A Continuous-time Markov Chain Approach to Data Analysis on Longitudinal Categorical Outcome

• Activity Number: 400
• Type: Contributed
• Date/Time: Wednesday, August 10, 2005 : 10:30 AM to 12:20 PM
• Sponsor: ENAR
• Author(s): Wenyaw Chan*+ and Yen-Peng Li
• Companies: The University of Texas Health Science Center at Houston and The University of Texas Health Science Center at Houston
• Keywords: Longitudinal outcome ; categorical data ; continuoustime Markov chains

Abstract

Analysis of categorical outcome in a longitudinal study has been an important statistical issue. Continuous outcome in a similar study design is commonly handled by the mixed effects model. The longitudinal binary outcome analysis often is handled by the generalized estimation equation (GEE) method. Neither can properly answer the question brought up by a longitudinal study with the outcome having more than two categories, especially when the focus is on the rate of moving from one category to another. In this research, the longitudinal model that has three categories in the outcome variable will be examined. A continuous-time Markov chain model will be used to perform the analysis. This model permits an unbalanced number of measurements and an uneven duration between two consecutive measurements. Using maximum likelihood approach, we can estimate the transition rates at the unobserved time of transition. These rates are assumed to be a function of the linear combination of independent variables. For group comparison, the odds ratios and their confidence intervals can be calculated. Empirical studies were performed to compare the proposed method with other available methods.