OBJECTIVES: The Beliefs about Medication Questionnaire (BMQ) is a measure of patients' medication beliefs (perceived need for medications and perceived medication concerns), and it has been used widely in investigations of medication adherence. To date, there has been little assessment of the relative importance of necessity vs. concerns in determining medication adherence. Using a systematic literature review and subsequent meta-analysis, this study evaluates the impact of BMQ necessity, concerns, and necessity-concerns differential as a predictor of medication adherence. METHODS: Articles were identified through searches conducted on MEDLINE, CINAHL, Psych Info, EMBASE, International Pharmaceutical Abstracts, PubMed, and review of reference citations. Methodological variables, effect sizes of associations, diseases, and measures of adherence were abstracted from each eligible article. Studies were categorized by BMQ measures (necessity, concerns, and the differential), statistical significance (bivariate or multivariate significance, insignificance, or not applied), disease category, and adherence-outcome metric. The relative impact of BMQ measures on adherence across different categories was assessed. RESULTS: Across 77 studies, significant multivariate associations were reported between adherence and perceived concerns (57.4%), perceived necessity (75.6%), and the differential (88.9%). Two-thirds of the 33 multivariate analyses demonstrated higher effect sizes (odds ratios or standardized regression coefficients) between necessity and adherence than between concern and adherence. There was wide variation between BMQ measures and adherence across diseases and adherence metrics. For example, necessity was significantly associated with adherence in 100% of diabetes studies but 0% of renal studies. Self-reported adherence metrics and pill counts had the lowest and highest rates, respectively, of statistical significance with BMQ. CONCLUSIONS: Perceived need for medications is a more potent predictor of adherence than medication concerns. Perceived need for medication is a mutable patient belief.  Adherence interventions may improve their effectiveness if perceived need for medications became a central theoretical and interventional focus.