• Elizabeth J Unni, Jeffery L Olson, and Karen B Farris.
• Roseman University of Health Sciences , South Jordan, UT , USA.
• Curr Med Res Opin. 2014 Feb 1; 30 (2): 211-21.

BackgroundMedication non-adherence is a complex phenomenon that requires tailored interventions to improve it. A new self-reported measure of medication non-adherence was previously reported based on the commonly reported reasons underlying non-adherence with the intention to match the items in the scale with tailored interventions.ObjectiveThe objectives were to revise the original Medication Adherence Reasons Scale (MAR-Scale) based on expert opinion and cognitive interviewing, and establish the psychometric properties of the revised scale.MethodsA cross-sectional design was used in cholesterol lowering and asthma maintenance medications in collaboration with an integrated medical center in the Mountain West. In the first phase, the original MAR-Scale was revised based on expert opinion and cognitive interviewing. In the second phase, the revised MAR-Scale was tested for psychometric properties in a random sample of 350 subjects on each medication.ResultsRevisions based on expert opinion included asking a global question about adherence in the past 7 days, simplifying the items and converting them into first person sentences, objective anchoring of the scale, and expanding the 'forgetfulness' item. Cognitive interviewing added one additional item to the survey. The revised MAR- Scale identified 50% of the cholesterol lowering respondents and 68% of the asthma maintenance respondents as non-adherents. An exploratory factor analysis identified four domains in the scale, with Cronbach's alpha ranging from 0.848-0.953 in cholesterol lowering and 0.827-0.891 in asthma maintenance medications. The scale also exhibited significant correlations with few other self-reported measures, consistent with hypotheses. A key limitation of the study was the moderate response rate to the survey for both medications.ConclusionThe revised MAR-Scale demonstrates better psychometric properties than the original.

18 keywords...

hide…