Joint seminar with CANSSI Ontario STatistics Seminar (CAST)
Kuan Liu
University of Toronto
Room: M3 3127
Bayesian Causal Methods for Cognitive Aging and Modifiable Risk Factors
Dementia is a major and growing public health challenge in Canada. Almost three-quarters of a million Canadians currently live with dementia, and this number is expected to approach one million by 2030 as our population ages. Prevention efforts focus on delaying onset and slowing progression by targeting modifiable risk factors that shape cognitive aging, with a key aim of preventing mild cognitive impairment, an intermediate stage between normal cognition and dementia that is critical to reducing the burden of dementia and maintaining healthy brain aging. This talk is motivated by large aging cohorts, such as the Canadian Longitudinal Study on Aging, that measure multiple cognitive outcomes, lifestyle factors and clinical risk factors. I will introduce two Bayesian causal approaches for this setting that can be used to study questions such as how modifiable risk factors relate to the development of mild cognitive impairment. The first approach is a longitudinal Bayesian framework for estimating causal dose–response relationships with repeated outcomes and a time-varying continuous exposure. The second approach is an ongoing Bayesian causal latent class approach with the final goal of causal trajectory modelling for multivariate longitudinal cognitive measures. Access to longitudinal cohort data for these projects is underway, and the performance of both approaches will be demonstrated using simulation studies mimicking health cohort data.