Designing and Psychometric Properties of Self-Care Tool for Adults With Pre-Diabetes: Exploratory Sequential Mixed Method

Abstract

Background Self-care is one of the most critical factors in disease prevention. Adults with pre-diabetes are at 5 to 15 times higher risk of developing type 2 diabetes compared with others. Without self-care behaviours to promote health and prevention, more than 70 will ultimately develop type 2 diabetes during their lives. Objective This study aimed to design and psychometrically evaluate the self-care of adults with pre-diabetes. Methods This study was a sequential exploratory mixed-methods study. In the first phase of the mixed-methods study, a qualitative study was conducted with a directed content analysis approach according to Riegel et al.'s middle-range theory as a guide. This qualitative-directed content analysis was conducted on prediabetes from June 2023 to October 2023. The experiences of 39 adults with pre-diabetes and 6 healthcare workers were assessed through individual, face-to-face, semi-structured interviews. The data were analysed based on the Elo and Kyng & auml;s's method. The psychometric properties of the primary tool were evaluated in the second phase. Face and content validity, item analysis, structural validity, internal consistency, relative and absolute reliability, interpretability, responsiveness, and feasibility were evaluated, and the scoring method was determined. Results The concept of self-care in prediabetes includes behaviours that are performed to return blood sugar to a normal state in a routine and usual way (self-maintenance) and behaviours in response (self-management) to the changes that have been detected following the follow-up and interpretation of symptoms, periodic examinations and tests (self-monitoring). The primary tool entered the psychometric evaluation phase with 57 items (blueprint). After performing face and content validity and item analysis, the number of items was reduced to 29 items. Exploratory factor analysis was performed with 29 items and 207 people with prediabetes, and finally, three subscales with 19 items were formed, which explain 38 of the total extracted variance. The results of confirmatory factor analysis with 200 samples indicated the acceptable fit of the model. The Cronbach's alpha of all subscales was higher than 0.7, and the intraclass correlation coefficient of the scale was higher than 0.90. The standard error of measurement was 1.340, the minimum detectable change was 6.57, and the minimal important change was 3.71. The total score of the questionnaire had no ceiling and floor effect; the percentage of unanswered items was within the acceptable range. Conclusion The results show that the self-care questionnaire for prediabetes has good psychometric properties and can measure self-care in adults with pre-diabetes.