5-Year Dynamic Prediction of Dementia Using Repeated Measures of Cognitive Tests and a Dependency Scale
Langue
EN
Article de revue
Ce document a été publié dans
American Journal of Epidemiology. 2021-11-09
Résumé en anglais
The progression of dementia prevalence over the years and the lack of efficient treatments to stop or reverse the cognitive decline make dementia a major public health challenge in the developed world. Identifying subjects ...Lire la suite >
The progression of dementia prevalence over the years and the lack of efficient treatments to stop or reverse the cognitive decline make dementia a major public health challenge in the developed world. Identifying subjects at high risk of developing dementia could improve the management of these patients and help selecting the target population for preventive clinical trials. We used joint modeling to build a dynamic prediction tool of dementia based on the change over time of two neurocognitive tests (the Mini-Mental State Examination and the Isaacs Set Tests) as well as an autonomy scale (the Instrumental Activities of Daily Living). The model was estimated on the French cohort PAQUID (1988-2015) and validated both by cross-validation and externally on the French cohort 3C (1999-2018). We evaluated its predictive abilities through Area Under the Receiver Operating Characteristics Curve and Brier score accounting for right censoring and competing risk of death and obtained Area Under the Curve equal to 0.95 in average for the risk of dementia in the next 5 or 10 years. This tool is able to discriminate a high risk group of subjects from the rest of the population. This could be of great help in clinical practice and research.< Réduire
Mots clés en anglais
Alzheimer disease
Cognition
Dementia
Dependency
Joint model
Prediction
Unités de recherche