Applying Machine Learning for the Prediction of Tuberculosis Among People Living with HIV from Health Facilities in Addis Ababa, Ethiopia

<title>Abstract</title> <p>Background Tuberculosis (TB) is a major global public health problem and one of the leading causes of death worldwide, particularly in the era of Human Immunodeficiency Virus (HIV). TB is among the leading causes of death for people living with HIV (PLWHIV). Early diagnosis of TB among PLWHIV remains challenging due to non-specific clinical symptoms, coexisting opportunistic infections, and limitations of conventional diagnostic and screening approaches. The objective of this study was to develop and validate a machine learning model to predict TB cases among PLWHIV using smart care data from health facilities. Methods The research employs quantitative analysis of secondary data from four health facilities that provided HIV care and treatment to PLWHIV and reported to the Addis Ababa Health Bureau. The study uses an experimental research design applying machine learning and deep learning models. Results From four facilities, 16,698 adult patients currently on ART were included; 59% were female, and 85% were married. In this study, the majority of the models performed better when the Hybrid method combined SMOTE with Edited Nearest Neighbors (ENN), and XGBoost had the strongest influence on the predictor of TB occurrence across most evaluation matrices. The XGBoost accuracy was 98%, recall 80%, precision 25%, F1 score 38%, F Beta 30%, MCC 52%, and AUC 96%. ART regimen, taking Tuberculosis Preventive Therapy (TPT), cotrimoxazole prophylaxis, months on ART, patient weight, and functional status are the strongest risk factors for developing TB and for protection against TB development. Conclusion Overall, XGBoost outperforms the other models across accuracy, precision, F1 score, and ROC AUC. Using the XGBoost model, the ART regimen, TPT, cotrimoxazole prophylaxis, months on ART, and patient weight were the strongest predictors of whether a patient would be protected against TB or would develop TB.</p>