Predictors of HIV-1 RNA Viral Suppression in ART Treated PWLH in sub-Saharan Africa

Abstract Background: HIV remains a significant health issue, especially in sub-Saharan Africa. There are 39 million PLWH globally. Treatment with ART improves patient outcomes by suppressing the HIV RNA viral load. However, not all patients treated with ART suppress HIV RNA viral load. This research paper explores the potential predictors of viral suppression in ART-treated PLWH. Viral suppression was defined as having less than 1,000 copies/ml during the study period. Methods: We used retrospective data from the 4,713 ART-treated PLWH. The participants were enrolled through population-based surveys conducted in Zambia and Malawi. Ensemble machine learning (ML) methods that combine multiple ML algorithms were applied to identify predictors of viral suppression. Ensemble methods are used to reduce bias and variance and improve accuracy, thereby offering superior predictions compared to individual ML algorithms. Results: The age of the participants ranged from 15 to 64 years, with a majority being females. In Malawi, the bagging, boosting, and stacking models each achieved 94% accuracy. In Zambia, the accuracies were 90% (bagging), 88% (boosting), and 91% (stacking). Feature importance plots were used to identify the important features associated with VL suppression. Our findings indicated that ART status was strongly associated with higher odds of viral suppression: Malawi [OR= 106.69,95% CI: 52.82–215.4, p p p Conclusions: Our findings indicate that ART status, age, CD4 T-cell count, early initiation of ART, and ART adherence were significant predictors of viral suppression. Besides ART treatment, HIV prevention programs in the region should integrate education on early ART initiation and adherence in PLWH.