Aging is a complex, lifelong process characterized by the accumulation of chronic conditions, functional decline, and increasing heterogeneity in health trajectories. Improving our understanding of aging is essential for identifying critical windows for prevention and intervention in older populations. However, longitudinal aging studies pose substantial statistical challenges due to intermittent observation, partial observability of health trajectories, and the presence of multiple biological processes evolving continuously over time.
In this seminar, I will present my current research on the development and application of statistical methodologies to address etiological questions in aging research and to construct longitudinal aging metrics that support inference, individualized prediction, and data-driven evaluation of preventive interventions for healthy aging.
The focus will be on extensions of classical multistate models that address interval censoring and hidden states, as well as on multivariate mixed latent formulations motivated by research questions in dementia, multimorbidity, and biological aging. Simulation studies and applications to population-based aging cohorts are used to evaluate model performance and to illustrate the practical implementation of these methods.