Design of an AI model for predicting antiretroviral resistance in HIV patients: a case study in cameroon

Abstract Antiretroviral (ARV) resistance poses a major challenge for HIV patient management in Cameroon, with resistance prevalence reaching 15% in treatment-naive patients and up to 93.3% among those experiencing therapeutic failure. Resistance detection currently relies on expensive genotypic tests (approximately 165 USD per test – very expensive compared to the cost of living in Cameroon), which are largely inaccessible and performed only at the transition from second- to third-line therapy. Of the 448,911 patients on ARV treatment, approximately 4,500 require annual testing; however, national testing capacity is limited to only 1,000–1,500 tests per year, creating a coverage deficit of 66–78%. This study presents the design and implementation of an artificial intelligence (AI) model to predict ARV resistance in people living with HIV in Cameroon, using routine clinical data and machine learning algorithms. Data from two local sources—the Centre International de Référence Chantal Biya (CIRCB, n = 1,420 patients) and the Étoug-Ébé Electronic Medical Record (EMR, n = 4,982 patients)—were processed and used to train nine machine learning algorithms. XGBoost achieved the best overall performance on the held-out test set ( n = 272): AUC = 0.763 (95% CI: 0.678–0.849), precision 96.3%, recall (sensitivity) 73.6%, specificity 68.2% (95% CI: 47.3–83.7%), negative predictive value 18.5% (95% CI: 11.6–28.3%), and F1-score 0.834. A functional pilot interface (Streamlit) was developed for real-time clinical decision support. Integration into the national EMR system is proposed as a future step to improve therapeutic management and reduce treatment failures. So, AI-based prediction of ARV resistance using routine clinical data is a feasible and cost-effective approach in resource-limited settings. This model could support earlier treatment decisions, reduce reliance on expensive genotypic testing, and improve HIV patient management in Cameroon.