Adherence Definitions, Measurement Modalities, and Psychometric Properties in HIV, Diabetes, and Nutritional Supplementation Studies: A Scoping Review.

Measuring adherence has been a priority for researchers to help inform effective care for patients regularly consuming medications for chronic conditions. As a widely accepted "gold standard" adherence measure or operational definition does not exist, studies measure adherence using different modalities, which may lead to different conclusions about adherence patterns. The purpose of the scoping review was to identify modalities used to measure adherence to HIV medication, diabetes medication, and nutritional supplementation and explore the variation in adherence definitions, measurement modalities, and psychometric properties being reported across studies. Comprehensive searches were performed in PubMed, Scopus, and PsycINFO from January 2012 to January 2022. We included studies reporting psychometric properties of adherence/compliance to HIV medication, diabetes medication, or nutritional supplements. In total, we included 88 studies in the review. The 8-item Morisky Medication Adherence Scale (MMAS-8) was the most frequently used self-reported measure. We found almost no relationship between country income level and triangulation levels. The operational definition of adherence fell into four categories: numerical, dichotomous, ranked ordinal, and undefined. The amount of variation in an adherence definition category within a modality depended on whether the measures within the modality could be assessed numerically and whether widely accepted cutoffs existed for the measure. Across studies, 46 (52%) reported both validity and reliability, 28 (31%) reported validity only, and 14 (16%) reported reliability only. Fourteen types of validity and six types of reliability were identified across the studies. Measuring adherence accurately and reliably continues to be a challenge for research in HIV, diabetes, and nutritional supplementations. When reporting adherence measurements, we suggest including adherence results from multiple measures and modalities, presenting adherence results numerically, and reporting multiple types of validity and reliability.