Medication nonfulfillment rates and reasons: narrative systematic review

Background: The literature on nonfulfillment of prescription medications spans over three decades of work. There is a wide variation in reported nonfulfillment rates, but no previous study has systematically reviewed this literature to explore the reasons behind this variation. Objective: The objective of this study was to review estimates of medication nonfulfillment rates and published reasons for nonfulfillment and explore whether nonfulfillment rates vary by study variables. Methods: Articles were identified through searches conducted on MEDLINE, CINAHL, Psych Info, and EMBASE, and review of relevant reference citations. Methodological variables, nonfulfillment rate, and unit of analysis (i.e., patient or prescription) were abstracted from each article selected for review. Mean and median nonfulfillment rates for groups categorized by unit of analysis and selected methodological variables (method for assessing nonfulfillment, sample characteristics, disease subgroup, sample size, country of data collection, recall period or time allowed before classifying as nonfulfillment, and year of study) were calculated. Reasons for nonfulfillment were abstracted from all articles that included a relevant discussion. Findings: A total of 79 studies reporting pure nonfulfillment rates (59 at the patient level and 20 at the prescription level) and six studies reporting nonfulfillment rates in combination with nonpersistence rates were included. There was a wide variation in nonfulfillment rates reported by the studies - from 0.5% to 57.1%. The three primary reasons for nonfulfillment identified from this review were perceived concerns about medications, lack of perceived need for medications, and medication affordability issues. Conclusion: To the best of the authors' knowledge, this study is the first narrative systematic review on nonfulfillment of prescription medications. Despite the wide variation in individual study rates, the mean and median rates across different modes of data collection and sources of data were in a relatively narrow range (11% to 19%) and surprisingly close to the overall mean (16.4%) and median (15%.0) rates for all studies. The reasons for nonfulfillment identified through this review address barriers to nonfulfillment at the patient, physician, and health system level and thus bear important implications for policy makers.