B0470
Title: Dynamic survival risk prediction: Leveraging an array of time-varying biomarkers
Authors: Zhiyang Zhou - University of Wisconsin-Milwaukee (United States) [presenting]
Abstract: The lifetime risk pooling project (LRPP) combines individual-level observations from twenty community-based studies on cardiovascular disease, which is the leading cause of death worldwide. The project involves approximately three hundred thousand participants and includes long-term follow-ups on a host of time-varying risk factors (such as blood pressure and cholesterol levels), which are recorded sparsely and irregularly. LRPP enables the development of personalized dynamic prediction models for cardiovascular risk. However, traditional joint models become computationally cumbersome when accommodating all the longitudinal risk factors covered by LRPP. To address this, a more interpretable model is developed with an efficient algorithm. Also, the proposal demonstrates competitive prediction accuracy through numerical studies.