Predictive Analytics Using Machine Learning to Identify ART Clients at Health System Level at Greatest Risk of Treatment Interruption in Mozambique and Nigeria.

BACKGROUND: A core objective of HIV/AIDS programming is keeping clients on treatment to improve their health outcomes and to limit spread. Machine learning and artificial intelligence can combine client, temporal, and locational attributes to identify which clients are at greatest risk of loss to follow-up (LTFU) and enable health providers to direct support interventions accordingly. SETTING: The analysis was part of a project funded by U.S. President's Emergency Plan for AIDS Relief and United States Agency for International Development, Data for Implementation, and applied to data from publicly available sources (health facility data, geospatial data, and satellite imagery) and de-identified electronic medical record data on antiretroviral therapy clients in Nigeria and Mozambique. METHODS: The project applied binary classification techniques using temporal cross-validation to predict the risk that patients would be LTFU. Classifiers included logistic regression, neural networks, and tree-based models. RESULTS: Models showed strong predictive power in both settings. In Mozambique, the best-performing model, a Random Forest, achieved an area under the precision-recall curve of 0.65 compared against an underlying LTFU rate of 23%. In Nigeria, the best-performing model, a boosted tree, achieved an area under the precision-recall curve of 0.52 compared against an underlying LTFU rate of 27%. CONCLUSIONS: Machine-learned models outperformed current classification techniques and showed potential to better direct health worker resources toward patients at greatest risk of LTFU. Moreover, models performed equally across sex and age groups, supporting the model's generalizability and wider application.

Implementation of differentiated service delivery for paediatric HIV care and treatment: opportunities, challenges and experience from seven sub-Saharan African countries.

Differentiated service delivery (DSD) models for HIV often exclude children and adolescents. Given that children and adolescents have lower rates of HIV diagnosis, treatment and viral load suppression, there is a need to use DSD to meet the needs of children and adolescents living with HIV. This commentary reviews the concept of DSD, examines the application of DSD to the care of children and adolescents living with HIV, and describes national guidance on use of DSD for children and adolescents and implementation of DSD for HIV care and treatment in children and adolescents in Elizabeth Glaser Pediatric AIDS Foundation (EGPAF)-supported programmes in seven sub-Saharan countries between 2017 and 2019. Programme descriptions include eligibility criteria, location and frequency of care delivery, healthcare cadre delivering the care, as well as the number of EGPAF-supported facilities supporting each type of DSD model. A range of DSD models were identified. While facility-based models predominate, several countries support community-based models. Despite significant uptake of various DSD models for children and adolescents, there was variable coverage within countries and variability in age criteria for each model. While the recent uptake of DSD models for children and adolescents suggests feasibility, more can be done to optimise and extend the use of DSD models for children and adolescents living with HIV. Barriers to further DSD uptake are described and solutions proposed. DSD models for children and adolescents are a critical tool that can be optimised to improve the quality of HIV care and outcomes for children and adolescents.