Concordance of adherence measurement using self-reported adherence questionnaires and medication monitoring devices

The primary objective of this review was to identify and examine the literature on the association between medication adherence self-reported questionnaires (SRQs) and medication monitoring devices. The primary literature search was performed for 1980-2009 using PubMed, PubMed In Process and Non-Indexed, Ovid MEDLINE, Ovid MEDLINE In-Process, PsycINFO (EBSCO), CINAHL (EBSCO), Ovid HealthStar, EMBASE (Elsevier) and Cochrane Databases and using the following search terms: 'patient compliance', 'medication adherence', 'treatment compliance', 'drug monitoring', 'drug therapy', 'electronic', 'digital', 'computer', 'monitor', 'monitoring', 'drug', 'drugs', 'pharmaceutical preparations', 'compliance' and 'medications'. We identified studies that included SRQs and electronic monitoring devices to measure adherence and focused on the SRQs that were found to be moderately to highly correlated with the monitoring devices. Of the 1679 citations found via the primary search, 41 full-text articles were reviewed for correlation between monitoring devices and SRQs. A majority (68%) of articles reported high (27%), moderate (29%) or significant (12%) correlation between monitoring devices (37 using Medication Event Monitoring System [MEMS?] and four using other devices) and SRQs (11 identified and numerous other unnamed SRQs). The most commonly used SRQs were the Adult/Pediatric AIDS Clinical Trial Group (AACTG/PACTG; 24.4%, 10/41) followed by the 4-item Morisky (9.8%, 4/41), Brief Medication Questionnaire (9.8%, 4/41) and visual analogue scale (VAS; 7.3%, 3/41). Although study designs differed across the articles, SRQs appeared to report a higher rate of medication adherence (+14.9%) than monitoring devices. In conclusion, several medication adherence SRQs were validated using electronic monitoring devices. A majority of them showed high or moderate correlation with medication adherence measured using monitoring devices, and could be considered for measuring patient-reported adherence prospectively.