Bayesian Mixed Hidden Markov Models (BMHMM): A Multilevel Approach to Modeling Childhood Asthma

• Activity Number: 257
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Yue Zhang*+ and Kiros Berhane
• Companies: University of Southern California and University of Southern California
• Keywords: Hidden Markov Model ; Multi-Level Model ; Misclassification ; MCMC ; Bayesian Inference ; Asthma

Abstract

We propose a Bayesian Mixed Hidden Markov Model (BMHMM) in a multi-level setting. The Hidden Markov Model (HMM) is a natural approach for handling data with misclassification. The BMHMM generalizes the traditional HMM by introducing random effects into three sets of HMM parameters for joint estimation of prevalence, transition and emission (misclassification) probabilities, allowing for cluster level heterogeneity based on a multi-level model structure. We propose an MCMC algorithm for posterior computation and discuss model identifiability issues. This method is applied to data from the Southern California Children's Health Study (CHS), where questionnaire based information on asthma in children may be observed with misclassification, and treats observed asthma as outcome conditional on the true asthma state. Risk factors for both asthma transition and misclassification are presented.