|
Abstract:
|
Large amounts of longitudinal health records are now collected in private and public healthcare systems. Data from sources such as electronic health records, healthcare databases and mobile health records are available to inform clinical and public health decision-making. In many situations, such data enable the dynamic monitoring of the underlying disease process that governs the observations. However, this process is not observed directly, and so inferential methods are needed to ascertain progression. Multi-state models capture the status of individuals longitudinally as a discrete-time realization of a continuous-time Markov process. We construct a continuous-time hidden Markov model (CTHMM) with inference based on Markov chain Monte Carlo (MCMC), which provides a fully Bayesian analysis that yields much richer inference than pure-likelihood methods. In addition, we relax the assumption that the number of health states is known by implementing trans-dimensional MCMC that can explore a model space where it is allowed to vary under a CTHMM. Finally, we apply the proposed method to a large COPD dataset from a Canadian healthcare system in Quebec.
|
