# A Multistate Continuous Time-Inhomogeneous Markov Model for Assessing the CD4 Count Dynamics of HIV/AIDS Patients Undergoing Antiretroviral Therapy in KwaZulu-Natal, South Africa

## TL;DR

This study uses a Markov model to track CD4 count changes in HIV patients in South Africa, finding that gender and location affect disease progression.

## Contribution

A novel multistate continuous time-inhomogeneous Markov model is applied to assess CD4 dynamics in HIV patients under ART.

## Key findings

- Males had a higher risk of CD4 decline compared to females.
- Rural patients showed slower CD4 recovery compared to urban patients.
- Regimen, location, gender, and age significantly influence HIV progression.

## Abstract

Monitoring CD4 count levels is essential for tracking the progression of HIV in patients. This study aimed to identify the key factors influencing HIV progression by incorporating time-varying factors and transition probabilities. The data for this study were obtained from the Centre for the AIDS Programme of Research in South Africa (CAPRISA), which enrolled 3325 patients aged 14 to 76 who initiated antiretroviral therapy (ART) and were followed up with between June 2004 and August 2013. The dataset included clinical, demographic, and treatment information to capture a comprehensive picture of HIV progression. To analyze the factors associated with HIV progression, this study employed time-inhomogeneous Markov models, which allow for incorporating covariates that change over time and transition probabilities. These models provided a robust framework to assess how various factors, such as CD4 count, viral load, and treatment adherence, evolve and influence disease progression. The results indicated that males had a significantly higher risk of moving from a normal (more than 500 cells/mm3) to mild state (351–500 cells/mm3) than females [HR: 1.614, 95% CI (1.281, 2.034)]. Rural patients had a significantly higher risk compared to urban patients of transiting from a mild state (351–500 cells/mm3) to an advanced state (200–350 cells/mm3) with a 95% confidence interval of (0.641, 1.009) [HR: 0.805, 95% CI (0.641, 1.009)]. The multistate model identified regimen, location, gender, and age as significant clinical variables influencing HIV progression. Rural patients and males showed slower transitions to CD4 count recovery. These findings provide valuable insights for disease management, treatment planning, and understanding the long-term prognosis for individuals living with HIV. Improving healthcare access, increasing educational efforts targeting men, reducing stigma, and fostering supportive environments can play a crucial role in enhancing CD4 count recovery and overall health outcomes for people living with HIV.