Outcome Measures for Interventions to Reduce Inappropriate Chronic Drugs: A Narrative Review

Inappropriate prescribing is a highly important problem, given the growing aging multimorbid population with associated polypharmacy. An increasing number of studies have recently developed and tested interventions to withdraw inappropriate drugs, a process called deprescribing. However, we still lack complete information on the types and prevalence of measures used to assess the success of such interventions.

To categorize and synthesize the full spectrum of measures used in intervention studies focused on reducing inappropriate prescribing of chronic drugs in adults, to standardize measurements in future studies and help researchers design studies inclusive of the important measure types.

We searched Ovid/MEDLINE to identify intervention studies focused on deprescribing chronic drugs in adults, published between 2010 and 2019.

We extracted data on study characteristics, intervention components, and outcome measures. We categorized and synthesized the measures using a comprehensive and systematic framework, separating measures of intended and unintended consequences.

Most (90/93) studies used measures of appropriate prescribing, such as drug cessation or dose reduction. The following measures were used infrequently across studies: patient-reported experience, preferences, and outcome (12 (13%), 2 (2%), and 25 (27%) studies, respectively); provider-reported experience (11 (12%) studies); patient-provider interaction (4 (4%) studies); and measures of unintended consequences (24 (26%) studies). Studies varied in the type and number of measures assessed, ranging from 1 to 20 different measures by study.

To ensure initiation, success, and long-term sustainability of deprescribing, it is important to assess the success of intervention studies using clinically relevant patient- and provider-centered measures. This categorized synthesis of outcome measures used in deprescribing studies may facilitate implementation of important measure types (e.g., patient-reported measures and measures of unintended consequences) in future studies. J Am Geriatr Soc 68:2390-2398, 2020.