B1130
Title: Multistate Markov models: Application to dementia progression
Authors: Jonathan Williams - North Carolina State University (United States) [presenting]
Abstract: Multistate Markov models are a canonical parametric approach for data modeling to draw inferences on the role of ageing in the development of dementia. Two fundamental obstacles to such approaches are described, and tools for remediation are provided. The first is a delayed enrollment bias likely to ensue in prospective studies where some or all subjects are not observed at baseline. The second is the unbiased estimation of a time-inhomogeneous infinitesimal generator matrix. Continuous-time Markov processes describe data observed irregularly over time, as is often the case in longitudinal medical and biological data sets, for example. Assuming that a continuous-time Markov process is time-homogeneous, a closed-form likelihood function can be derived from the Kolmogorov forward equations for a system of differential equations with a well-known matrix-exponential solution. Unfortunately, however, the forward equations do not admit an analytical solution for continuous-time, time-inhomogeneous Markov processes, and so researchers and practitioners often make the simplifying assumption that the process is piecewise time-homogeneous. Intuitions and illustrations of the potential biases for parameter estimation are provided, that may ensue in the more realistic scenario that the piecewise-homogeneous assumption is violated, and a solution for likelihood computation is advocated in a truly time-inhomogeneous fashion.