Preliminary development of the Medication Nonpersistence Scale.

Objectives: To develop the Medication Nonpersistence Scale (MNPS)-a multi-item self-reported scale to measure medication persistence.

Setting: Six hundred seventy-five patients patronizing 3 separate independent community pharmacies in the southeastern United States participated in this research.

Practice Innovation: The MNPS, a self-reported measure, developed to provide an estimate of, and reasons for, medication nonpersistence.

Evaluation: Cross-sectional survey data were linked with retrospective prescription fill data obtained from 3 independent community pharmacies in the southeastern United States. The MNPS factor structure was studied by means of confirmatory factor analysis (CFA), and its scale reliability and convergent validity were evaluated with the use of the results of this analysis. Its concurrent validity was tested against a standardized days-to-discontinuation measure calculated over the past 12 months, and an attempt was made to arrive at an optimum cutoff point to identify patients who have been nonpersistent with their medications.

Results: The survey yielded 675 usable patients. The CFA confirmed a single-factor solution with good model fit (root mean square error of approximation = 0.06 [90% CI 0.05-0.07]; comparative fit index = 0.96). Moderate to strong evidence of scale reliability (Cronbach alpha = 0.75; construct reliability = 0.94; index of composite reliability developed for binary items = 0.91), convergent validity (standardized factor loadings >0.5 and statistically significant), and concurrent validity (unstandardized regression coefficient = -3.97; P = 0.03) was observed. Individuals who score 1 or higher on the MNPS were considered to be nonpersistent.

Conclusion: The MNPS demonstrated good psychometric properties and offers a useful first step toward the self-reported measurement of medication persistence in clinical practice and research.