Compliance from self-reported versus pharmacy claims data with metered-dose inhalers

OBJECTIVE: To determine how well self-reported compliance with metered-dose inhalers (MDis) agrees with compliance determined from pharmacy claims data. Additional analyses were conducted to establish psychometric properties of the questionnaire. METHODS: A cross-sectional survey was performed of all adult enrollees with asthma of a managed care organization. The overall return rate was 63.4% (696/1098). Patents with persistent asthma and at least one prescription claim for a controller MDI during the four months preceding the survey were included. Pharmacy claims data were used to calculate a compliance ratio. Self-reported compliance was based on a four-item questionnaire. The primary end points were the percent-agreement and the overall correlation between the two measures. The questionnaire's internal reliability and construct validity were assessed using Cronbach's cc and Pearson's correlations, respectively. Mean item response scores for compliant patients were compared with scores of noncompliant patients using the Student's t-test. RESULTS: One hundred questionnaire respondents met the study criteria. The mean +/- SD compliance ratio was 0.52 +/- 0.27, the mean SD self-reported compliance score was 4.07 +/- 1.0. The overall percent agreement and correlation between self-reported compliance score and the compliance ratio were 75.5% and r = 0.348 (p = 0.01), respectively. The percent agreement was highest (85%) when noncompliant patients (by compliance ratio <0.8) reported being noncompliant (questionnaire score <5). Cronbach's cc was 0.86. CONCLUSIONS: This study demonstrated a moderate correlation between self-reported asthma compliance and the compliance ratio determined using claims data. The questionnaire may be useful as a screening tool to detect noncompliant behavior in populations or individual patients with asthma.