Evaluating a Two-Step Approach: Combining Brief Risk Indicators and Short Cognitive Assessments for Improving Early Detection of Cognitive Impairment

Background Scalable approaches for early detection of cognitive impairment are urgently needed. A two-step screening approach, in which short cognitive assessments are administered only to individuals at risk, offers an efficient alternative to universal screening. This study evaluated combinations of brief self-report risk indicators and short cognitive assessments to identify models that accurately classify cognitive impairment. Methods Participants aged 60+ from dementia specialty settings were classified as cognitively normal (CN; n = 87) or cognitively impaired (CI; n = 66). All completed three risk indicators, a modified self-report Brief Dementia Screening Indicator (BDSI), self-report of personal memory concerns, self-report of memory concerns expressed by others (other concern), and two short cognitive assessments (Quick Mild Cognitive Impairment screen [Qmci] and Speeded Matching [SM]). Logistic regressions using continuous variables and composite binary predictors predicted cognitive status. Results Models significantly predicted cognitive status ( P < .001; AUCs = 0.84-0.91 for continuous models and 0.59-0.79 for composite models). Among risk indicators, the BDSI was least informative, while other concern was the strongest predictor. Qmci and SM performed comparably as continuous variables. The best-performing composite predictor derived from other concern and Qmci yielded OR = 28.70 ( P < .001), sensitivity = 63.6%, specificity = 94.3%. Conclusion A two-step approach using brief self-report risk indicators and short cognitive assessments shows promise for efficient detection of cognitive impairment, providing initial validation in dementia specialty settings to support evaluation in primary care.