Utility of a Machine-Guided Tool for Assessing Risk Behavior Associated With Contracting HIV in Three Sites in South...

Area of Science:

• Digital health
• Machine learning applications in public health
• Epidemiology of infectious diseases

Background:

• Mobile technology facilitates health programs, with automated risk prediction models showing success in identifying individuals at high risk of HIV.
• Machine-guided tools utilize algorithms to compute HIV risk scores based on questionnaire responses.

Purpose of the Study:

• To demonstrate machine learning's capability in developing machine-guided tools for HIV risk assessment.
• To statistically analyze the correlation between specific behavioral patterns and HIV risk.

Main Methods:

• Recruitment of 200 HIV-negative adults across diverse South African settings.
• Participants completed a questionnaire via an application, underwent two HIV tests, and provided user experience feedback.
• Risk assessment scores were compared against HIV test results to evaluate the tool's prediction accuracy and reliability.

Main Results:

• Ethical approval obtained; data collection is ongoing with expected completion in late 2021.
• The study will report on the machine-guided tool's performance, usability, user satisfaction, and areas for enhancement.
• Analysis will include reliability calculations based on risk assessment scores and ground truth from HIV testing.

Conclusions:

• Machine-guided risk assessment tools offer a cost-effective approach to HIV screening.
• These tools can enhance the efficiency of targeted counseling and testing interventions.
• The development of such tools is crucial for preventing the spread of HIV.