Development and psychometric properties of the Chinese version of the Impaired Awareness Hypoglycemia Assessment Scale

Objectives: This study was conducted to translate Clark’s and Gold’s questionnaires on impaired awareness of hypoglycemia (IAH) from English to Chinese (Clarke-TW and Gold-TW, respectively) and evaluate their reliability and validity. Methods: A bilingual expert group translated and back-translated the original questionnaires to develop the Clarke-TW and Gold-TW questionnaires. Convenient sampling was performed across 19 pharmacies, clinics, and public health centers in Tainan between September to November 2022. Thus, we recruited 898 patients (age ≥ 20 years) with type 2 diabetes who had received insulin injection or oral sulfonylureas over the previous 3 months. Face-to-face interviews were conducted with these individuals. IAH prevalence, internal consistency, test–retest reliability, and construct validity were evaluated for both questionnaires. Results: The bilingual expert group confirmed the content validity of the two questionnaires. For Clark-TW, the Kuder–Richarson-20 coefficient was 0.481 and the 4-week test–retest intraclass correlation coefficient was 0.660. The intraclass correlation coefficient for Gold-TW was 0.653. IAH prevalence measured using Clarke-TW and Gold-TW exhibited no association with sex (p = .32 and .70, respectively) or employment status (p = .08 and .29, respectively). However, routine self-monitoring of blood glucose and a retinal examination in the previous year were significantly associated with reduced IAH prevalence regardless of the measurement tool used (p < .001); this finding supports the questionnaires’ known-group validity. Factor analyses identified two factors from Clarke-TW; principal axis factoring indicated that the two-factor model fit was adequate. Conclusions: Both Clarke-TW and Gold-TW have moderate test–retest reliability and satisfactory construct validity. The overall IAH prevalence measured using Clarke-TW and Gold-TW was 36.1% (95% confidence interval: 32.9%-39.3%) and 49.4% (95% confidence interval: 46.1%-52.3%), respectively. Both questionnaires are straightforward and convenient and can be used in community surveys to estimate IAH prevalence among patients with diabetes.