Race-Specific Dementia Risk Prediction Using 2024 Lancet Commission Risk Factors and Resting Heart Rate: A Survival Analysis of 55,004 NACC Participants
Alaka, S. A.; Ngan, S.-F. C.; Iyappan, R.; Nwaeze, J.; D'Amore, B.; Katoueezadeh, M.; Thinakaran, Y.; Laein, M. H.; Baker, J.; Sze, S. K.
Show abstract
Persistent racial disparities in dementia raise concerns regarding the validity and generalizability of existing prognostic models across diverse populations. We evaluated the utility of the 2024 Lancet Commission risk factors and resting heart rate (RHR) for race-specific dementia risk prediction using data from 55,004 participants in the National Alzheimer's Coordinating Center cohort. Cox proportional hazards and Random Survival Forest (RSF) models were developed separately for Black, White, Asian, and American Indian participants to predict 1-, 3-, and 5-year time to dementia. RSF consistently outperformed Cox models across all racial groups and prediction horizons, achieving 5-year AUCs of 0.88-0.91 compared with 0.69-0.81 for Cox models. Inclusion of RHR modestly and consistently improved predictive performance across racial groups. Predictor importance varied between racial groups, suggesting heterogeneity in dementia risk profiles and disease presentation. These findings support the potential utility of RHR as a complementary prognostic biomarker and highlight the importance of equitable, personalized dementia risk prediction across diverse populations.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Profiles of cognitive change in preclinical Alzheimer's disease using change-point analysis 96%
- Affective Neuropsychiatric Symptom Metrics in the National Alzheimers Coordinating Center Dataset 95%
- Blood Biomarkers for Diagnosis & Differential Diagnosis of Alzheimers Disease in Real-World Clinical Populations: A Systematic Review 95%
Similar papers in this journal
- Risk models based on non-cognitive measures may identify presymptomatic Alzheimer’s disease 96%
- Examining a Preclinical Alzheimer’s Cognitive Composite for Telehealth Administration, the tPACC, for Reliability between In-Person and Remote Cognitive Testing with Neuroimaging Biomarkers 95%
- Inferring Alzheimer’s disease pathologic traits from clinical measures in living adults 95%
Similar papers in this journal
- Interpretable multivariate survival models: Improving predictions for conversion from mild cognitive impairment to Alzheimers disease (AD) via data fusion and machine learning 95%
- Examining heterogeneity in dementia using data-driven unsupervised clustering of cognitive profiles 94%
- c-Triadem: A constrained, explainable deep learning model to identify novel biomarkers in Alzheimer’s disease 94%
Similar papers in this journal
- Predicting future cognitive impairment in preclinical Alzheimer's disease using multimodal imaging: a multisite machine learning study 96%
- Integrating plasma, MRI, and cognitive biomarkers for personalized prediction of decline across cognitive domains 96%
- Vascular risk burden is a key player in the early progression of Alzheimer’s disease 95%