MARKOV CHAIN MODELING OF HIV, TUBERCULOSIS AND HEPATITIS-B TRANSMISSION: A STUDY OF A REGIONAL HOSPITAL IN GHANA

ABSTRACT

This study demonstrated the application of Markov S-I-R model and other statistical methods

in exploring HIV, Tuberculosis (TB) and Hepatitis B (HB)disease outcomes using Ghanaian

data. Secondary datasets from cohort studies were collected from records from the Ashanti Regional

Hospital. Relevant disease metrics as well as transition probabilities were generalized

for each disease. Method of Competing risks was used to further estimate the crude, partial

crude and net probabilities of death across age groups; whereas, the conditional relationship

among the three diseases was also established using Bayesian networks. In addition, some

significant demographic characteristics of individuals on the prevalence of these diseases were

determined using Classification tree. The Markov Chain S-I-R model revealed that Hepatitis B

(HB) was more infectious over time than Tuberculosis (TB) and HIV within the study population;

although the probability of first infection of these diseases were relatively low. However,

individuals infected with HIV comparatively had lower life expectancies than those infected

with TB and HB. The Competing risk method revealed that individuals between the ages of 20

and 50 years had a greater chance of dying from these diseases on the average. In addition,

TB was found to be very prevalent among HIV infected individuals as opposed to Hepatitis

B from the fitted Bayesian network. It was deduced from the Classification tree that females

within the study population were likely to contract HIV as opposed to males; whereas, males

were rather prone to contracting TB. Also, sex and age of patients were found to contribute

significantly to the prevalence of HIV and TB as compared to marital status and educational

level. But, none of the demographic characteristics influenced Hepatitis B prevalence. Future

studies should expand the application of Markov modeling to disease dynamics in Ghana by

considering several major hospitals in the country.