Estimating the survival times of HIV patients using a markov chain

• Main Authors: Crews, Trenton, Khamis, Imad, Daoud, Jamal Ibrahim
• Format: Article
• Language: English
• Published: INSI Publications 2011
• Subjects: HA Statistics
• Online Access: http://irep.iium.edu.my/3730/1/Jamal-10.pdf; http://irep.iium.edu.my/3730/; http://www.insipub.com/online.html

In this paper, real-world data are used to estimate CDC (Center for Disease Control) stage transition times for patients diagnosed with HIV using statistical methods. The method that will be used for these estimations will be an analysis of a finite state Markov chain. For the purposes of this thesis, two populations will be analyzed. The two populations are men and women. A one-step transition matrix will be derived for both of the populations. Using this one-step transition matrix, a steady-state probability vector will also be derived. The data and the corresponding results will then be used to make a comparison between the two populations. The data were collected by observing individuals infected with HIV. Individuals who were active in the study were observed every six months. During these observations, each individual’s CDC stage was recorded. Involvement in the study was voluntary resulting in many individuals only being observed once which meant that no data concerning their transition from one CDC stage to another could be acquired. Even with the large number of individuals who discontinued their participation in the study, there were still a large number of individuals who maintained their participation throughout the duration of the study and ensured that the variance and error estimates of the results would be negligible.