifty four unintentional discrepancies identified in the control arm were 23 potentially related to medicines to be prescribed post discharge i.e. the remainder related to 24 medicines prescribed during admission only. Due to the limited number of GP records 25 provided at 3 months data was only available for 82 (53.2%) unintentional discrepancies in 26 the control arm at 3 months and 37 out of the original 154 (24%) were found from the medical 27 records provided to have been propogated into the patient notes. 28 29 Table 3 provides a comparison of patient outcomes for the intervention and control group. 30 The results suggest that LOS, mortality and rehospitalisation rates tend to be lower in the 31 intervention arm although statistical significance was not demonstrated as 95% Confidence 32 Intervals overlapped. Based on those who responded, both groups had a higher mean quality of life (based on the VAS) at the 3 month follow up with improvement higher in the control 33 arm. This difference between groups was not significant. http://bmjopen.bmj.com/ 34 35 With regard to the response rates for the resource use data there were complete data 36 available (for those upon whom it was requested) for length of stay, medication data as part of 37 the original admission, re admissions (to the same hospital) and the intervention pharmacist 38 times. Of the 62 controls for whom pharmacist/pharmacy technician review occured the times 39 were missing for N=27. At 3 month follow up, medication data were retrieved from GPs for 86 40 participants (94.5% of those from whom it was requested) in each arm and 133 participants 41 completed and returned both the EQ 5D 3L and the health resource use questionnaires (66 42 Intervention, 67 Control) (73.5% of those requested from all living participants). on September 27, 2021 by guest. Protected copyright. 43 44 Based on the pilot data we calculate that for a full trial 1120 patients would need to be 45 recruited to detect a 6% reduction in 3 month re admission rates from a starting point of 26% with 90% power, using a 5% significance level and assuming 5% loss to follow up. 46
47 Discussion 48 The results from this study which was performed to inform the design of a future randomised 49 controlled trial suggest that even though medicines reconciliation activities are taking place 50 such a trial is feasible with reasonable recruitment and retention rates and that both cost and 51 outcome data can be effectively obtained. 52 53 We consider that emergency rehospitalisation within 3 months would be the most appropriate 54 primary outcome measure for such a trial, as unlike the other outcome measures tested, it 55 reflects all of the MR activity which occurs in both secondary and primary care. Furthermore, 56 data collection was complete and this is a patient orientated outcome, unlike medication 57 errors which whilst representing a patient safety issue are a measure of the prescribing 58 process. 59 60 8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 20 BMJ Open
BMJ Open Version 1 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 4 Due to the need to recruit patients before they received MR as usual care this reduced the 5 generalisability of the sample with all of those who had already received MR being 6 automatically excluded. This was further compounded by recruitment activities which were 7 focussed towards mornings.
8 With a requirement that all patients receive MR within 24 hours of admission it is not 9 surprising that two thirds of the control arm received some form of MR during their 10 hospitalisation, a quarter of which received this within 24 hours. The majority of 11 discrepancies identified in the control arm by the researchers were however found not to have 12 been resolved and therefore reasons for the relative ineffectiveness of the control arm MR 13 requires further elucidation. 14 15 The averageFor time spent peer on the intervention review found within this study only was very similar to that 16 reported in a MR time and motion study(23) but three times greater than that in the control 17 arm. It has been suggested that organisations are probably unlikely to repeat the benefits 18 from MR services reported in the literature if there are deficiencies in intervention intensity 19 and breadth.(24) The results from the control arm of this study support this assertion and 20 suggest that if MR of a similar nature to that seen in the intervention arm was to be shown to be cost effective then this would require significantly more pharmacist time. 21
22 The proportion of patients screened for eligibility and eventually recruited to the study was 23 20% and this could have been improved by increasing the number of weekends covered 24 (from 80% to 100%). Although some patients were not suitable for inclusion in the study as 25 they had already been reviewed by a member of the pharmacy team prior to being 26 approached by the RA, there was a sufficient number of patients available for recruitment. 27 28 When identifying unintentional errors we have assumed that the MRP in the intervention arm 29 had created an accurate list with no errors. This is an unrealistic assumption and ideally 30 within a definitive study the data should have been reviewed independent of the service and 31 blinded to group allocation. The unblinded identification of medication discrepancies and 32 inability to confirm intentional or unintentional nature of errors in many instances also means that the data on unintentional discrepancies must be treated with further caution. 33 http://bmjopen.bmj.com/ 34 The process of collating all medicines related data at different time points and sealing it until 3 35 months post discharge provides an opportunity to identify discrepancies without adversely 36 affecting the intervention. Whilst the identification of unintentional discrepancies should be 37 undertaken blind of allocation the resources required for this may not be warranted when 38 considering that the intentional or unintentional nature of the discrepancy cannot always be 39 accurately determined. 40 41 Considering evidence published post study completion, limiting the study population to those 42 over the age of 70 and including a post discharge telephone call as part of the intervention on September 27, 2021 by guest. Protected copyright. 43 may have further enhanced the intervention.(13) 44 45 In line with previous research the intervention prevented a large number of unintentional medicines related discrepancies both during admission and post discharge.(3 6) The data 46 obtained suggests that just less than a quarter of unintentional discrepancies identified at 47 discharge were found to actually translate into primary care records at three months. Whilst 48 reasons for non translation were not elucidated the research suggests that the use of number 49 of unintentional discrepancies at discharge as an outcome measure may over emphasise the 50 problem. 51 52 Not all primary care medication lists were made available to researchers and approaches to 53 addressing this will require consideration for a future definitive study. Utilising data on 54 ‘unplanned readmission at 3 months’ would seem to be the most appropriate primary 55 outcome measure for a future definitive study as it is a patient orientated outcome which is 56 most likely to reflect the effect of errors which occur at all stages of the process. Similar to 57 hospital readmission, LOS is a cost which would be captured in any cost effectiveness 58 analysis. LOS was however found to be largely skewed by a small number of individuals who 59 60 9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 20
BMJ Open Version 1 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 were admitted for extended periods. Furthermore, it is not affected by errors which translate 4 into primary care i.e. does not reflect the full impact of the service. Within the UK hospitals are 5 penalised for unplanned readmissions within one month of discharge and therefore collection 6 of this data may also be warranted for a UK based definitive study. 7 8 Unplanned readmission has been used as a primary outcome measure within other similar MR trials (25 29). Whilst differences in unplanned readmission at three months have been 9 demonstrated,(28, 29) this has frequently not been the case at either one or six months.(25 10 27) Within the first month patients may still be using the medication they were discharged with 11 and consequently this may lessen the impact of errors on the discharge letter resulting from 12 incorrect translation into primary care. Kwan et al. after systematically reviewing the literature 13 suggest that unplanned readmission data should probably be collected for more than 30 days 14 post discharge.(14) Unplanned readmission at six months will be affected by more factors 15 unrelated Forto the index admissionpeer than at review3 months and therefore onlyit may be more difficult to 16 identify the impact of MR intervention using this outcome measure. 17 18 Quality of data collection with respect to LOS, mortality and readmission rates was high as 19 this information was available from hospital records. Resource utilisation data and quality of 20 life scores were however only available for two thirds of participants at follow up. 21 Consequently researchers powering a definitive study on readmission rate will need to consider the possibility of there being more missing data for the cost effectiveness analysis 22 and imputation of missing data (30) may be necessary to address this. 23 24 Acknowledgements: 25 The authors would like to thank Dr Tracey Sach for her contribution in the design of the health 26 economics element of the project and in the development of the original grant proposal, 27 Antony Warford for his contribution to the management, delivery and dissemination of the 28 project as the patient and public involvement representative and Kellie Hempstead and the 29 pharmacists volunteering to provide the intervention at weekends in particular Lynn Martin for 30 their efforts in completing this study. The input by the Norwich Clinical Trials Unit throughout 31 the study is also acknowledged. 32 33 Contributors: http://bmjopen.bmj.com/ 34 BC, EH, DW, JD, RH participated in the conception and design of this study. AB, BC, DW,GB, 35 IN, LI participated in the analysis of the data and interpretation of the results AB, BC, DW 36 drafted the article and all other authors performed critical review of the article. All authors approved the submission of this article 37
38 Funding: 39 This article presents independent research commissioned by the National Institute for Health 40 Research (NIHR) under Research for Patient Benefit. The views expressed in this publication 41 are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department on September 27, 2021 by guest. Protected copyright. 42 of Health. 43 44 Competing interests: 45 None 46 47 Ethics approval: 48 Ethical approval was received on 14th June 2012 from NRES Committee East of England 49 Essex Research Ethics Committee. Reference number: 12/44/0143 50 51 Provenance and peer review 52 The trial was registered on 5th August 2012: ISRCTN23949491 53 The trial received competitive funding from NIHR Research for Patient Funding stream and 54 was peer reviewed as part of this process. 55 56 57 58 59 60 10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 20 BMJ Open
BMJ Open Version 1 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Figure 1 Outline of service standard operating procedure 4 5 Admission 6 • Patient interviewed to confirm allergy status and medicines being taken 7 • GP letter/fax reviewed to confirm allergy status and medicines prescribed 8 • Medical notes reviewed to confirm any intentional medication changes due to clinical 9 status of patient 10 • Other sources of information may have been used depending on availability and 11 relevance to current admission by virtue of the date, and included o previous electronic discharge letter 12 o 13 clinic letter o patient’s own drugs (PODs) 14 o patients relative or carer 15 oFor nursing home peer record review only 16 o medication administration record (MAR) 17 o modified dosage system (MDS) 18 • An accurate medication list was documented in the medical notes by the intervention 19 pharmacist 20 21 Discharge 22 • Discharge letter compared with the final drug chart to ensure correct 23 • Medical notes reviewed 24 • Changes to medicines field on the electronic system was completed by the intervention 25 pharmacist to communicate intentional medication changes to the GP 26 • All medicines were added to the discharge letter to ensure a complete record at the point 27 of transfer of care 28 29 30 31 32 33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40 41 42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 20
BMJ Open Version 1 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Figure 2 Consort diagram 4 5 6 Enrolment 7 Assessed for eligibility (n=919) 8 Excluded (n=719) 9 • Not meeting inclusion criteria (n=224) • Declined to participate (n=110) 10 • Other reasons (n=385) 11 Allocation Randomised (n=200) 12 13 Excluded post randomisation (n=1 #95) Excluded post randomisation (n=1 #154) Not meeting inclusion Not meeting inclusion 14 15 For peer review only
16 Intervention arm (n=96) Control arm (n=102) 17 • Received intervention < 24 hr (n=95) • < 24 hour by pharmacist (n=13) 18 • Received intervention > 24 hr (n=1) #75 • < 24 hour by technician (n=22) • > 24 hour by pharmacist (n=18) 19 • > 24 hour by technician (n=9) 20 • No MR performed (n=40) 21 Discharge follow up 22 23 24 Lost to follow up (n=0) Lost to follow up (n=0) Death (n=2) #108 #136 Death (n=1) #20 25
26 3 month follow up 27
28 Lost to follow up (n=1) #70 Lost to follow up (n=1) #83 29 Withdrawn (n=0) Withdrawn (n=1) #173 30 Death (n=4) #51, 68, 82, 97 Death (n=7) #16, 55, 76, 80, 86, 99, 198
31 Data availability 32 33 Complete self reported cost data (n=66) Complete self reported cost data (n=66) http://bmjopen.bmj.com/ 34 EQ5D data available at 3 months (incl. EQ5D data available at 3 months (incl. deaths) (n=72) deaths) (n=75) 35 Medication data received from GPs at 3 Medication data received from GPs at 3 36 months (n=78) months (n=85) Medication data not provided at 3 months Medication data not provided at 3 months 37 (n=12 plus n=6 RIPs) (n=9 plus n=8 RIPs) 38 39 40 41 42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 20 BMJ Open
Version 27 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Table 1 Comparison of demographics at baseline: 4 5 Demographic Measure Intervention Control (n=102) 6 (n=96) 7 Female No. (%) 45 (46.9) 60 (58.8) 8 Age (years) Mean (SD) 67.6 (19.0) 65.4 (20.2) 9 Regular medicines Mean (SD) 5.84 (4.07) 6.67 (4.64) 10 As required medicines Mean (SD) 0.85 (2.08) 0.95 (2.53) 11 12 Ward 13 1 No. (%) 26 (26.8) 27 (26.2) 14 2 No. (%) 30 (30.9) 30 (29.1) 15 For peer3 reviewNo. (%) 10 (10.3) only 9 (8.7) 16 4 No. (%) 14 (14.4) 16 (15.5) 17 5 No. (%) 16 (16.4) 14 (13.5) 18 6 No. (%) 1 (1.03) 7 (6.8) 19 20 EQ5D Quality of life (Visual Mean (SD) 55.9 (23.2) 54.7 (23.5) 21 Analogue Scale) 22 Reason for admission 23 Lower respiratory tract infection No. (%) 9 (9.2) 11 (10.7) 24 Troponin –ve chest pain No. (%) 7 (7.2) 4 (3.9) 25 Heart failure No. (%) 2 (2.1) 5 (4.9 ) 26 Exacerbation Chronic No. (%) 4 (4.1) 3 (2.9) 27 Obstructive Pulmonary Disease 28 Other No. (%) 73 (75) 79 (76.7) 29 30 31 32 33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40 41 42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 20
Version 27 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Table 2 Comparison of unintentional discrepancies (UDs) 4 5 Unintentional Intervention Control 6 discrepancies 7 No. UDs No. No. per No. UDs No. No. per patients patient patients patient 8 9 Admission 255 95 2.80 309 102 3.0 10 Resolved during 250 95 2.74 Unknown 11 hospital stay 12 13 Remaining at 2 91 0.02 268 99 2.71 14 discharge 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40 41 42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1.51 (4.36) 1.51
20.84 (38.75)20.84
Mean differenceMean (SE of thedifference)
7.15 (26.2)7.15 8 (7.8)8 27 (26.7)27 37 (36.6)37 203.9 (246.8)203.9 117 (13 1546) 117 137.3) 109.3 (87.0 to 109.3 Control
68 95 101 101 102 102 102 N
5.64 (23.6) 5.64
6 (6.3) 6 17 (17.7)17
30 (31.6)30 224.8 (293.1)224.8 94.0 (12 – 94.0 1077) 99.6 (76.59 to 129.63)99.6 Intervention
15 63 95 95 95 95 95 95 N BMJ Open http://bmjopen.bmj.com/
Mean (SD)Mean No. (%) No.
No. (%) No.
No. (%) No. Median (range)Median Geometric (95%MeanGeometric CI) Measure Arithmetic mean (SD)mean Arithmetic on September 27, 2021 by guest. Protected copyright.
For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only
Visual Visual Scale Analogue
readmissions(emergency) readmissions
Quality of life Quality of life from changebaseline (high better) score Mortality Hospital Hospital Length of (hours) stay Outcome Version 27 27 Version Table 3 Table 3 outcome Comparison of measures Page 15 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 16 of 20
Version 27 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 References 4 1. World Health Organisation. High 5s Fact Sheet: The High 5s Assuring Medication 5 Accuracy at Transitions of Care: Medicines Reconciliation Standard Operating Procedure 6 [Available from: http://www.who.int/patientsafety/implementation/solutions/high5s/h5s fact 7 sheet.pdf?ua=1. 8 2. The University of Sheffield SoHaRRS. A systematic review of the effectiveness and costeffectiveness of interventions aimed at preventing medication error (medicines 9 reconciliation) at hospital admission. NICE. 2009. 10 3. Tim Dornan DA, Heather Heathfield, Penny Lewis, Jon Miles, David Taylor, Mary 11 Tully, Val Wass. Final report. An in depth investigation into causes of prescribing errrors by 12 foundation trainees in relation to their medical education. EQUIP Study. General Medical 13 Council: 2009. 14 4. Gray E, Ullah S, Lwin A, Javed S, Farrow C, Southgate J, et al. Prescription Errors in 15 Patients onFor an Acute Medicalpeer Assessment review Unit. Royal College ofonly Physicians of Edinburgh. 16 2007:37:305 8. 17 5. McLeod SE, Lum E, Mitchell C. Value of Medication Reconciliation in Reducing 18 Medication Errors on Admission to Hospital. Journal of Pharmacy Practice and Research. 19 2008;38(3):196 9. 20 6. Pippins JR, Gandhi TK, Hamann C, Ndumele CD, Labonville SA, Diedrichsen EK, et 21 al. Classifying and predicting errors of inpatient medication reconciliation. J Gen Intern Med. 2008;23(9):1414 22. 22 7. The Audit Comission. A spoonful of sugar Medicines Management in NHS Hospitals. 23 London. 2001. 24 8. Dean B, Schachter M, Vincent C, Barber N. Prescribing errors in hospital inpatients: 25 their incidence and clinical significance. Qual Saf Health Care. 2002;11(4):340 4. 26 9. Department of Health. Management of Medicines A resource to support 27 implementation of the wider aspects of medicines management for the National Service 28 Frameworks for Diabetes, Renal Services and Long term Conditions. 2004. 29 10. Karnon J, McIntosh A, Dean J, Bath P, Hutchinson A, Oakley J, et al. Modelling the 30 expected net benefits of interventions to reduce the burden of medication errors. J Health 31 Serv Res Policy. 2008;13(2):85 91. 32 11. National Institute for Health and Clinical Excellence. Technical patient safety solutions 33 for medicines reconciliation on admission of adults to hospital. NICE; 2007. http://bmjopen.bmj.com/ 34 12. NICE. Guide to the methods of technology appraisal 2013. National Institute of Health and Clinical Excellence (NICE) publications, 2013. 35 13. Mueller SK, Sponsler KC, Kripalani S, Schnipper JL. Hospital based medication 36 reconciliation practices: a systematic review. Arch Intern Med. 2012;172(14):1057 69. 37 14. Kwan JL, Lo L, Sampson M, Shojania KG. Medication reconciliation during transitions 38 of care as a patient safety strategy: a systematic review. Ann Intern Med. 2013;158(5 Pt 39 2):397 403. 40 15. Mekonnen AB, McLachlan AJ, Brien JA. Effectiveness of pharmacist led medication 41 reconciliation programmes on clinical outcomes at hospital transitions: a systematic review 42 and meta analysis. BMJ open. 2016;6(2):e010003. on September 27, 2021 by guest. Protected copyright. 43 16. Hammad EA, Bale A, Wright DJ, Bhattacharya D. Pharmacy led medicine 44 reconciliation at hospital: a systematic review of effects and costs. Research in Social and 45 Administrative Pharmacy. 46 17. Karnon J, Campbell F, Czoski Murray C. Model based cost effectiveness analysis of interventions aimed at preventing medication error at hospital admission (medicines 47 reconciliation). Journal of evaluation in clinical practice. 2009;15(2):299 306. 48 18. Farmer D. 2009 Eastern England Collaborative Audit on Medicines Reconciliation 49 2009. 50 19. Medical Research Council. Developing and evaluating complex interventions 2015 51 [Available from: www.mrc.ac.uk/complexinterventionsguidance. . 52 20. Brooks R. EuroQol: the current state of play. Health Policy. 1996;37(1):53 72. 53 21. Gray AM, Clarke PM, Wolstenholme JL, Wordsworth S. Applied methods of cost 54 effectiveness analysis in healthcare: Oxford University Press; 2011. 55 22. Moloney ED, Bennett K, Silke B. Effect of an acute medical admission unit on key 56 quality indicators assessed by funnel plots. Postgraduate Medical Journal. 2007;83(984):659 57 63. 58 59 60 16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 20 BMJ Open
Version 27 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 23. Meguerditchian AN, Krotneva S, Reidel K, Huang A, Tamblyn R. Medication 4 reconciliation at admission and discharge: a time and motion study. BMC Health Services 5 Research. 2013;13(1):1 11. 6 24. Pevnick JM, Shane R, Schnipper JL. The problem with medication reconciliation. 7 BMJ quality & safety. 2016. 25. Gillespie U, Alassaad A, Henrohn D, Garmo H, Hammarlund Udenaes M, Toss H, et 8 al. A comprehensive pharmacist intervention to reduce morbidity in patients 80 years or older: 9 a randomized controlled trial. Arch Intern Med. 2009;169(9):894 900. 10 26. Spinewine A, Swine C, Dhillon S, Lambert P, Nachega JB, Wilmotte L, et al. Effect of 11 a collaborative approach on the quality of prescribing for geriatric inpatients: a randomized, 12 controlled trial. J Am Geriatr Soc. 2007;55(5):658 65. 13 27. Stowasser DS, M and Collins, DM. A randomised controlled trial of medication liaison 14 services patient outcomes Journal of Pharmacy Practice and Research. 2002;32(2):133 40. 15 28. MakowskyFor MJ, peerKoshman SL, Midodzi review WK, Tsuyuki RT. Capturingonly outcomes of clinical 16 activities performed by a rounding pharmacist practicing in a team environment: the 17 COLLABORATE study [NCT00351676]. Med Care. 2009;47(6):642 50. 18 29. Boso Ribelles V, Montero Hernandez M, Font Noguera I. Evaluation of a plan for 19 cardiology medication reconciliation on admission, and patient information at discharge in a 20 teaching hospital. European Journal of Hospital Pharmacy Practice. 2011;17(1):26 30. 30. Rubin DB. Multiple Imputation for Non response in Surveys: Wiley and Sons; 1987. 21
22 23 24 25 26 27 28 29 30 31 32 33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40 41 42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from Page 1 1 Page Page 18 of 20 * * Reported Reported on page No page on N/A N/A 5 5
7 7 11 11 5 5 4 4 4 5 5 5 5 5 5 6 6 5 5 1 1 5 5 N/A N/A 5 5 3 (for specific guidance see CONSORT for abstracts) for see CONSORT guidance specific (for
BMJ Open http://bmjopen.bmj.com/ on September 27, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only actually administered administered actually Checklist item Checklist interventions interventions describing any steps taken to conceal the sequence until interventions were assigned assigned were interventions until sequence the conceal to stepstaken any describing were assessed assessed were 9 9 containers), numbered sequentially as (such sequence randomallocation the implement to used Mechanism 1a 1a titlethe in trial randomised as a Identification 6a 6a they when and how including measures, outcome secondary and primary pre-specified defined Completely 7b 7b guidelines stopping and analyses interimof any explanation applicable, When 6b 6b reasons with commenced, trial the after outcomes trial to changes Any 4b 4b collected were data the where locations and Settings 3b 3b reasons criteria), with aseligibility (such commencement trial after methods to changes Important 2a 2a of rationale explanation and background Scientific 2b 2b hypotheses or objectives Specific 8a 8a sequence allocation random the generate to used Method 8b size) block and blocking as (such restriction ofany details of randomisation; Type 1b 1b conclusions and results, methods, design, of trial summary Structured No No 11a 11a those providers, care participants, example, (for interventions to after assignment blinded was who Ifdone, N/A Item Item CONSORT CONSORT 2010 checklist of to include wheninformation randomised reporting a trial generation generation concealment concealment mechanism Implementation Implementation 10 to participants assigned and who participants, enrolled who sequence, random allocation the Whogenerated Sequence Sequence Allocation Allocation Blinding Blinding
Randomisation: Randomisation: Sample size size Sample 7a determined was size sample How Outcomes Outcomes Interventions Interventions 5 were they andwhen how including allowreplication, to details sufficient with group foreach Theinterventions Participants Participants 4a for criteria participants Eligibility Trial design design Trial 3a ratio allocation including factorial) parallel, as (such design oftrial Description Methods Methods Background and Background objectives Introduction Introduction
Title and abstract abstract and Title Section/Topic Section/Topic CONSORT 2010 checklist checklist 2010 CONSORT 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from Page 2 2 Page N/A N/A
15 15 7 7 7 7 N/A N/A online online N/A N/A N/A N/A Info submitted submitted Info online
9 9 15 15 8 8 Info submitted submitted Info 15 15 8 8 13 13 N/A N/A
(for specific guidance see CONSORT for harms) for CONSORT see guidance specific (for BMJ Open http://bmjopen.bmj.com/ on September 27, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only by original assigned groups assignedgroups original by pre-specified from exploratory fromexploratory pre-specified assessing outcomes) and how how and outcomes) assessing precision (such as 95% confidence interval) interval) confidence 95% as (such precision were analysed for the primary outcome primary for analysed the were 25 25 funders of role ofdrugs), supply as (such support other and funding of Sources 23 23 registry trial ofname and number Registration 24 24 ifavailable accessed, be can protocol trial full the Where 19 19 group each in effects unintended or harms important All 17b 17b sizesrecommended is effect relative and absolute of both presentation outcomes, binary For 14b 14b stopped was or ended trial the Why 12b 12b analyses adjusted and analyses subgroup as such analyses, foradditional Methods 11b 11b ofinterventions similarity ofthe description Ifrelevant, 17a 17a andits size effect estimated the and group, eachfor results outcome, andsecondary primary each For 13a 13a and treatment, intended received assigned, randomly were who participants of numbers the group, each For 13b 13b reasons with together randomisation, after exclusions and losses group, each For
Funding Registration Other information information Other Interpretation Interpretation 22 evidence relevant other considering and harms, and benefits balancing results, with consistent Interpretation 10 Limitations Limitations Generalisability 21 20 findings trial of the applicability) validity, (external Generalisability ofanalyses multiplicity if relevant, and, imprecision, bias, potential sources of addressing limitations, Trial 9 Discussion Discussion Harms Harms Outcomes and and Outcomes estimation Numbers analysed analysed Numbers 16 was analysis the whether and analysis each in included (denominator) participants of number group, each For Baseline data data Baseline 15 group each for characteristics clinical and demographic baseline showing Atable Recruitment Recruitment 14a follow-up and recruitment periods of the defining Dates Participant flow (a (a flow Participant strongly is diagram recommended) Results Results Statistical methods methods Statistical 12a outcomes secondary and primary for compare groups to used methods Statistical CONSORT 2010 checklist checklist 2010 CONSORT Protocol Ancillary analyses analyses Ancillary 18 distinguishing analyses, adjusted and analyses subgroup including performed, analyses other any of Results Page 19 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from Page 3 3 Page Page 20 of . www.consort-statement.org BMJ Open http://bmjopen.bmj.com/ on September 27, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only recommend reading recommend extensionsCONSORT for cluster trials,randomised non-inferiority trials, and equivalence non-pharmacological treatments, herbal and interventions, pragmatic trials. extensions Additional are forthcoming: for thosefor date references and up to this relevant checklist, to see *We recommendstrongly *We this reading conjunction statement in the with 2010 CONSORT Explanationand forElaboration important clarifications items.allthe If on relevant, also we CONSORT 2010 checklist checklist 2010 CONSORT 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from
Pharmacist provided medicines reconciliation within 24 hours of admission and on discharge: A randomised controlled pilot study
For peer review only
Journal: BMJ Open
Manuscript ID bmjopen-2016-013647.R1
Article Type: Research
Date Submitted by the Author: 07-Nov-2016
Complete List of Authors: Cadman, Brit; Cambridge University Hospitals NHS Foundation Trust, Pharmacy Wright, David; University of East Anglia, School of Pharmacy Bale, Amanda; Cambridge University Hospitals NHS Foundation Trust, Pharmacy Barton, Garry; University of East Anglia, Norwich Medical School Hammad, Eman; University of Jordan Desborough, James; University of East Anglia, School of Pharmacy Holland, Richard; University of East Anglia, Norwich Medical School Howe, Helen; Cambridge University Hospitals NHS Foundation Trust, Pharmacy
Nunney, Ian; University of East Anglia, Norwich Medial school http://bmjopen.bmj.com/ Irvine, Lisa; University of East Anglia, Norwich Medical School
Primary Subject Health services research Heading:
Secondary Subject Heading: Evidence based practice, Health economics, Pharmacology and therapeutics
medicines reconciliation, pharmacy, cost-effectiveness, randomized Keywords: controlled trial, pilot study on September 27, 2021 by guest. Protected copyright.
For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 21 BMJ Open
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Pharmacist provided medicines reconciliation within 24 hours of admission 4 and on discharge: A randomised controlled pilot study 5 1 2 1 3 4 2 6 Brit Cadman , David Wright , Amanda Bale , Garry Barton , James Desborough , Eman A 5 3 1 3 3 4 7 Hammad , Richard Holland , Helen Howe , Ian Nunney , Lisa Irvine 8 9 1 Pharmacy Department, Cambridge University Hospitals, Cambridge, UK 10 2 School of Pharmacy, University of East Anglia, UK 11
12 3 Norwich Medical School, University of East Anglia, UK 13 14 4 Norwich Clinical Trials Unit, University of East Anglia, UK 15 For peer review only 16 5 School of Pharmacy, University of Jordan, Jordan 17 18 6 Patient and Public Involvement Representative 19 20 21 Correspondence to: Professor David Wright, School of Pharmacy (Room 0.17), University of 22 East Anglia, Norwich Research Park, Norwich, NR4 7TJ [email protected] Telephone: 01603 592042Fax: 01603 592003 23
24 Ms Brit Cadman 25 Consultant Pharmacist and Principal Investigator 26 Pharmacy Department 27 Box 104 28 Cambridge University Hospitals 29 Cambridge 30 CB2 0QQ 31 [email protected] 32 01223 217980 33 http://bmjopen.bmj.com/ 34 Ms Amanda Bale 35 Senior Research Assistant Pharmacy Department 36 Box 104 37 Cambridge University Hospitals 38 Cambridge 39 CB2 0QQ 40 [email protected] 41 01223 217980 42 on September 27, 2021 by guest. Protected copyright. 43 Dr Garry Barton 44 Reader in Health Economics 45 Norwich Medical School 46 Faculty of Medicine and Health Sciences 47 University of East Anglia Norwich 48 NR4 7TJ 49 [email protected] 50 01603 591936 51 52 Dr James Desborough 53 Senior Lecturer in Pharmacy Practice 54 School of Pharmacy (Room 0.14) 55 University of East Anglia 56 Norwich Research Park 57 Norwich 58 NR4 7TJ 59 60 1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 21
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 [email protected] 4 01603 593413 5 6 Dr Eman Hammad 7 Assistant Professor of Clinical Pharmacy and Health Economics Department of Biopharmaceutics and Clinical Pharmacy 8 Faculty of Pharmacy University of Jordan 9 Amman 11942 10 Jordan 11 [email protected] 12 +9626 5355000 (ext.23369) 13 14 Professor Richard Holland 15 ProfessorFor of Public Health peer Medicine review only 16 Norwich Medical School (Room 1.07A) 17 University of East Anglia 18 Norwich Research Park 19 Norwich 20 NR4 7TJ [email protected] 21 01603 593574 22 23 Professor Helen Howe 24 Professor of Pharmacy Practice 25 School of Pharmacy 26 University of East Anglia 27 Norwich 28 NR4 7TJ 29 [email protected] 30 31 Mr Ian Nunney 32 Lecturer Norwich Medical School (Room 0.30) 33 University of East Anglia http://bmjopen.bmj.com/ 34 Norwich Research Park 35 Norwich 36 NR4 7TJ 37 [email protected] 38 01603 597282 39 40 Miss Lisa Irvine 41 Senior Research Associate in Health Economics 42 Norwich Medical School on September 27, 2021 by guest. Protected copyright. 43 University of East Anglia 44 Norwich Research Park 45 Norwich NR4 7TJ 46 [email protected] 47 01603 591280 48 49 MeSH words: Medication reconciliation, cost effectiveness, pharmacy, randomised controlled 50 trial, pilot study. 51 52 Word count: 4165 excluding figures, tables and references 53 54 55 56 57 58 59 60 2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 21 BMJ Open
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Abstract 4 5 Background 6 The UK government currently recommends that all patients receive medicines reconciliation 7 (MR) from a member of the pharmacy team within 24 hours of admission and subsequent 8 discharge. The cost effectiveness of this intervention is unknown. A pilot study to inform the 9 design of a future randomised controlled trial to determine effectiveness and cost 10 effectiveness of a pharmacist delivered service was undertaken.
11 Method 12 Patients were recruited seven days a week from five adult medical wards in one hospital over 13 a 9 month period and randomised using an automated system to intervention (MR within 24 14 hours of admission and at discharge) or usual care which may include MR (control). 15 RecruitmentFor and retention peer rates were determined. review Length of stay(LOS), only quality of life (EQ 5D 16 3L), unintentional discrepancies(UDs) and emergency re admission(ER) within 3 months 17 were tested as outcome measures. The feasibility of identifying and measuring intervention 18 associated resources was determined. 19 20 Result 21 200 patients were randomised to either intervention or control. Groups were comparable at 22 baseline.95(99%) of patients in the intervention received MR within 24 hours, whilst 62(60.8%) of control patients received MRat some point during admission. The intervention 23 resolved 250 of the 255 UDs identifed at admission. Only 2 UDs were identifed in the 24 intervention group at discharge compared with 268 in the control. The median LOS was 94 25 hours in the intervention arm and 118 hours in the control, with ER rates of 17.9% and 26 26.7%,respectively. Assuming 5% loss to follow up 1120 patients (560 in each arm) are 27 required to detect a 6% reduction in 3 month ER rates. 28 29 Conclusions 30 The results suggest that changes in outcome measures resulting from MR within 24 hours 31 were in the appropriate direction and readmission within 3 months is the most appropriate 32 primary outcome measure. A future study to determine cost effectiveness of the intervention 33 is feasible and warranted. http://bmjopen.bmj.com/ 34 35
36 Strengths and limitations 37 • Pilot randomised controlled trial 38 • Intervention fidelity enhanced through competency assessment of medicine reconciliation 39 providers 40 • Robust process for identifying unintentional discrepancies at each stage developed to 41 prevent results contamination on September 27, 2021 by guest. Protected copyright. 42 • Pragmatic design and therefore elements of the intervention found within the control arm 43 • Limited response to request for patient data 3 months post discharge 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 21
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Background 4 Medicines Reconciliation (MR) is defined by the World Health Organisation as ‘the formal 5 process in which healthcare professionals partner with patients to ensure accurate and 6 complete medication information transfer at the interfaces of care’.(1) Researchers have 7 identified unintentional error rates within hospital prescriptions on admission of between 30 8 70% (2 5) and these can include omission of usual medicines, prescription of incorrect 9 dosages and addition of medicines which have not been previously prescribed. The majority of these errors are believed to result from deficiencies in the MR process with such errors 10 known to contribute to patient morbidity and mortality and increase the length of hospital stay 11 (LOS).(6 10) 12 13 Consequently, within the UK, patient safety guidance issued by the National Patient Safety 14 Agency (NPSA) and the National Institute of Health and Care Excellence (NICE) 15 recommendedFor that policies peer for MR should review be implemented in hospitals only for all adult patient 16 admissions, and that pharmacy should be involved in the MR process within 24 hours of 17 admission.(11) NICE guidance on pharmacy involvement was based primarily on one 18 randomised controlled trial (RCT) which demonstrated that the inclusion of the pharmacist in 19 MR reduced the error rate from 44% to 19%.(2) Data on cost effectiveness which usually 20 underpins recommendations made by NICE (12) is not available and therefore whether this 21 intervention represents an appropriate use of NHS resources is unknown. 22 Recent systematic reviews of hospital based medicines reconciliation practices consistently 23 demonstrate a reduction in medicine discrepancies, however this has not been shown to 24 result in reductions in post discharge healthcare utilisation.(13, 14) When considering only 25 pharmacist led medicines reconciliation interventions a 19% reduction in the rate of all cause 26 readmissions was seen but similarly pooled data on mortality and composite readmission and 27 emergency department visit did not find in favour of either pharmacist led intervention or 28 usual care.(15) 29 30 With the significant resources required for pharmacists to deliver medicines reconciliation 31 services it is not adequate to demonstrate reduction in errors (which may or may not result in 32 adverse drug events), it is also important in resource limited health systems to show that such 33 investment constitutes value for money. There is currently a recognised lack of evidence http://bmjopen.bmj.com/ 34 supporting the cost effectiveness of medicines reconciliation interventions provided by pharmacists.(16) A model built in 2008 to estimate the likely cost effectiveness of preventing 35 medication error at hospital admission using medicines reconciliation suggested that a 36 pharmacist based intervention is likely to be cost effective but was based solely on USA error 37 data and made assumptions concerning the severity of errors prevented.(17) 38 39 Despite national guidance few hospitals are providing MR as envisioned by NICE for all 40 patient admissions.(18) Thus, it would appear that further high quality evidence which 41 demonstrates cost effectiveness is required to ensure that resources are appropriately 42 allocated to this service in order to meet national recommendations. The most suitable on September 27, 2021 by guest. Protected copyright. 43 approach to determine the cost effectiveness of a complex intervention such as medicines 44 reconciliation is to perform a RCT which collates data on cost from the appropriate 45 perspective and outcomes which are most proximal to the intervention. Within the UK it is 46 additionally necessary to collect data on quality of life and mortality to enable the cost per 47 quality adjusted life year to be estimated. Recent guidance however suggests that before an RCT of such nature is undertaken feasibility testing and piloting are required.(19) 48
49 Aims 50 The aim of this study was therefore to pilot a randomised controlled trial to inform the design 51 of a future definitive study to determine the clinical and cost effectiveness of pharmacy 52 providing a full MR service on admission and at discharge. The objectives of the pilot study 53 were to 54 • identify the most suitable outcome measure for a future definitive trial with respect to 55 proximity and response to the intervention and quality of data obtained 56 • determine potential recruitment and retention rates 57 58 59 60 4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 21 BMJ Open
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 • develop and test the process for measuring resource utilisation associated with the 4 intervention and use of other NHS services 5 6 7 Method 8 9 Study design and location 10 The trial was randomised controlled pilot study undertaken at Cambridge University Hospitals 11 NHS Foundation Trust (CUHFT) on five adult medical wards from a range of medical 12 specialities where patients did not routinely receive MR from a pharmacist within 24 hours of 13 admission. One similar ward was identified as a “backup”, in the eventuality that one of the study wards was closed for any reason (e.g. norovirus outbreak) during the recruitment 14 period. 15 For peer review only 16 Intervention 17 A standard operating procedure (SOP) based on hospital guidelines, outlined in Figure 1, was 18 used to deliver MR by a trained medicines reconciliation pharmacist (MRP) within 24 hours of 19 admission (including weekends) and at the point of transfer of care out of hospital, or as soon 20 as possible following patient discharge from hospital to the next care provider. The five MRPs, 21 all clinical pharmacists employed within the hospital, covered for each other’s holidays, sick 22 leave and absences wherever possible. 23 24 MRPs recorded all unintentional discrepancies (UDs), defined as differences between patient 25 records with no identifiable rationale, they identified between the information they collated and 26 the inpatient medication chart on admission and again any differences between the inpatient chart and discharge letter. MRPs followed up on all identified UDs to ensure that they were 27 addressed prior to discharge. 28 29 Control 30 Patients in the control arm received usual care which may or may not consist of medicines 31 reconciliation and where it was provided it may not have occurred within 24 hours and could 32 either be delivered by a pharmacist or pharmacy technician. The MRPs within the 33 intervention arm did not deliver MR to control patients and the SOP used for study http://bmjopen.bmj.com/ 34 intervention purposes was not automatically followed within the control arm. For the purposes 35 of the study all MR details regarding interventions undertaken within the control arm were 36 recorded and costed. 37 38 Intervention fidelity To enhance intervention fidelity all MRPs were observed by the Principal Investigator (PI) on 39 at least 3 occasions to confirm adherence to the SOP. All MRPs had provided MR to more 40 than 30 patients in the year previous to delivering the intervention for the trial. 41 on September 27, 2021 by guest. Protected copyright. 42 Recruitment 43 A recruitment target of 200 patients was set for the nine month pilot phase. Study wards 44 were visited every morning by the research assistant (RA) during the study period to identify 45 potential participants. The nurse in charge of the ward confirmed that it was appropriate for 46 the patient to be approached to be consented to participate in the study. Patients were 47 recruited based on the following inclusion and exclusion criteria. 48 49 Inclusion criteria: 50 • Adult (≥18 years of age) 51 • Admitted with at least one prescribed medicine to one of the five medical wards 52 • Patient had not already received MR from the pharmacy team as part of routine 53 pharmaceutical input at the time of recruitment 54 • Identified from hospital computer system as having been admitted straight from the 55 Emergency Department (ED) to one of the five participating wards within the previous 24 56 hours 57 58 59 60 5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 21
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 With the intervention required to be delivered within 24 hours of admission, patients were 4 given a maximum of 2 hours to consider whether they wished to participate. Intervention 5 patients received MR in accordance with the SOP. Information obtained by the RA for the 6 purposes of the study was given to the MRP pharmacist prior to the visit to prevent 7 duplication of effort and to ensure the patient was not interviewed for the same information twice. 8
9 Randomisation was performed utilising the Norwich Clinical Trials Unit automated service with 10 patients stratified by ward. When wards were later closed for infection control reasons, 11 participants on the “backup” ward were randomised and stratified as if they had entered the 12 closed ward. 13 14 Patients randomised to the control group received usual care; this may have included 15 elements Forof MR by members peer of the pharmacy review team and in some onlycases may have occurred 16 within 24 hours of admission but post randomisation. Whilst this may have affected patients in 17 the control arm an intention to treat analysis was performed and consequently for the purpose 18 of the analysis patients they remained in their allocated arm. 19 20 Outcome measures Although undertaken as a pilot study with study aims to identify the most suitable outcome 21 measure, length of stay (LOS) was nominally selected as the primary outcome measure for 22 this pilot trial Secondary outcome measures were unplanned (emergency) re admission at 3 23 months, quality of life (EQ 5D 3L) and unintentional discrepancies. 24 25 Data collection 26 To enable comparison of intervention and control groups, age, gender, primary reason for 27 admission, all co morbidities and the admission ward was recorded. Additionally consented 28 patients or their third party consultees completed a quality of life score (EQ 5D 3L) (20) on 29 admission, including the related visual analogue scale (VAS). 30 31 LOS, reported in hours, was calculated as the difference in time from arrival at the hospital to 32 the time of discharge as recorded in the hospital information system (HISS). Unplanned re admissions to the intervention hospital within the 3 months post discharge were also obtained 33 from HISS. EQ 5D 3L responses were also obtained via postal survey 3 months post http://bmjopen.bmj.com/ 34 discharge, allowing the calculation of Quality Adjusted Life Years (QALYs)(21) in the 35 subsequent economic evaluation. 36 37 Unintentional discrepancy identification 38 To enable the identification of UDs the following information was photocopied by the research 39 associate (RA) for all consented patients (both intervention and control) and stored securely: 40 • All versions of the inpatient medication chart(s) and discharge letters, 41 • Medical notes during admission on September 27, 2021 by guest. Protected copyright. 42 • GP medication list on admission 43 • GP medication list at 3 month post discharge (when received from GP surgery) 44 • Any additional medicines related information brought in by the patient on admission e.g. 45 copies of labels from patient medicines, handwritten or typed medicine lists. 46 47 The RA was a trained nurse and consequently was experienced in medical data collation and 48 extraction. 49 50 3 months post discharge the stored information was used to develop an ‘accurate medication 51 list’ for the control arm patients by the research team on admission and at discharge. These 52 were then compared with the inpatient chart on admission and discharge letter to identify any 53 discrepancies. Medical notes were subsequently reviewed, unblinded to group allocation, to 54 enable differentiation between those which were unintentional i.e. discrepancies which could 55 not be explained from the information available and those which were intentional.
56 The GP medication list obtained 3 months post discharge (where available) was used to 57 enable the identification of discrepancies which still remained at 3 months. Without access to 58 59 60 6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 21 BMJ Open
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 GP medical notes it was not possible to establish whether identified discrepancies were 4 unintentional. 5 6 All potential discrepancies which were identified at the 3 month point as having translated into 7 the patient notes were reported to the patient’s GP.
8 Resource use 9 The time taken by the pharmacy team to deliver the MR was monitored in both the 10 intervention and control group (if applicable). Additionally, the following items were requested 11 in both groups: 12 • Time in hospital (Length of stay) 13 • Medication (in patient medication and GP medication list at 3 months) 14 • Re hospitalisations 15 • OtherFor health carepeer contacts review only 16 17 With the exception of the final item, which included all health professional contacts and was 18 requested from the participant at 3 months post discharge, these were extracted from medical 19 records. 20 21 Sample size calculation 22 As a pilot study a formal power calculation was not performed. The consequences for the 23 precision of the primary outcome variable (LOS) of the choice of sample size were however estimated. Summary statistics on LOS taken from a study undertaken at St James Hospital, 24 Dublin (22) gave quartiles for two groups as (3, 7, 5) and (2, 5, 12) days. To derive an 25 estimate of variability from this, an underlying log normal distribution was assumed, which is 26 consistent with the position of the medians, and its standard deviation estimated using the 27 geometric mean of the ratios between the upper and lower quartiles. This produced an 28 estimated SD of 1.26 which implied that a comparison between 2 groups of 100 patients, 29 would provide an expected half width for the 95% confidence interval of the difference 30 between group means (on the natural log scale) of 0.35. Translated back to the confidence 31 interval for the ratio of average length of stay between the intervention and control groups 32 would extend by a factor of about 1.4 either side of the point estimate. 33 http://bmjopen.bmj.com/ 34 Data analysis 35 As a pilot study descriptive statistics were utilised to determine the suitability of the different outcome measures and report the variation in the differences in order to determine a sample 36 size for a future randomised controlled trial. Similarly, completion rates are reported for each 37 source of resource use data and the EQ 5D 3L. 38 39 The primary outcome variable, length of stay, was reported using median, arithmetic mean 40 and geometric mean. The rate of Trust re admissions, Trust emergency re admissions and 41 mortality are also reported for both arms, along with the mean change in the VAS from the 42 EQ 5D 3L. on September 27, 2021 by guest. Protected copyright. 43 44 Results 45 Nine hundred and nineteen patients were assessed for eligibility of which 224 did not meet 46 inclusion criteria. 200 out of 310 patients who were subsequently approached by the RA 47 consented to take part in the study (Figure 2). Of those patients identified as potentially 48 eligible but not approached this was primarily because either the nurse in charge of the ward 49 advised that the patient was not suitable to be approached or that the end of the 24 hour window for intervention was due to expire. 50
51 Recruitment took place between July 2012 and April 2013 (9 months and 2 weeks), resulting 52 in a recruitment rate of 5.2 patients per seven days. 53 54 There was one post randomisation exclusion in each arm (they were subsequently found not 55 to meet the inclusion criteria), and were accordingly excluded from all analyses. Table 1 56 thereby summarises the characteristics of the remaining 198 participants (96 intervention 102 57 control). The groups were broadly comparable. 58 59 60 7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 21
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 One patient in the intervention arm did not receive the MR as he was not on the ward when 4 the pharmacist visited. Data were erroneously not collected for this patient post 5 randomisation. The time taken by the pharmacist to deliver the intervention was recorded for 6 the remaining 95 intervention participants, where the mean total time was 48.6 minutes 7 (range 2 to 195 minutes). In the control group 62(60.8%) participant received some form of medicines reconciliation, 31(30.4%) from a pharmacist (mean reported time 15 minutes) and 8 31(30.4%) from a pharmacy technician (mean reported time 12 minutes) 25(24.5%) control 9 patients received MR within the 24 hour window. 10 11 During the initial hospital admission three participants died, with a further 11 deaths during the 12 three months follow up period (see Figure 2). In total 6 intervention and 8 control patients 13 died during the trial period (p=0.78, Fisher’s Exact). Additionally, in the control arm, one 14 participant was lost to follow up (address not known) and one participant withdrew. After 15 taking accountFor of the 1 peerpatient who did notreview receive the MR and a onlyfurther loss to follow up in 16 the intervention arm, this left a total of 88 available cases in the intervention arm and 92 17 available cases in the control arm at final assessment point. In terms of outcomes, there was 18 complete data available on LOS and re admission data. 19 20 Table 2 provides a summary of the unintentional discrepancies which were identified at each stage. 16 (16%) intervention patients had no discrepancies at admission and discharge, 21 compared to 12 (12%) control patients. Overall, 2 UDs were known to remain at discharge in 22 the intervention arm compared with 268 in the control group. Neither of the UDs in the 23 intervention arm identified at discharge were present at 3 months. 24 25 One hundred and fifty four unintentional discrepancies identified in the control arm were 26 potentially related to medicines to be prescribed post discharge i.e. the remainder related to 27 medicines prescribed during admission only. Due to the limited number of GP records 28 provided at 3 months data were only available for 82 (53.2%) unintentional discrepancies in 29 the control arm at 3 months and 37 out of the original 154 (24%) were found from the medical 30 records provided to have been propogated into the patient notes. 31 32 Table 3 provides a comparison of patient outcomes for the intervention and control group. The results suggest that LOS, mortality and rehospitalisation rates tend to be lower in the 33 intervention arm although statistical significance was not demonstrated as 95% Confidence http://bmjopen.bmj.com/ 34 Intervals overlapped. Based on those who responded, both groups had a higher mean quality 35 of life (based on the VAS) at the 3 month follow up with improvement higher in the control 36 arm. This difference between groups was not significant. 37 38 With regard to the response rates for the resource use data there were complete data 39 available (for those upon whom it was requested) for length of stay, medication data as part of 40 the original admission, re admissions (to the same hospital) and the intervention pharmacist 41 times. Of the 62 controls for whom pharmacist/pharmacy technician review occured the times 42 were missing for N=27. At 3 month follow up, medication data were retrieved from GPs for 86 on September 27, 2021 by guest. Protected copyright. 43 participants (94.5% of those from whom it was requested) in each arm and 133 participants 44 completed and returned both the EQ 5D 3L and the health resource use questionnaires (66 45 Intervention, 67 Control) (73.5% of those requested from all living participants).
46 Based on the pilot data we calculate that for a full trial 1120 patients would need to be 47 recruited to detect a 6% reduction (conservative assumption, one standard error below 9% 48 reduction seen) in 3 month unplanned re admission rates from a starting point of 26% with 49 90% power, using a 5% significance level and assuming 5% loss to follow up. 50 51 52 53 54 55 56 57 58 59 60 8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 21 BMJ Open
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Discussion 4 The results from this study which was performed to inform the design of a future randomised 5 controlled trial suggest that even though medicines reconciliation activities are taking place 6 such a trial is feasible with reasonable recruitment and retention rates and that both cost and 7 outcome data can be effectively obtained. 8 We consider that emergency rehospitalisation within 3 months would be the most appropriate 9 primary outcome measure for such a trial, as unlike the other outcome measures tested, it 10 reflects all of the MR activity which occurs in both secondary and primary care. Furthermore, 11 data collection was complete and this is a patient orientated outcome, unlike medication 12 errors which whilst representing a patient safety issue are a measure of the prescribing 13 process. 14 15 Due to theFor need to recruit peer patients before review they received MR as usualonly care this reduced the 16 generalisability of the sample with all of those who had already received MR being 17 automatically excluded. This was further compounded by recruitment activities which were 18 focussed towards mornings. 19 20 With a hospital requirement that all patients receive MR within 24 hours of admission it is 21 unsurprising that two thirds of the control arm received some form of MR during their hospitalisation, a quarter of which received this within 24 hours. There is no specific 22 requirement for MR on discharge however and this may explain some of the differences seen 23 between the two groups. The majority of discrepancies identified in the control arm by the 24 researchers were found not to have been resolved and therefore reasons for the relative 25 ineffectiveness of the control arm MR requires elucidation through a detailed process 26 evaluation. 27 The average time spent on the intervention found within this study was very similar to that 28 reported in a MR time and motion study(23) but three times greater than that in the control 29 arm. The study SOP required pharmacists to undertake initial MR, follow up on all 30 interventions to ensure that discrepancies had been addressed, assess all discharge letters 31 for accuracy and correct them. Patients in the control arm frequently did not always receive 32 all four elements and this probably explains most of the difference in time provided. It has 33 been suggested that organisations are probably unlikely to repeat the benefits from MR http://bmjopen.bmj.com/ 34 services reported in the literature if there are deficiencies in intervention intensity and breadth.(24) The results from the control arm of this study support this assertion and suggest 35 that if MR of a similar nature to that seen in the intervention arm was to be shown to be cost 36 effective then this would require significantly more pharmacist time. 37 38 The proportion of patients screened for eligibility and eventually recruited to the study was 39 20% and this could have been improved by increasing the number of weekends covered 40 (from 80% to 100%). Although some patients were not suitable for inclusion in the study as 41 they had already been reviewed by a member of the pharmacy team prior to being 42 approached by the RA, there was a sufficient number of patients available for recruitment. on September 27, 2021 by guest. Protected copyright. 43 44 When identifying unintentional discrepancies, we have assumed that the MRP generated list 45 in the intervention arm and the RA generated list in the control arm were accurate. Both are 46 unrealistic assumptions and ideally within a definitive study all data should have been reviewed independent of the service and blinded to group allocation. The unblinded 47 identification of medication discrepancies and inability to confirm intentional or unintentional 48 nature of errors in many instances also means that the data on unintentional discrepancies 49 must be treated with further caution. 50 51 The process of collating all medicines related data at different time points and sealing it until 3 52 months post discharge provides an opportunity to identify discrepancies without adversely 53 affecting the intervention. Whilst the identification of unintentional discrepancies should be 54 undertaken blind of allocation the resources required for this may not be warranted when 55 considering that the intentional or unintentional nature of the discrepancy cannot always be 56 accurately determined. 57 58 59 60 9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 21
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Considering evidence published post study completion, limiting the study population to those 4 over the age of 70 and including a post discharge telephone call as part of the intervention 5 may have further enhanced the intervention.(13) 6 7 In line with previous research the intervention prevented a large number of unintentional medicines related discrepancies both during admission and post discharge.(3 6) The data 8 obtained suggests that just less than a quarter of unintentional discrepancies identified at 9 discharge were found to actually translate into primary care records at three months. Whilst 10 reasons for non translation were not elucidated the research suggests that the use of number 11 of unintentional discrepancies at discharge as an outcome measure may over emphasise the 12 problem. 13 14 Not all primary care medication lists were made available to researchers and approaches to 15 addressingFor this will require peer consideration review for a future definitive study. only Utilising data on 16 ‘unplanned readmission at 3 months’ would seem to be the most appropriate primary 17 outcome measure for a future definitive study as it is a patient orientated outcome which is 18 most likely to reflect the effect of errors which occur at all stages of the process. Similar to 19 hospital readmission, LOS is a cost which would be captured in any cost effectiveness 20 analysis. LOS was however found to be largely skewed by a small number of individuals who were admitted for extended periods. Furthermore, it is not affected by errors which translate 21 into primary care i.e. does not reflect the full impact of the service. Within the UK hospitals are 22 penalised for unplanned readmissions within one month of discharge and therefore collection 23 of this data may also be warranted for a UK based definitive study. 24 25 Unplanned readmission has been used as a primary outcome measure within other similar 26 MR trials (25 29). Whilst differences in unplanned readmission at three months have been 27 demonstrated,(28, 29) this has frequently not been the case at either one or six months.(25 28 27) Within the first month patients may still be using the medication they were discharged with 29 and consequently this may lessen the impact of errors on the discharge letter resulting from 30 incorrect translation into primary care. Kwan et al. after systematically reviewing the literature 31 suggest that unplanned readmission data should probably be collected for more than 30 days 32 post discharge.(14) Unplanned readmission at six months will be affected by more factors 33 unrelated to the index admission than at 3 months and therefore it may be more difficult to http://bmjopen.bmj.com/ 34 identify the impact of MR intervention using this outcome measure.
35 Quality of data collection with respect to LOS, mortality and readmission rates was high as 36 this information was available from hospital records. Resource utilisation data and quality of 37 life scores were however only available for two thirds of participants at follow up. 38 Consequently researchers powering a definitive study on readmission rate will need to 39 consider the possibility of there being more missing data for the cost effectiveness analysis 40 and imputation of missing data (30) may be necessary to address this. 41 42 This study was a pilot study and was not designed to obtain a definitive answer to whether on September 27, 2021 by guest. Protected copyright. 43 MR provided to all patients within 24 hours of admission was cost effective. However, it was 44 conducted as an RCT, conforming to expected standards and consequently it can be 45 assumed that it would be possible to perform a full scale RCT in the future. An internal pilot 46 would be warranted in such a trial as the trial should be multi centre in nature and consequently local variations in recruitment rates and service delivery would require 47 identification and appropriate local adaptation. Additionally, costs related to service delivery 48 may differ between settings and consequently time for service delivery would require 49 estimation. 50 51 The pilot study has shown that it is feasible to perform an RCT to determine the effectiveness 52 and cost effectiveness of a pharmacist MR within 24 hours of admission and again at 53 discharge. This study has demonstrated that this form of intervention does appear to reduce 54 medication errors at discharge, and may reduce LOS and hospital readmissions. We 55 consider that unplanned readmission at 3 months is the most suitable primary outcome 56 measure but LOS, errors, mortality and quality of life should be captured. 57 58 59 60 10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 21 BMJ Open
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Additionally a thorough process evaluation is warranted in order provide a more complete 4 understanding of the intervention. Pharmacists delivering the intervention arm within a 5 definitive study should follow an SOP and be trained to undertake the role to ensure 6 standardisation in delivery. Whilst the results of this pilot emulate other studies where 7 prescribing errors were reduced in the intervention arm there is an additional cost associated with providing the intervention and therefore high quality evidence from a multi centre RCT is 8 now needed to determine both its effectiveness and cost effectiveness. 9 10 Acknowledgements: 11 The authors would like to thank Dr Tracey Sach for her contribution in the design of the health 12 economics element of the project and in the development of the original grant proposal, 13 Antony Warford for his contribution to the management, delivery and dissemination of the 14 project as the patient and public involvement representative and Kellie Hempstead and the 15 pharmacistsFor volunteering peer to provide the interventionreview at weekends only in particular Lynn Martin for 16 their efforts in completing this study. The input by the Norwich Clinical Trials Unit throughout 17 the study is also acknowledged. 18 19 Contributors: 20 BC, EH, DW, JD, RH participated in the conception and design of this study. AB, BC, DW,GB, 21 IN, LI participated in the analysis of the data and interpretation of the results AB, BC, DW 22 drafted the article and all other authors performed critical review of the article. All authors approved the submission of this article 23
24 25 Funding: This article presents independent research commissioned by the National Institute for Health 26 Research (NIHR) under Research for Patient Benefit. The views expressed in this publication 27 are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department 28 of Health. 29 30 Competing interests: 31 None 32 33
Ethics approval: http://bmjopen.bmj.com/ 34 Ethical approval was received on 14th June 2012 from NRES Committee East of England 35 Essex Research Ethics Committee. Reference number: 12/44/0143 36 37 Provenance and peer review 38 The trial was registered on 5th August 2012: ISRCTN23949491 39 The trial received competitive funding from NIHR Research for Patient Funding stream and 40 was peer reviewed as part of this process. 41 42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 21
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Figure 1 Outline of service standard operating procedure 4 5 Admission 6 • Patient interviewed to confirm allergy status and medicines being taken 7 • GP letter/fax reviewed to confirm allergy status and medicines prescribed 8 • Medical notes reviewed to confirm any intentional medication changes due to clinical 9 status of patient 10 • Other sources of information may have been used depending on availability and 11 relevance to current admission by virtue of the date, and included o previous electronic discharge letter 12 o 13 clinic letter o patient’s own drugs (PODs) 14 o patients relative or carer 15 oFor nursing home peer record review only 16 o medication administration record (MAR) 17 o modified dosage system (MDS) 18 • An accurate medication list was documented in the medical notes by the intervention 19 pharmacist 20 21 Discharge 22 • Discharge letter compared with the final drug chart to ensure correct 23 • Medical notes reviewed 24 • Changes to medicines field on the electronic system was completed by the intervention 25 pharmacist to communicate intentional medication changes to the GP 26 • All medicines were added to the discharge letter to ensure a complete record at the point 27 of transfer of care 28 29 30 31 32 33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40 41 42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 21 BMJ Open
BMJ Open Version 2 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Figure 2 Consort diagram 4 5 6 Enrolment 7 Assessed for eligibility (n=919) 8 Excluded (n=719) 9 • Not meeting inclusion criteria (n=224) • Declined to participate (n=110) 10 • Other reasons (n=385) 11 Allocation Randomised (n=200) 12 13 Excluded post randomisation (n=1 #95) Excluded post randomisation (n=1 #154) Not meeting inclusion Not meeting inclusion 14 15 For peer review only
16 Intervention arm (n=96) Control arm (n=102) 17 • Received intervention < 24 hr (n=95) • < 24 hour by pharmacist (n=13) 18 • Did not receive intervention as > 24 hr • < 24 hour by technician (n=22) (n=1) #75 • > 24 hour by pharmacist (n=18) 19 • > 24 hour by technician (n=9) 20 • No MR performed (n=40)
21 Discharge follow up 22 23 24 Lost to follow up (n=0) Lost to follow up (n=0) Death (n=2) #108 #136 Death (n=1) #20 25
26 3 month follow up 27
28 Lost to follow up (n=1) #70 Lost to follow up (n=1) #83 29 Withdrawn (n=0) Withdrawn (n=1) #173 30 Death (n=4) #51, 68, 82, 97 Death (n=7) #16, 55, 76, 80, 86, 99, 198
31 Data availability 32 33 Complete self reported cost data (n=66) Complete self reported cost data (n=66) http://bmjopen.bmj.com/ 34 EQ5D data available at 3 months (incl. EQ5D data available at 3 months (incl. deaths) (n=72) deaths) (n=75) 35 Medication data received from GPs at 3 Medication data received from GPs at 3 36 months (n=78) months (n=85) Medication data not provided at 3 months Medication data not provided at 3 months 37 (n=12 plus n=6 RIPs) (n=9 plus n=8 RIPs) 38 39 40 41 42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 21
Version 27 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Table 1 Comparison of demographics at baseline: 4 5 Demographic Measure Intervention Control (n=102) 6 (n=96) 7 Female No. (%) 45 (46.9) 60 (58.8) 8 Age (years) Mean (SD) 67.6 (19.0) 65.4 (20.2) 9 Regular medicines Mean (SD) 5.84 (4.07) 6.67 (4.64) 10 As required medicines Mean (SD) 0.85 (2.08) 0.95 (2.53) 11 12 Ward 13 1 No. (%) 26 (26.8) 27 (26.2) 14 2 No. (%) 30 (30.9) 30 (29.1) 15 For peer3 reviewNo. (%) 10 (10.3) only 9 (8.7) 16 4 No. (%) 14 (14.4) 16 (15.5) 17 5 No. (%) 16 (16.4) 14 (13.5) 18 6 No. (%) 1 (1.03) 7 (6.8) 19 20 EQ5D Quality of life (Visual Mean (SD) 55.9 (23.2) 54.7 (23.5) 21 Analogue Scale) 22 Reason for admission 23 Lower respiratory tract infection No. (%) 9 (9.2) 11 (10.7) 24 Troponin –ve chest pain No. (%) 7 (7.2) 4 (3.9) 25 Heart failure No. (%) 2 (2.1) 5 (4.9 ) 26 Exacerbation Chronic No. (%) 4 (4.1) 3 (2.9) 27 Obstructive Pulmonary Disease 28 Other No. (%) 73 (75) 79 (76.7) 29 30 31 32 33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40 41 42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 21 BMJ Open
Version 27 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 Table 2 Comparison of unintentional discrepancies (UDs) 4 5 Unintentional Intervention Control 6 discrepancies 7 No. UDs No. No. per No. UDs No. No. per patients patient patients patient 8 9 Admission 255 95 2.80 309 102 3.0 10 Resolved during 250 95 2.74 Unknown 11 hospital stay 12 13 Remaining at 2 91 0.02 268 99 2.71 14 discharge 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40 41 42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from Page 16 of 21 1.51 (4.36) 1.51
20.84 (38.75)20.84
Mean differenceMean (SE of thedifference)
7.15 (26.2)7.15 8 (7.8)8 27 (26.7)27 37 (36.6)37 203.9 (246.8)203.9 117 (13 1546) 117 137.3) 109.3 (87.0 to 109.3 Control
68 95 101 101 102 102 102 N
5.64 (23.6) 5.64
6 (6.3) 6 17 (17.9)17
30 (31.6)30 224.8 (293.1)224.8 94.0 (12 – 94.0 1077) 99.6 (76.59 to 129.63)99.6 Intervention
16 63 95 95 95 95 95 95 N BMJ Open http://bmjopen.bmj.com/
Mean (SD)Mean No. (%) No.
No. (%) No.
No. (%) No. Median (range)Median Geometric (95%MeanGeometric CI) Measure Arithmetic mean (SD)mean Arithmetic on September 27, 2021 by guest. Protected copyright.
For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only
Visual Visual Scale Analogue
readmissions(emergency) readmissions
Quality of life Quality of life from changebaseline (high better) score Mortality Hospital Hospital Length of (hours) stay Outcome Version 27 27 Version Table 3 Table 3 outcome Comparison of measures 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 17 of 21 BMJ Open
Version 27 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 References 4 1. World Health Organisation. High 5s Fact Sheet: The High 5s Assuring Medication 5 Accuracy at Transitions of Care: Medicines Reconciliation Standard Operating Procedure 6 [Available from: http://www.who.int/patientsafety/implementation/solutions/high5s/h5s fact 7 sheet.pdf?ua=1. 8 2. The University of Sheffield SoHaRRS. A systematic review of the effectiveness and costeffectiveness of interventions aimed at preventing medication error (medicines 9 reconciliation) at hospital admission. NICE. 2009. 10 3. Tim Dornan DA, Heather Heathfield, Penny Lewis, Jon Miles, David Taylor, Mary 11 Tully, Val Wass. Final report. An in depth investigation into causes of prescribing errrors by 12 foundation trainees in relation to their medical education. EQUIP Study. General Medical 13 Council: 2009. 14 4. Gray E, Ullah S, Lwin A, Javed S, Farrow C, Southgate J, et al. Prescription Errors in 15 Patients onFor an Acute Medicalpeer Assessment review Unit. Royal College ofonly Physicians of Edinburgh. 16 2007:37:305 8. 17 5. McLeod SE, Lum E, Mitchell C. Value of Medication Reconciliation in Reducing 18 Medication Errors on Admission to Hospital. Journal of Pharmacy Practice and Research. 19 2008;38(3):196 9. 20 6. Pippins JR, Gandhi TK, Hamann C, Ndumele CD, Labonville SA, Diedrichsen EK, et 21 al. Classifying and predicting errors of inpatient medication reconciliation. J Gen Intern Med. 2008;23(9):1414 22. 22 7. The Audit Comission. A spoonful of sugar Medicines Management in NHS Hospitals. 23 London. 2001. 24 8. Dean B, Schachter M, Vincent C, Barber N. Prescribing errors in hospital inpatients: 25 their incidence and clinical significance. Qual Saf Health Care. 2002;11(4):340 4. 26 9. Department of Health. Management of Medicines A resource to support 27 implementation of the wider aspects of medicines management for the National Service 28 Frameworks for Diabetes, Renal Services and Long term Conditions. 2004. 29 10. Karnon J, McIntosh A, Dean J, Bath P, Hutchinson A, Oakley J, et al. Modelling the 30 expected net benefits of interventions to reduce the burden of medication errors. J Health 31 Serv Res Policy. 2008;13(2):85 91. 32 11. National Institute for Health and Clinical Excellence. Technical patient safety solutions 33 for medicines reconciliation on admission of adults to hospital. NICE; 2007. http://bmjopen.bmj.com/ 34 12. NICE. Guide to the methods of technology appraisal 2013. National Institute of Health and Clinical Excellence (NICE) publications, 2013. 35 13. Mueller SK, Sponsler KC, Kripalani S, Schnipper JL. Hospital based medication 36 reconciliation practices: a systematic review. Arch Intern Med. 2012;172(14):1057 69. 37 14. Kwan JL, Lo L, Sampson M, Shojania KG. Medication reconciliation during transitions 38 of care as a patient safety strategy: a systematic review. Ann Intern Med. 2013;158(5 Pt 39 2):397 403. 40 15. Mekonnen AB, McLachlan AJ, Brien JA. Effectiveness of pharmacist led medication 41 reconciliation programmes on clinical outcomes at hospital transitions: a systematic review 42 and meta analysis. BMJ open. 2016;6(2):e010003. on September 27, 2021 by guest. Protected copyright. 43 16. Hammad EA, Bale A, Wright DJ, Bhattacharya D. Pharmacy led medicine 44 reconciliation at hospital: a systematic review of effects and costs. Research in Social and 45 Administrative Pharmacy. 46 17. Karnon J, Campbell F, Czoski Murray C. Model based cost effectiveness analysis of interventions aimed at preventing medication error at hospital admission (medicines 47 reconciliation). Journal of evaluation in clinical practice. 2009;15(2):299 306. 48 18. Farmer D. 2009 Eastern England Collaborative Audit on Medicines Reconciliation 49 2009. 50 19. Medical Research Council. Developing and evaluating complex interventions 2015 51 [Available from: www.mrc.ac.uk/complexinterventionsguidance. . 52 20. Brooks R. EuroQol: the current state of play. Health Policy. 1996;37(1):53 72. 53 21. Gray AM, Clarke PM, Wolstenholme JL, Wordsworth S. Applied methods of cost 54 effectiveness analysis in healthcare: Oxford University Press; 2011. 55 22. Moloney ED, Bennett K, Silke B. Effect of an acute medical admission unit on key 56 quality indicators assessed by funnel plots. Postgraduate Medical Journal. 2007;83(984):659 57 63. 58 59 60 17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 21
Version 27 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from 1 2 3 23. Meguerditchian AN, Krotneva S, Reidel K, Huang A, Tamblyn R. Medication 4 reconciliation at admission and discharge: a time and motion study. BMC Health Services 5 Research. 2013;13(1):1 11. 6 24. Pevnick JM, Shane R, Schnipper JL. The problem with medication reconciliation. 7 BMJ quality & safety. 2016. 25. Gillespie U, Alassaad A, Henrohn D, Garmo H, Hammarlund Udenaes M, Toss H, et 8 al. A comprehensive pharmacist intervention to reduce morbidity in patients 80 years or older: 9 a randomized controlled trial. Arch Intern Med. 2009;169(9):894 900. 10 26. Spinewine A, Swine C, Dhillon S, Lambert P, Nachega JB, Wilmotte L, et al. Effect of 11 a collaborative approach on the quality of prescribing for geriatric inpatients: a randomized, 12 controlled trial. J Am Geriatr Soc. 2007;55(5):658 65. 13 27. Stowasser DS, M and Collins, DM. A randomised controlled trial of medication liaison 14 services patient outcomes Journal of Pharmacy Practice and Research. 2002;32(2):133 40. 15 28. MakowskyFor MJ, peerKoshman SL, Midodzi review WK, Tsuyuki RT. Capturingonly outcomes of clinical 16 activities performed by a rounding pharmacist practicing in a team environment: the 17 COLLABORATE study [NCT00351676]. Med Care. 2009;47(6):642 50. 18 29. Boso Ribelles V, Montero Hernandez M, Font Noguera I. Evaluation of a plan for 19 cardiology medication reconciliation on admission, and patient information at discharge in a 20 teaching hospital. European Journal of Hospital Pharmacy Practice. 2011;17(1):26 30. 30. Rubin DB. Multiple Imputation for Non response in Surveys: Wiley and Sons; 1987. 21
22 23 24 25 26 27 28 29 30 31 32 33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40 41 42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from Page 1 1 Page * * Reported Reported on page No page on N/A N/A 5 5
7 7 11 11 5 5 4 4 4 5 5 5 5 5 5 6 6 5 5 1 1 5 5 N/A N/A 5 5 3 (for specific guidance see CONSORT for abstracts) for see CONSORT guidance specific (for
BMJ Open http://bmjopen.bmj.com/ on September 27, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only actually administered administered actually Checklist item Checklist interventions interventions describing any steps taken to conceal the sequence until interventions were assigned assigned were interventions until sequence the conceal to stepstaken any describing were assessed assessed were 9 9 containers), numbered sequentially as (such sequence randomallocation the implement to used Mechanism 1a 1a titlethe in trial randomised as a Identification 6a 6a they when and how including measures, outcome secondary and primary pre-specified defined Completely 7b 7b guidelines stopping and analyses interimof any explanation applicable, When 6b 6b reasons with commenced, trial the after outcomes trial to changes Any 4b 4b collected were data the where locations and Settings 3b 3b reasons criteria), with aseligibility (such commencement trial after methods to changes Important 2a 2a of rationale explanation and background Scientific 2b 2b hypotheses or objectives Specific 8a 8a sequence allocation random the generate to used Method 8b size) block and blocking as (such restriction ofany details of randomisation; Type 1b 1b conclusions and results, methods, design, of trial summary Structured No No 11a 11a those providers, care participants, example, (for interventions to after assignment blinded was who Ifdone, N/A Item Item CONSORT CONSORT 2010 checklist of to include wheninformation randomised reporting a trial generation generation concealment concealment mechanism Implementation Implementation 10 to participants assigned and who participants, enrolled who sequence, random allocation the Whogenerated Sequence Sequence Allocation Allocation Blinding Blinding
Randomisation: Randomisation: Sample size size Sample 7a determined was size sample How Outcomes Outcomes Interventions Interventions 5 were they andwhen how including allowreplication, to details sufficient with group foreach Theinterventions Participants Participants 4a for criteria participants Eligibility Trial design design Trial 3a ratio allocation including factorial) parallel, as (such design oftrial Description Methods Methods Background and Background objectives Introduction Introduction
Title and abstract abstract and Title Section/Topic Section/Topic CONSORT 2010 checklist checklist 2010 CONSORT Page 19 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from Page 2 2 Page Page 20 of 21 N/A N/A
15 15 7 7 7 7 N/A N/A online online N/A N/A N/A N/A Info submitted submitted Info online
9 9 15 15 8 8 Info submitted submitted Info 15 15 8 8 13 13 N/A N/A
(for specific guidance see CONSORT for harms) for CONSORT see guidance specific (for BMJ Open http://bmjopen.bmj.com/ on September 27, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only by original assigned groups assignedgroups original by pre-specified from exploratory fromexploratory pre-specified assessing outcomes) and how how and outcomes) assessing precision (such as 95% confidence interval) interval) confidence 95% as (such precision were analysed for the primary outcome primary for analysed the were 25 25 funders of role ofdrugs), supply as (such support other and funding of Sources 23 23 registry trial ofname and number Registration 24 24 ifavailable accessed, be can protocol trial full the Where 19 19 group each in effects unintended or harms important All 17b 17b sizesrecommended is effect relative and absolute of both presentation outcomes, binary For 14b 14b stopped was or ended trial the Why 12b 12b analyses adjusted and analyses subgroup as such analyses, foradditional Methods 11b 11b ofinterventions similarity ofthe description Ifrelevant, 17a 17a andits size effect estimated the and group, eachfor results outcome, andsecondary primary each For 13a 13a and treatment, intended received assigned, randomly were who participants of numbers the group, each For 13b 13b reasons with together randomisation, after exclusions and losses group, each For
Funding Registration Other information information Other Interpretation Interpretation 22 evidence relevant other considering and harms, and benefits balancing results, with consistent Interpretation 10 Limitations Limitations Generalisability 21 20 findings trial of the applicability) validity, (external Generalisability ofanalyses multiplicity if relevant, and, imprecision, bias, potential sources of addressing limitations, Trial 9 Discussion Discussion Harms Harms Outcomes and and Outcomes estimation Numbers analysed analysed Numbers 16 was analysis the whether and analysis each in included (denominator) participants of number group, each For Baseline data data Baseline 15 group each for characteristics clinical and demographic baseline showing Atable Recruitment Recruitment 14a follow-up and recruitment periods of the defining Dates Participant flow (a (a flow Participant strongly is diagram recommended) Results Results Statistical methods methods Statistical 12a outcomes secondary and primary for compare groups to used methods Statistical CONSORT 2010 checklist checklist 2010 CONSORT Protocol Ancillary analyses analyses Ancillary 18 distinguishing analyses, adjusted and analyses subgroup including performed, analyses other any of Results 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2016-013647 on 16 March 2017. Downloaded from Page 3 3 Page . www.consort-statement.org BMJ Open http://bmjopen.bmj.com/ on September 27, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only recommend reading recommend extensionsCONSORT for cluster trials,randomised non-inferiority trials, and equivalence non-pharmacological treatments, herbal and interventions, pragmatic trials. extensions Additional are forthcoming: for thosefor date references and up to this relevant checklist, to see *We recommendstrongly *We this reading conjunction statement in the with 2010 CONSORT Explanationand forElaboration important clarifications items.allthe If on relevant, also we CONSORT 2010 checklist checklist 2010 CONSORT Page 21 of 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60