OBJECTIVES

The study tested the four-factor structure of the Medication Adherence Reasons Scale (MAR-Scale) in additional comorbid conditions and frequencies of administration. The 20-item MAR-Scale includes one global item assessing overall adherence and 19 items describing reasons for not taking a prescribed medication to treat specific medical conditions.

METHODS

Respondents re-contacted from the 2017 United States National Health and Wellness Survey, an Internet-based survey of demographically representative adults (≥18 years), who self-reported taking medications to treat one of three conditions (pain, migraine, and sleep conditions) were invited to complete the MAR-Scale for orally administered medications. Separate MAR-Scales assessed medicines taken daily or weekly (i.e., adherence reported during the previous seven days or four weeks, respectively). Six categorical confirmatory factor analyses (CCFAs), one per condition and frequency, used a mean- and variance-adjusted weighted least squares estimator to assess response fit to the four-factor configuration identified in prior research: logistic issues, belief issues, forgetfulness issues, and long-term concerns. A CCFA was deemed to fit the data well if global fit indices exceeded the following criteria: root mean square error of approximation (RMSEA) <0.08 and comparative fit index (CFI) >0.95, and standardized factor loadings (FLs) exceeding 0.50 (and statistically significant at p<0.05, indicating strong relationships between items and the latent variables they were intended to measure).

RESULTS

CCFAs for each condition and frequency fit the data well. Across all models, RMSEAs and CFAs exceeded prespecified criteria (RMSEA range: 0.049-0.053 daily; 0.057-0.062 weekly; CFI range: 0.963-0.978 daily; 0.978-0.982 weekly). Additionally, all FLs were strong (Range: 0.734-0.985 daily; 0.696-0.970 weekly) and significant (p<0.001).

CONCLUSIONS

The MAR-Scale four-factor structure replicated beyond orals daily administration, applying also to novel weekly administration. The four factors also replicated in three novel conditions beyond several tested previously, thus extending the scope of settings to which the MAR-Scale can be applied.