Back

Facility-Level Structural Drivers of HIV Treatment Outcomes: A Multi-Level Analysis of 27,288 Patients from a Nigerian HIV Programme and Implications for PEPFAR and Global Fund Programming

Chinthala, L. K.

2026-05-19 health systems and quality improvement
10.64898/2026.05.15.26353326 medRxiv
Show abstract

Objective: To identify facility-level characteristics - including care level, ownership type, and funding model - associated with poor HIV treatment outcomes, and quantify their independent contributions after adjustment for patient-level clinical factors. Design: Retrospective cross-sectional analysis using multivariable logistic regression with HC3 cluster-robust standard errors to account for facility-level clustering. Setting: HIV care facilities in the Nigerian national HIV programme, spanning primary health centres, secondary health facilities, and tertiary hospitals. Participants: 27,288 HIV-positive patients enrolled on ART, from a publicly available de-identified Quality of Care dataset. Main outcome measures: Composite poor outcome (poor ART adherence, treatment interruption, or mortality); individual outcomes including poor adherence rate, mortality, ART interruption, and diagnosis-to-ART delay exceeding 90 days. Results: Primary health centres had 15.4% composite poor outcome versus 10.2% at tertiary hospitals. After adjustment for patient age, sex, WHO stage, and CD4 count, primary health centre patients had 95% higher odds of poor outcome (OR=1.95; 95%CI 1.45-2.61; p<0.001). NGO-funded facilities had 24% higher odds (OR=1.24; 95%CI 1.10-1.39; p<0.001) and federally funded facilities 25% higher odds (OR=1.25; 95%CI 1.06-1.48; p=0.008). Female sex was independently protective (OR=0.87; 95%CI 0.79-0.96; p=0.003). Diagnosis-to-ART delays exceeded 90 days in 47.3% of patients, with significant variation by facility level (chi-squared=49.4, p<0.001). Conclusions: Facility level and funding model independently predict HIV treatment outcomes after patient-level adjustment. Primary health centres and NGO/federally funded facilities may require targeted quality improvement support. These findings have direct implications for PEPFAR, the Global Fund, and national HIV programme managers.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.