Purpose: The purpose of this study was to examine the characteristics of the Spanish-language diabetes self-efficacy scale (DSES-S) and the English-language version (DSES).
Methods: This study consists of secondary data from 3 randomized studies that administered the DSES-S and DSES at 2 time points. The scales consist of 8 Likert-type 10-point items. Principal component analysis was applied to determine if the scales were unitary or consisted of subscales. Univariate statistics were used to describe the scales. Sensitivity to change was measured by comparing randomized treatment with control groups, where the treatment included methods designed to enhance self-efficacy. General linear models were used to examine the association between the scales and the 8 medical outcomes after controlling for demographic variables.
Results: Principal component analysis indicated that there were 2 subscales for both versions: self-efficacy for behaviors and self-efficacy to manage blood levels and medical condition. The measures had similar means across the 3 studies, high internal consistent reliability, values distributed across the entire range, and they showed no evidence of floor effects and little evidence of ceiling effects. The measures were sensitive to change. They were associated with several health indicators and behaviors at baseline, and changes were associated with changes in health measures.
Conclusions: The self-efficacy measures behaved consistently across the 3 studies and were highly reliable. Associations with medical indicators and behaviors suggested validity, although further study would be desirable to compare other measures of self-efficacy for people with type 2 diabetes. These brief scales are appropriate for measuring self-efficacy to manage diabetes.
© 2016 The Author(s).