Objective: Adhering to prescribed medication is often a problem for patients with diabetes yet there is no consensus on how best to measure adherence in this patient population. This systematic literature review critically reviewed and summarized the methods used to measure medication adherence in patients with diabetes (on oral hypoglycemic agents [OHAs] and/or insulin) in original research published between 2007-2013.

Study design: Literature review.

Methods: A systematic search for methods to assess medication adherence in patients with type I or type II diabetes was conducted using PubMed, EMBASE, PsychInfo, and Cochrane databases. Two researchers independently screened abstracts for initial eligibility and then applied the inclusion/exclusion criteria to the relevant full-text articles.

Results: Fifty-nine articles met the criteria for inclusion. Subjective assessment (observer-reported and patient-reported), pill counts, Medication Event Monitoring System (MEMS), cell-phone real-time assessment, and logbooks were used in prospective studies. In pharmacy claims databases, medication possession ratios (MPRs), or some derivation thereof, were utilized. Each method has strengths and weaknesses, but few approaches specifically addressed issues unique to assessing insulin adherence. Three novel approaches (using cell-phone real-time assessment, computerized logbooks, and a questionnaire about different dosing irregularities) provided insight on timing and dosing issues that could be useful for highlighting interventions to improve insulin adherence.

Conclusion: No gold standard exists for measuring medication adherence in patients with diabetes. The plethora of adherence methods precludes the comparison of adherence rates across studies. Greater consistency is therefore needed in adherence measurement, including question content, recall period, and response options for self-report measures. Novel methods for understanding adherence to variable-dosed insulin require further research. Researchers should select a methodology that best fits their research question, study design, patient population and resources.