Harnessing Predictive Modeling to Advance HIV Self-Testing in SubSaharan Africa: A Viewpoint on Equity-Driven Implementation

Abstract

ABSTRACT
Predictive modeling presents a transformative opportunity to enhance HIV self-testing (HIVST) uptake across SubSaharan Africa (SSA). While machine learning techniques such as Random Forest (RF) and Classification and
Regression Trees (CART) offer powerful tools for identifying high-risk populations and optimizing HIVST
distribution, their adoption in public health remains limited. This Viewpoint examines how stigma, economic
constraints, and urban-centric data biases hinder the integration of predictive analytics into HIVST and argues for
equity-driven implementation strategies. The authors argue that leveraging predictive modeling requires an
ethical, community-driven approach that prioritizes fairness, transparency, and real-world applicability. Without
inclusive implementation strategies, predictive analytics risks reinforcing disparities rather than reducing them.This
article presents a strategic framework for integrating machine learning into HIVST policy and practice while
addressing concerns around data bias, public trust, and stakeholder engagement. By bridging the gap between
artificial intelligence (AI) and global health equity, predictive modeling can serve as a catalyst for achieving UNAIDS’
2030 goals for broad, equitable HIV testing access.