RESUMO
OBJECTIVE: With the rising burden of dementia globally, there is a need to harmonize dementia research across diverse populations. The Addenbrooke's Cognitive Examination-III (ACE-III) is a well-established cognitive screening tool to diagnose dementia. But there have been few efforts to standardize the use of ACE-III across cohorts speaking different languages. The present study aimed to standardize and validate ACE-III across seven Indian languages and to assess the diagnostic accuracy of the test to detect dementia and mild cognitive impairment (MCI) in the context of language heterogeneity. METHODS: The original ACE-III was adapted to Indian languages: Hindi, Telugu, Kannada, Malayalam, Urdu, Tamil, and Indian English by a multidisciplinary expert group. The ACE-III was standardized for use across all seven languages. In total, 757 controls, 242 dementia, and 204 MCI patients were recruited across five cities in India for the validation study. Psychometric properties of adapted versions were examined and their sensitivity and specificity were established. RESULTS: The sensitivity and specificity of ACE-III in identifying dementia ranged from 0.90 to 1, sensitivity for MCI ranged from 0.86 to 1, and specificity from 0.83 to 0.93. Education but not language was found to have an independent effect on ACE-III scores. Optimum cut-off scores were established separately for low education (≤10 years of education) and high education (>10 years of education) groups. CONCLUSIONS: The adapted versions of ACE-III have been standardized and validated for use across seven Indian languages, with high diagnostic accuracy in identifying dementia and MCI in a linguistically diverse context.