Testing complex interventions for polypharmacy in real-world conditions

Polypharmacy, commonly defined as the use of at least five medications, affects approximately 37% of adults.^1,2 Although polypharmacy can be considered appropriate in some scenarios to manage multiple co-occurring chronic conditions, polypharmacy is also associated with risks, such as inappropriate prescribing, impairment in cognition and function, frailty, falls, hospitalisations, and mortality.^3–6 Health-care professionals and researchers have developed a wide variety of interventions to reduce risks from polypharmacy.^7,8 The cornerstone of many of these interventions involves a pharmacist reviewing a comprehensive list of a patient’s medications and assessing for potential inappropriate prescribing, followed by discussion with the patient. Additional components have been layered on top of these medication reviews, such as specific training for health-care professionals, tools built into the electronic health record software, greater collaboration with the prescriber, patient-facing educational materials, solicitation of patient’s goals of care and shared decision making, and involvement of family members. Some studies testing these interventions have shown positive effects in terms of reducing risks from polypharmacy, but others have not.^7,8 The mixed evidence is most likely a reflection of the great heterogeneity in both the interventions tested and the outcomes assessed.

The study by Rupert Alistair Payne and colleagues⁹ published in The Lancet Healthy Longevity builds on this mixed evidence. In a large, pragmatic, open-label cluster-randomised trial across 37 English general practices, the authors tested whether a complex, multicomponent intervention reduces potentially inappropriate prescribing. The intervention included several components and is described in detail in the protocol.¹⁰ First, a pharmacist reviewed a patient’s list of medications, discussed with the general practitioner, had a 20-min conversation with the patient, and then potentially had a follow-up conversation with the patient, when needed. Second, a bespoke informatics tool was developed to identify potentially inappropriate prescribing and generate a clinical summary. Third, a 4-h training session was provided for the pharmacist conducting the mediation review, following which a 2-h training session was provided to the practice (for both the pharmacist and the general practitioner). Fourth, financial incentives were provided to the practice (£60 per patient). Fifth, monthly feedback on potentially inappropriate prescribing was provided to the practices. For a pragmatic trial, the study was able to achieve high completion rates for the different intervention components. For example, 75% or more of the intervention sites achieved 81% or more of patient participation in conversations with the pharmacist as part of the medication review.

The main finding was that the intervention did not reduce potentially inappropriate prescribing as compared with usual care. However, of note, the usual care arm was subject to a national policy encouraging structured medication reviews and an increased number of primary care pharmacists. As a result, that the practices randomised to the usual care arm already had heightened scrutiny of potentially inappropriate prescribing is possible. Additionally, potentially inappropriate prescribing was measured on the basis of a wide variety of indicators, such as ones related to electrolyte imbalances, blood pressure treatment, respiratory decompensation, and more. The broad measure reflects the complicated nature of studying a broad primary care population with polypharmacy and at risk of a wide range of potentially inappropriate prescribing. Although some studies have narrowed focus on a specific condition or subset of indicators, a strength of this study is examining what would be done in real-world practice, when many conditions and a wide range of potentially inappropriate prescribing would be managed by the general practitioner and pharmacist.

The pharmacists providing the medication review also differed in terms of their connection to the general practice–some were embedded in the general practice, some had some connection but worked across various general practices, and some had no previous connection with the general practice. This variation again reflects real-world practices, which present great variation in the extent to which pharmacists are embedded in general practices. Whether any differences in effect exist based on the pharmacist connection to the general practice could be interesting to explore in the future.