Abstrakt

A Markov model to estimate mortality due to HIV/AIDS using CD4 cell counts based states and viral load: a principal component analysis approach

Delson Chikobvu, Claris Shoko

Background: Improvement of health in HIV/AIDS patients on Highly Active Antiretroviral Therapy (HAART) is characterised by an increase in CD4 cell counts and a decrease in viral load to undetectable levels. In modelling HIV/AIDS progression in patients, researchers mostly deal with either viral load only or CD4 cell counts only as they expect these two variables to be collinear.

Methods: In this study, a cohort of 320 HIV/AIDS patients under HAART follow-up from a wellness clinic in Bela-Bela, South Africa is used. A time homogeneous Markov model is developed to explain and predict probability of death from HIV/AIDS. Principal component variables are created by fitting a regression model of viral load on CD4 cell counts.

Results: Inclusion of a viral load principal component improves the efficiency of the model. The new viral load covariate helps to explain the component of mortality/transition, which could not be explained by the CD4 cell counts alone. CD4 cell counts are categorised to define the states for the Markov model. Results show that the expected percentage prevalence gives almost a perfect fit of the observed data.

Conclusion: The orthogonal viral load covariate along with CD4 baseline, gender, non-adherence to treatment and age in years (y) variables play a significant role in modelling HIV/AIDS progression based on both CD4 cell counts and viral load monitoring.