Pharmacist-independent prescriber deprescribing in UK care homes: Contextual factors associated with increased activity

Aims: The Care Home Independent Pharmacist Prescriber Study (CHIPPS) process evaluation hypothesized that contextual factors influenced the likelihood of deprescribing by pharmacist-independent prescribers. The aim of this paper is to test this hypothesis


Funding information
This study is funded by the National Institute for Health Research (NIHR) for Health Research (NIHR) Translating Research into Policy funding scheme (Award ID: NIHR202053). The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.
Aims: The Care Home Independent Pharmacist Prescriber Study (CHIPPS) process evaluation hypothesized that contextual factors influenced the likelihood of deprescribing by pharmacist-independent prescribers. The aim of this paper is to test this hypothesis.
Methods: From CHIPPS study data, medications deprescribed totalled 284 for 370 residents in UK care homes. Regression analysis was used to describe the relationship between the number of medicines stopped and contextual factors (number of residents cared for, pharmacist employment within associated medical practice, previous care home experience, hours active within trial, years' experience as a pharmacist and as a prescriber).
Results: Number of residents and pharmacist-independent prescriber employment within a medical practice were positive predictors of deprescribing.
Conclusion: Previous experiences were not related to deprescribing likelihood.
Increasing the number of residents increases the opportunity for deprescribing and therefore this relationship is intuitive. The location within a medical practice is an interesting finding that requires further exploration to understand its exact nature. (or deprescribe) without medical authorization for any condition within their clinical competence 6,7 This extended role means that pharmacist-independent prescribers (PIPs) are well placed to deliver a full medication management service to care homes. 8  Deprescribing was the most frequently recorded activity undertaken within CHIPPS, 9 with substantial variation identified in the rate of deprescribing between PIPs. 9 It was hypothesized that 1 or more contextual factors might be prominent in causing the variation and understanding this could be used to inform future interventions and policy recommendations to promote deprescribing in care homes.
During the development phase, employment of PIPs within medical practices was identified as the preferred model for CHIPPS. A medical practice is an organization consists of 1 or more of general practitioners (GPs) who provide primary care to a particular group of patients. 10 This study aimed to quantify and characterize deprescribing activity performed by PIPs in terms of proactive or reactive deprescribing and identify any relationships between contextual factors and deprescribing. 5  Only those interventions (n = 284) associated with a medication being stopped by PIPs were included in our analysis.

| METHOD
Our analysis comprised characterizing deprescribing interventions according to whether they were proactive or reactive and identifying any relationships between contextual factors and rate of deprescribing.

| Phase 1: characterizing deprescribing activity
Deprescribing interventions were extracted from the CHIPPS trial database including the medication name, British National Formulary (BNF) medication classification, 11 dose and the rationale for the intervention documented by the PIP. 9 Two clinical pharmacists (M.A. and S.S.) independently categorized PIP medication discontinuation interventions as reactive or proactive deprescribing according to the accepted definition. 5 Medication deprescribed had been classified according to BNF. 11 Medication discontinuation activities related to a change from regular to when required dosage and those stopped at the end of a short treatment course, such as an antibiotic, were excluded as these do not meet the criteria for reactive or proactive deprescribing. 5 Scott et al. defined proactive deprescribing as deprescribing a medication when future benefits are unlikely to outweigh future harms, and reactive deprescribing as deprescribing a medication in response to an adverse clinical trigger. 5 Inter-rater reliability on deprescribing categorization was assessed using Cohen's κ, with κ = 0.6-0.8 considered good and κ > 0.8 excellent. 12 Any disagreements were resolved through discussion and referral to a third reviewer (D.W., pharmacist).
The authors also checked whether deprescribing activities had been sustained at the end of the CHIPPS trial (6 months).

| Phase 2: identifying relationships
We explored any relationships between the deprescribing activities characterized in phase 1 and all contextual factors captured in the CHIPPS trial 13 : • Number of years qualified as a pharmacist All 6 contextual factors were entered into the regression analysis.
Following backward elimination, only the number of residents in assigned care home and whether the pharmacist was employed in a medical practice were significantly associated with increased deprescribing interventions (see Table S2 for full analysis). The resulting model predicted 50.5% (adjusted R 2 ) of variance in the number of deprescribing interventions. The contextual factors predicting a PIP performing deprescribing interventions are provided in Table 2. The full regression analysis is provided in Table S3. Because the bootstrap results were similar to the parametric modelling, the results were resistant to deviations from distributional assumptions, and only the parametric modelling is provided for brevity.

| DISCUSSION
This study found that PIPs' interventions were dominated by deprescribing that was proactive in nature, that is, to prevent future medication-related harm. Interestingly medicines for use for conditions associated with the central nervous system were the most commonly deprescribed and that experience of the PIP either as a pharmacist, prescriber or of working with care homes were not predictors of deprescribing activity.
Increasing the number of residents to be reviewed will understandably increase deprescribing activity as the opportunity for deprescribing increases. The most interesting finding however was the fact that those PIPs already employed in medical practices were more likely to deprescribe medicines that those whose main employment contract was elsewhere.
This analysis is based on a small number of PIPs working within a relatively small number of care homes. Furthermore, the PIPs were self-selected for the trial and received specific training which, while