Explainable Machine Learning for Predicting Treatment Outcomes in Female Genital Schistosomiasis and HIV Co-morbidity: A Narrative Review and Translational Framework

Abstract

Background: Female genital schistosomiasis (FGS) remains a neglected tropical disease with profound implications for women's health, particularly in sub-Saharan Africa where it is highly co-endemic with HIV. The World Health Organization estimates that over 56 million women are at risk of FGS, with prevalence reaching up to 75% in high-burden lake regions. Epidemiological evidence demonstrates that FGS increases susceptibility to HIV acquisition through genital lesions, mucosal immune changes, and chronic inflammation, thereby amplifying the dual burden of morbidity. Despite large-scale praziquantel mass drug administration, reinfection is common and integration with HIV prevention and care remains limited. Aim: This review explores the potential of explainable machine learning (ML) approaches to predict treatment outcomes and reinfection risk in women living with FGS–HIV co-morbidity, offering a framework for precision public health interventions. Methods: A narrative literature synthesis was undertaken across databases including PubMed, Scopus, and Web of Science, focusing on FGS epidemiology, treatment outcomes, HIV risk, and applications of ML in infectious disease modeling. Evidence was critically appraised and used to develop a conceptual ML framework. Results: Current evidence highlights significant gaps in the integration of FGS and HIV care, limited application of ML to FGS, and persistent reinfection despite preventive chemotherapy. However, emerging ML studies in malaria, tuberculosis, trachoma, and HIV prediction illustrate the feasibility of interpretable models for infectious diseases. Conclusion: Explainable ML has the potential to enable patient-level risk prediction, strengthen clinical decision-making, and inform integrated FGS–HIV strategies, ultimately advancing precision interventions in high-burden settings.