Construct Validity of the Brazilian Version of the Self-Efficacy for Appropriate Medication Adherence Scale.

Abstract

This study aimed to evaluate the reliability and the construct validity of the Brazilian version of the Self-Efficacy for Appropriate Medication Adherence Scale (SEAMS). One hundred forty-seven patients with coronary artery disease participated in the study. Reliability was estimated by composite reliability (CR). Convergent validity was estimated by the factor loading and average variance extracted (AVE). The discriminant validity was assessed by cross-factor loadings, comparing the square roots of AVEs by correlation coefficients between latent constructs and disattenuated correlation. The values of CR of .88 and .99 for Factors 1 and 2, respectively, indicate internal consistency reliability. Values of AVE = 0.52 for Factor 1 and 0.69 for Factor 2 to the convergent validity of the model. Cross-factor loadings were found ≥.50 for all items of the scale in areas that were allocated in the original instrument, square roots of AVEs higher than the correlations between the constructs and desatenuada correlation (r = .83), suggesting discriminant validity fashion model. Confirmatory factor analysis partially yielded the construct validity of the Brazilian version of the SEAMS.