RESUMO
The purpose of this research was to develop survey instruments to evaluate diabetes knowledge and self-efficacy in a diverse population, and investigate the psychometric properties of data obtained with these instruments using Rasch measurement. Two-hundred and fifty-five urban-dwelling participants with diabetes were recruited to complete surveys through independent interviews. To evaluate the association of health literacy on metabolic control, formal literacy and hemoglobin A1c fingerstick testing were performed. Rasch analysis of the data yielded item and person calibrations for self-efficacy and knowledge, with variable maps created to provide both norm and criterion-referenced interpretations. Knowledge scale person separation reliability was 0.50 and item separation reliability was 0.98; while self-efficacy scale person separation reliability was 0.72 with item separation reliability of 0.92. Statistically significant partial correlations were observed between knowledge and health literacy (r = 0.41, p<.001), and self-efficacy and hemoglobin A1c (r = -0.33, p<.001). However, there was no correlation between diabetes knowledge and hemoglobin A1c (r = 0.035, p = 0.29), or health literacy and A1c (r = 0.022, p = 0.36). Diabetes knowledge varied, with non-English speaking individuals having lower measures than English speakers (t(252) = -4.86, p<.001). Non-English speaking individuals also had lower self-efficacy measures than English speakers (t(251) = -2.68, p = .008). Current knowledge deficits and perceptions of self-management may be estimated visually through variable mapping, which may help in individualizing informational needs for people with diabetes.