Variations of a Commonly Used Medication Adherence Assessment Scale: Do Changes in Scale Change Structure Results?

Background: Medication nonadherence is a major barrier to both patients and health care professionals when trying to manage medical conditions. An appropriate self-report adherence tool would be helpful in determining a patient's medication adherence. Objectives: To observe variations in scale scores based on modifications to an Original Adherence Scale, with the hypothesis that making modifications to the Original Adherence Scale will create variations in the percentage of adherent patients. Methods: This cross-sectional study utilized mailed surveys to people identified in a prescription claims administrative dataset who had a pharmacy claim for at least 2 antihypertensive medications. One thousand people were equally divided and randomly placed in 1 of 4 groups: Original Adherence Scale Group, Time Reference Scale Group, 4-Point Likert-Type Scale Group, Multiple Medication Scales Group. Each scale underwent assessment of internal reliability using Cronbach's α. Changes made to the Original 4-item scale included altering the time reference period from 3 months to 7 days, changing response options from Yes/No to a Likert-type scale, and incorporating multiple scales so that the respondent may report on up to 4 different options. Results: There were 437 surveys completed appropriately, yielding a 46.4% response rate. The overall scale scores indicating perfect adherence was 51.8% for the 4-Point Likert-Type Scale Group, 66.5% for the Multiple Medication Scales Group, 68.8% for the Original Adherence Scale Group, and 78.9% for the Time Reference Scale Group. Conclusion: When there are more selection options, a change in time reference, or more medications reported, the amount of adherent patients varied.