A MARKOV MODEL FOR SEQUENCES OF ORDINAL DATA FROM A RELAPSING-REMITTING DISEASE

Many chronic diseases follow a course with multiple relapses into peri ods with severe symptoms alternating with periods of remission; experi mental allergic encephalomyelitis, the animal model for multiple scler osis, is an example of such a disease. A finite Markov chain is propos ed as a model for analyzing sequences of ordinal data from a relapsing -remitting disease. The proposed model is one in which the state space is expanded to include information about the relapsing-remitting stat us as well as the ordinal severity score, and a reparameterization is suggested that reduces the number of parameters needed to be estimated . The Markov model allows for a wide range of relapsing-remitting beha vior, provides an understanding of the stochastic nature of the diseas e process, and allows for efficient estimation of important characteri stics of the disease course (such as mean first passage times, occupat ion times, and steady-state probabilities). These methods are applied to data from a study of the effect of a treatment (transforming growth factor-beta(1)) on experimental allergic encephalomyelitis.