10th U6+ Consortium of African Universities International Conference

Malaria and sickle cell anemia (SCA) impose a compounded disease burden across sub-Saharan Africa, where overlapping clinical presentations, fever, anemia, and jaundice frequently obscure accurate diagnosis in resource-constrained settings. While artificial intelligence has demonstrated considerable promise in single-disease detection, existing diagnostic models operate in isolation, lack multi-class differentiation capabilities, and offer limited interpretability barriers that impede clinical adoption in co-endemic regions. This review synthesizes 87 peer-reviewed studies published between 2017 and 2025, examining ensemble learning and explainable AI (XAI) applications in malaria and SCA diagnostics across image-based and electronic health record modalities. Our analysis reveals three persistent gaps: the absence of multi-class frameworks capable of distinguishing malaria, SCA, co-infection, and other febrile illnesses; the under-utilization of stacking ensembles for integrating heterogeneous clinical and imaging data; and the neglect of clinician-centered XAI evaluation in low-resource settings. We propose a conceptual framework that combines stacking ensembles with SHAP-based explainability to enable hierarchical interpretation of diagnostic decisions. The findings indicate that while explainable ensemble approaches are methodologically mature in other healthcare domains, their application to malaria-SCA co-diagnosis remains unexplored representing both a critical translational gap and a significant opportunity for equity-focused AI research in East African health systems. This review contributes a research roadmap for developing context-aware, interpretable, and integrated diagnostic tools tailored to the epidemiological realities of co-endemic regions.