Semi-Markov and non-homogeneous Markov models in medicine
Longitudinal studies form the most interesting part of medical studies. All therapeutical trials and the majority of epidemiological studies can be considered as longitudinal studies but most often this aspect is ignored or grossly simplified when the data are analysed. For instance a very simple therapeutical trial can be described as follows. Patients are included in the trial, each at a certain moment of their history. When included they are assigned randomly to one of two treatment groups. A particular variable R, called here the response variable, can be observed till the end of the study. A simple method of analysis consists in comparing the values of R in the two groups observed after a time Te of treatment. A Pearson’s X2 or a Student’s t test for instance can be used according to the statistical nature of the variable.
KeywordsSojourn Time Markov Chain Model Hyperthyroidic Patient Homogeneous Markov Chain Marginal Total
Unable to display preview. Download preview PDF.
- Commenges, D., P. Barberger-Gateau, J.F. Dartigues, P. Loiseau and R. Salamon (1984). A non-homogeneous Markov chain model for follow-up studies with application to epilepsy. Meth.Inform. Med. 23, 109–114.Google Scholar
- Iosifescu, M. and P.Tautu (1973). Stochastic processes and applications in biology and medicine. Springer-Verlag.Google Scholar
- Kay, R. (1984). Multistate survival analysis: an application to breast cancer. Meth.Infor.Med. 23, 157–162.Google Scholar
- Mantel, N. (1966). Evaluation of survival data and two rank-order statistics arising in its consideration. Cancer Chemotherapy Report 50, 163–170.Google Scholar
- Mantel, N. and W. Haenszel (1959). Statistical aspects of the analysis of data from retrospective studies of disease. J.Nat.Cancer Inst. 22, 719–740.Google Scholar