The changes in DSM-5 diagnostic criteria for dementia (Major neurocognitive disorder (NCD)) and mild cognitive impairment (mild NCD) mandate a re-evaluation of screening instruments. This study attempted to validate screening instruments, identify optimum threshold, and describe their indices of efficacy.

Consecutive people above the age of 65 years attending the outpatient department of a general hospital were recruited. They were assessed using the Mini-Mental State Examination and the Vellore Screening Instruments for Dementia and were evaluated against the DSM-5 standard. Bivariate and multivariate statistics were obtained. Receiver-operating-characteristic curves were drawn, optimum thresholds obtained, sensitivity, specificity, and predictive values calculated.

One hundred and thirty four older people were recruited. The majority were women, married, with low levels of education, not employed, living with family, and had medical co-morbidity. A minority satisfied DSM-5 criteria for major (1.5%) and mild NCD (36.5%). The factors associated with NCD were older age, fewer years of education, and lower socio-economic status. MMSE, VSID patient, and VSID informant scores were significantly associated with NCD. The indices of efficacy for the MMSE and VSID patient version were modest for identifying Mild NCD. However, their performance in identifying major NCD was better. Nevertheless, optimal thresholds for recognition differed markedly from their originally recommended cut-offs.

The DSM-5 standards, with new and different cognitive domains, mandate a revaluation and recalibration of existing screening instruments. Ideally, new screening instruments, which match the cognitive domains and DSM-5 standard should be developed.