| Medication non-adherence and non-persistence are pressing issues in healthcare today. Their resultant health and economic ill-effects are well studied in numerous diseases. To help improve medication adherence and persistence, their effective and efficient measurement is essential. Thus, the purpose of this dissertation was to develop the Medication Non-Adherence Scale (MNAS) and the Medication Non-Persistence Scale (MNPS).;The scales were administered to patrons of three independent community pharmacies in the Southeastern United States. Their responses were anonymously linked to prescription fill data. The MNAS was validated against the past 6 months of prescription fill data, while the MNPS against the past 12 months. The MNAS was also tested against 3 months of prospective data to study its ability to predict future non-adherence.;Confirmatory factor analyses offered evidence for internal consistency reliability, and convergent and discriminant validity. The results indicated a 5 factor solution for the MNAS -- worries about side-effects, worries about addiction, worries about cost, lack of perceived need, and unintentional non-adherence - and a single factor solution for the MNPS. Linear regression analyses concluded that the scales demonstrated concurrent validity (MNAS: unstandardized regression coefficient=-0.50 (p<0.001); MNPS: unstandardized regression coefficient=-3.97 (p=0.03)). Another linear regression analysis also offered evidence for the MNAS having predictive validity (unstandardized regression coefficient=-0.62 ( p<0.001)). ROC curve analysis suggested that a score of greater than 16 on the MNAS indicated non-adherence in the past 6 months, and a score of greater than 20 indicated non-adherence in the next 3 months. A score of 1 or higher on the MNPS indicated non-persistence in the past 12 months.;The MNAS was seen to perform better than the Medication Adherence Reasons Scale (MAR-Scale: R2=0.016, standardized regression coefficient=-0.125; MNAS: R2=0.043, standardized regression coefficient=-0.208) and the 1986 Morisky scale (R2=0.018, standardized regression coefficient=-0.134) in estimating concurrent PDC, and better than the Adherence Estimator (AE: R2=0.010, standardized regression coefficient=-0.099; MNAS: R2=0.083, standardized regression coefficient=-0.288) in estimating future PDC. These estimates were also statistically significantly different from each other.;Thus, the MNAS and MNPS help fill vital gaps in adherence and persistence measurement, and may be used by healthcare practitioners and researchers to improve patient health. |
