Development and regional validation of a literacy-fair verbal cognitive screen for older adults in India
BackgroundBrief cognitive screens could narrow India's dementia detection gap, but widely used instruments depend on reading, writing, or clock-drawing and lose accuracy in the large illiterate older population; Alzheimer's disease and related dementias drive most of this burden.ObjectiveTo derive and validate a fully verbal, literacy-fair candidate cognitive screening model.MethodsIn LASI-DAD Wave 1 (n = 4096) we derived a screen from prop-free verbal tasks (LASSO then forward selection) against algorithmic DSM-5 classification, compared discrimination (AUC) with reconstructed Hindi Mental State Examination (HMSE), Mini-Cog, Community Screening Instrument for Dementia (CSID) and Telephone Interview for Cognitive Status (TICS), performed leave-one-region-out validation and addressed incorporation bias using clinician Clinical Dementia Rating (CDR) ratings (n = 2527).ResultsThe sample was 57.5% illiterate; weighted prevalence was 26.7% (any impairment) and 9.0% (dementia). A four-item model-word-list learning, orientation to time, object naming, animal fluency ("Verbal-4")-reached AUC 0.757 for any impairment and 0.785 for dementia, indistinguishable from the HMSE, higher than the Mini-Cog, CSID and TICS (all p < 0.001), and similar across literacy (0.778 versus 0.776). In the non-random CDR-rated subsample, dementia discrimination was 0.872 against the clinician CDR versus 0.775 against the DSM-5 algorithm. Any-impairment discrimination attenuated in South India (0.688), analyses suggesting substantial case-mix and/or reference-standard contributions. A positive screen predicted mortality (adjusted HR 1.63, 95% CI 1.39-1.90).ConclusionsThis candidate screening model matched full-battery accuracy without a literacy gradient, held against a partially independent clinician standard, and carried independent prognostic information; it requires prospective calibration into an operational score before field use.
Authors
- Aparajit Ballav Dey (ORCID: https://orcid.org/0000-0003-0979-9413)
- Sharmistha Dey (ORCID: https://orcid.org/0000-0003-4122-7415)
- Bhrigu Jain (ORCID: https://orcid.org/0000-0001-8785-3759)
Institutions
- All India Institute of Medical Sciences (IN)
Publication Details
- Journal
- Journal of Alzheimer s Disease
- Published
- 2026-11-17
- DOI
- https://doi.org/10.1177/13872877261507060
- Primary Topic
- Dementia and Cognitive Impairment Research
- Type
- article
- Field-Weighted Citation Impact
- 0.00