Appropriate polypharmacy in older people: an assessment of evidence and methodological approaches

Abstract

Introduction Improving appropriate polypharmacy in older individuals remains challenging. Numerous interventions have been designed to address this, yet their effectiveness is debated. Variations in outcomes and their related measurement create inconsistencies and hinder comparative analysis across studies. This thesis aims to examine various aspects of polypharmacy including the evaluation of effective interventions, the methodologies used in measuring outcomes, and the assessment of intervention fidelity and complexity. Methods A systematic review, conducted in accordance with the Cochrane Collaboration's methodology, assessed the effectiveness of interventions aimed at improving appropriateness of polypharmacy in older people. An extensive search across four databases, two trial registries, and other sources was performed to identify relevant trials. The Consensus-based Standards for the selection of health Measurement Instruments (COSMIN) guideline was applied to select outcome measurement instruments (OMIs) for use with a previously developed polypharmacy core outcome set (COS). OMIs were identified from systematic reviews targeting appropriate polypharmacy in older adults. The quality of the OMIs was assessed, and a Delphi study (two rounds) was conducted with a range of stakeholders to reach a consensus on OMIs for measuring outcomes in the polypharmacy COS. A paper-based assessment study applied the five domains of the National Institutes of Health Behavior Change Consortium (NIH-BCC) framework to examine intervention fidelity in the trials included in the systematic review. Finally, the ten dimensions (six core dimensions and four optional dimensions) of the intervention Complexity Assessment Tool for Systematic Reviews (iCAT_SR) tool were used to evaluate the complexity of interventions in the reviewed trials in the systematic review. Results A total of 18 randomised controlled trials (RCTs) was included in the systematic review. It was uncertain whether interventions included in these RCTs improved medication appropriateness (as measured by an implicit tool), reduced the number of potentially inappropriate medications (PIMs) or potential prescribing omissions (PPOs), or decreased the proportion of patients with one or more PIMs or PPOs (as measured by explicit tools). The certainty of evidence was very low. The systematic review also highlighted inconsistencies in outcome reporting and measurement due to high heterogeneity across the trials. Following the COSMIN guideline, 20 potential OMIs for the polypharmacy COS were identified and assessed for quality of measurement properties. Of these, seven OMIs [‘number of serious adverse drug reactions (ADRs)’, ‘medication appropriateness index’, ‘number of patients who fell’, ‘total number of prescriptions’, ‘EuroQol-5D’, ‘number of deaths’, and ‘number of symptoms of side effects’] met the COSMIN criteria for good measurement properties and demonstrated the best quality of evidence. These OMIs were presented to 50 participants from various countries and professional backgrounds who participated in the Delphi consensus study. Consensus was reached on three OMIs for the polypharmacy COS: ‘number of serious ADRs’, ‘number of deaths’, and ‘number of patients who fell’, corresponding to measuring ‘serious ADRs’, ‘mortality’, and ‘falls’, respectively. No consensus was reached for the remaining four OMIs. Using the NIH-BCC framework's five domains, intervention fidelity in the 18 trials was assessed, resulting in an overall poor fidelity rating. None of the trials covered all five domains. Barriers and facilitators to implementing effective interventions were also identified. Using the iCAT_SR tool the evaluation of each trial consistently revealed high complexity in only one core dimension: the behaviour or actions of the intervention recipients or participants targeted by the intervention. The four optional dimensions were assessed as unclear or incapable due to insufficient detail in study reports, making consistent application challenging across trials. Conclusion Evidence on the effectiveness of interventions aimed at improving the appropriateness of polypharmacy among older people remains uncertain. The identification of OMIs may support the use of the polypharmacy COS in future studies to measure and compare intervention effects, and facilitate the synthesis of trial findings, generating stronger evidence in systematic reviews. Poor fidelity and variation in intervention complexity in the included trials highlight the need for well-designed studies with highly standardised approaches to reporting these elements in interventions to improve the appropriateness of polypharmacy in older persons.