Does the Morisky Medication Adherence Scale Work for Measuring Adherence to Oral Contraceptives?
Abstract
Objective: The objective of this study is to evaluate the measurement properties of the Morisky Medication Adherence Scale in appraising adherence to oral contraceptives. Methods: Women attending a Midwest university were randomly sampled and invited to participate in an online survey (32.6% response rate). The 656 women reporting use of oral contraceptives responded to the 8-item Morisky Medication Adherence Scale (MMAS-8), and indicated the number of missed doses of contraceptive per month. A proxy for perfect adherence was computed (i.e., took contraceptive yesterday and no monthly missed doses). Evaluation of the MMAS-8 included Cronbach α, confirmatory and exploratory factor analysis, and a comparison with self-report of missing doses and perfect adherence. Results: The Cronbach α for the MMAS-8 was 0.637. The proposed one-factor structure of the MMAS-8 did not fit the data entirely. While the standardized root mean square residual was adequate (0.041), the comparative fit index (0.90) and root mean square error of approximation (0.07) did not meet criteria for good model fit. A principal component analysis with Varimax rotation was performed, and based upon a parallel analysis a one-factor solution was chosen accounting for 31.2% of the variance. Significantly more of those in the high adherence group (MMAS-8 score = 8) reported missing zero doses in a typical month (P <0.001) and last month (P <0.001). Using a proxy for perfect adherence, the MMAS-8 high adherence category had a sensitivity of 79.4% and specificity of 86.6%. Conclusions: This work reveals strengths and weaknesses of using the MMAS-8 in oral contraceptive adherence research. The internal consistency was lower than desired, and results from exploratory and confirmatory factor analysis suggest that item trimming may be necessary. Despite these caveats, MMAS-8 categories differentiated those who forget doses of their oral contraceptives. Further instrumentation work may reveal the MMAS-8 items most relevant to gauge oral contraceptive taking behaviors.