Concordance among three self-reported measures of medication adherence and pharmacy refill records

Abstract

Objective: To evaluate the level of agreement among three previously validated self-reported medication adherence measures and pharmacy refill records (RRs).

Design: Cross-sectional study.

Setting: Five primary care physician office sites in rural northeast Georgia.

Participants: 139 adult patients with one or more of these chronic diseases: hypertension, diabetes mellitus, hypercholesterolemia, hypothyroidism, or a condition requiring hormone replacement therapy.

Interventions: Study participants completed the Brief Medication Questionnaire (BMQ), the Medication Adherence Survey (MAS), and the Medical Outcomes Study (MOS) instruments; pharmacy RRs for the medication or medications being used to treat the target disease were obtained from pharmacies used by the study participants.

Main outcome measures: Adherence to medication therapy for target disease.

Results: Participants were nearly all white (98.6%), consistent with the Appalachian area in which the study was conducted, and mostly women (71.9%). While 91.4% of study participants reported taking their study medication most or all of the time, RRs showed mean adherence rates of 82.6%, 82.1%, 79.1%, 74.6%, and 69.8% for diabetes mellitus, hypertension, hypothyroidism, hypercholesterolemia, and hormone replacement therapy, respectively. Moderate correlations of .234, .261, and .213 were found between RRs and the MAS, MOS, and BMQ belief screen, respectively. Spearman correlations ranged from .091 between RRs and the BMQ regimen subscale to .313 between MOS and MAS. Pearson chi-square tests showed that only the BMQ belief subscale was significant in this study.

Conclusion: Because of the weak to moderate concordance found among validated measures of adherence, the selection of a useful adherence measure in pharmacy practice is difficult. These findings underscore the difficulty in both assessing patients' medication-taking behavior and assessing and comparing the results of adherence research. The development of valid and reliable measures for easily assessing medication adherence behavior in community pharmacies is needed.