Aims: Risk evaluation tools can be useful in primary care to identify high-risk patients and provide a framework for managing chronic disease risk factors . We aimed to develop risk stratification tools to facilitate timely detection of dementia in primary care.
Methods: Candidate factors associated with dementia derived from a harmonised dataset (Kimberley Healthy Adults Project, Koori Growing Old Well Study and the Torres Strait Dementia Prevalence Study) were weighted based on their multivariable logistic regression coefficients and used to calculate a risk score. A risk index was also constructed from the unweighted sum of risk factors present divided by the total number of factors assessed. Receiver operating characteristic (ROC) curves were generated using Youden’s method to investigate the best cut-point to identify people at high risk of dementia.
Results: Ten candidate items (age, sex, education, stroke, head injury with LOC, epilepsy, diabetes, smoking, past alcohol use and depression) were weighted to construct a Cognitive Assets and Risk Evaluation (CARE) score (AUC of 0.86, 95% CI 0.82-0.90; sensitivity 83.78%, specificity 76.95% at the optimal cut point of 32). As an index, these same factors had acceptable psychometric properties (AUC 0.80, 95% CI 0.75-0.85) and discriminatory powers were enhanced with the addition of other indicators and a four-item rapid screen based on the Kimberley Indigenous Cognitive Assessment.
Conclusions: We present exploratory work to support the development of dementia risk stratification tools that may be useful in primary care to identify Aboriginal and Torres Strait Islander patients for further dementia assessment.