Sensitivity and specificity of cognitive and functional screening instruments for dementia: the Indo-U.S. Dementia Epidemiology Study

Abstract

There is a shortage of adequate screening instruments for dementia in poorly educated populations and non-English-speaking groups. An epidemiological survey was conducted in a population-based, largely illiterate, sample of 5,126 individuals aged 55 and older in 28 villages in the rural community of Ballabgarh in northern India. All participants were administered a general mental status test, the Hindi Mental State Examination (HMSE), and a brief battery of neuropsychological tests. Their informants answered a questionnaire assessing functional ability, the Everyday Abilities Scale for India (EASI). Six hundred thirty-two participants underwent clinical diagnostic evaluation for dementia. We investigated whether the sensitivity, specificity, and predictive value for dementia of the mental status test could be improved by the addition of the brief neuropsychological test battery or the functional questionnaire, comparing the instruments alone and in combination. In participants who could be tested cognitively, the HMSE, the neuropsychological battery, and EASI had sensitivities of 81.3%, 81.3%, and 62.5%, respectively, with specificities of 60.2%, 74.5%, and 89.7%, respectively. The combination of all three was 93.8% sensitive and 41.8% specific. The sensitivity of the HMSE alone was nonsignificantly improved by the addition of either the EASI or the neuropsychological battery, whereas its specificity was significantly decreased by either addition. An advantage of the EASI was that it could also be administered to informants of subjects who were cognitively untestable. In this largely illiterate community, with a low prevalence of dementia, the combination of cognitive tests and a functional ability questionnaire had substantial value for population screening.