Increasing multimonth dispensing of antiretrovirals and assessing the effect on viral load suppression among children and adolescents receiving HIV services in Nigeria.

INTRODUCTION: Multimonth dispensing (MMD) enables less frequent clinic visits and improved outcomes for people living with HIV, but few children and adolescents living with HIV (CALHIV) are on MMD. At the end of the October-December 2019 quarter, only 23% of CALHIV receiving antiretroviral therapy (ART) through SIDHAS project sites in Akwa Ibom and Cross River states, Nigeria, were receiving MMD. In March 2020, during COVID-19, the government expanded MMD eligibility to include children and recommended rapid implementation to minimize clinic visits. SIDHAS provided technical assistance to 36 "high-volume" facilities-≥5 CALHIV on treatment-in Akwa Ibom and Cross River to increase MMD and viral load suppression (VLS) among CALHIV, toward PEPFAR's 80% benchmark for people currently on ART. We present change in MMD, viral load (VL) testing coverage, VLS, optimized regimen coverage, and community-based ART group enrollment among CALHIV from the October-December 2019 quarter (baseline) to January-March 2021 (endline) based on retrospective analysis of routinely collected program data. MATERIALS AND METHODS: We compared MMD coverage (primary objective), and optimized regimen coverage, community-based ART group enrollment, VL testing coverage, and VLS (secondary objectives), among CALHIV 18 years and younger pre-/post-intervention (baseline/endline) at the 36 facilities. We excluded children younger than two years, who are not recommended for or routinely offered MMD. The extracted data included age, sex, ART regimen, months of ART dispensed at last refill, most recent VL test results, and community ART group enrollment. Data on MMD-three or more months of ARVs dispensed at one time-were disaggregated into three to five months (3-5-MMD) vs. six or more months (6-MMD). VLS was defined as ≤1,000 copies. We documented MMD coverage by site, optimized regimen, and VL testing and suppression. Using descriptive statistics, we summarized the characteristics of CALHIV on MMD and non-MMD, number of CALHIV on optimized regimens, and proportion enrolled in differentiated service delivery models and community-based ART refill groups. For the intervention, SIDHAS technical assistance was data driven: weekly data analysis/review, site-prioritization scoring, provider mentoring, line listing eligible CALHIV, pediatric regimen calculator, child-optimized regimen transitioning, and community ART models. RESULTS: The proportion of CALHIV ages 2-18 receiving MMD increased from 23% (620/2,647; baseline) to 88% (3,992/4,541; endline), while the proportion of sites reporting suboptimal MMD coverage among CALHIV (<80%) decreased (100% to 28%). In March 2021, 49% of CALHIV were receiving 3-5-MMD and 39% 6-MMD. In October-December 2019, 17%-28% of CALHIV were receiving MMD; by January-March 2021, 99% of those 15-18 years, 94% 10-14 years, 79% 5-9 years, and 71% 2-4 years were on MMD. VL testing coverage remained high (90%), while VLS increased (64% to 92%). The proportion on pediatric-optimized regimens increased (58% to 79%). CONCLUSIONS: MMD was feasible among CALHIV without compromising VLS. Expanded eligibility criteria, line listing eligible children, monitoring pediatric antiretroviral stock, and data use contributed to positive results. Future efforts should address low 6-MMD uptake related to stock limitations and synchronize antiretroviral refill pickup with VL sample collection.

Predicting Treatment Interruption Among People Living With HIV in Nigeria: Machine Learning Approach.

BACKGROUND: Antiretroviral therapy (ART) has transformed HIV from a fatal illness to a chronic disease. Given the high rate of treatment interruptions, HIV programs use a range of approaches to support individuals in adhering to ART and in re-engaging those who interrupt treatment. These interventions can often be time-consuming and costly, and thus providing for all may not be sustainable. OBJECTIVE: This study aims to describe our experiences developing a machine learning (ML) model to predict interruption in treatment (IIT) at 30 days among people living with HIV newly enrolled on ART in Nigeria and our integration of the model into the routine information system. In addition, we collected health workers' perceptions and use of the model's outputs for case management. METHODS: Routine program data collected from January 2005 through February 2021 was used to train and test an ML model (boosting tree and Extreme Gradient Boosting) to predict future IIT. Data were randomly sampled using an 80/20 split into training and test data sets, respectively. Model performance was estimated using sensitivity, specificity, and positive and negative predictive values. Variables considered to be highly associated with treatment interruption were preselected by a group of HIV prevention researchers, program experts, and biostatisticians for inclusion in the model. Individuals were defined as having IIT if they were provided a 30-day supply of antiretrovirals but did not return for a refill within 28 days of their scheduled follow-up visit date. Outputs from the ML model were shared weekly with health care workers at selected facilities. RESULTS: After data cleaning, complete data for 136,747 clients were used for the analysis. The percentage of IIT cases decreased from 58.6% (36,663/61,864) before 2017 to 14.2% (3690/28,046) from October 2019 through February 2021. Overall IIT was higher among clients who were sicker at enrollment. Other factors that were significantly associated with IIT included pregnancy and breastfeeding status and facility characteristics (location, service level, and service type). Several models were initially developed; the selected model had a sensitivity of 81%, specificity of 88%, positive predictive value of 83%, and negative predictive value of 87%, and was successfully integrated into the national electronic medical records database. During field-testing, the majority of users reported that an IIT prediction tool could lead to proactive steps for preventing IIT and improving patient outcomes. CONCLUSIONS: High-performing ML models to identify patients with HIV at risk of IIT can be developed using routinely collected service delivery data and integrated into routine health management information systems. Machine learning can improve the targeting of interventions through differentiated models of care before patients interrupt treatment, resulting in increased cost-effectiveness and improved patient outcomes.