Machine Learning in HIV Care and Antiretroviral Therapy: Systematic Review

Area of Science:

• Medical Informatics
• Computational Biology
• Public Health

Background:

• Artificial intelligence (AI) and machine learning (ML) are increasingly integrated into medical fields.
• ML applications aim to improve patient management in diagnosis, prevention, and therapeutic care.

Purpose of the Study:

• To provide an overview of ML applications in HIV care using real clinical data.
• To identify areas for further exploration in AI-driven HIV management.

Main Methods:

• Systematic review following PRISMA 2020 guidelines.
• Analysis of four databases (PubMed, Embase, IEEE, Web of Science) until August 31, 2024.
• Keywords: Machine Learning, HIV, Antiretroviral Therapy. Excluded studies not focused on HIV, pre-exposure prophylaxis, in silico drug development, or biological mechanisms.

Main Results:

• 98 studies analyzed from 476 identified, detailing 6 major ML application categories in HIV care.
• Random forests were the most used algorithm (17.49%), effective in identifying biomarkers and predicting impairment.
• Support vector machines and logistic regression showed efficacy in predicting drug resistance, treatment outcomes, and adverse events.

Conclusions:

• ML methods vary in suitability for specific HIV concerns.
• Areas like treatment recommendations and adherence require further AI exploration.
• AI is a promising tool for developing clinical decision-support systems for people living with HIV.