Assessing the effectiveness of the ART programme in the Western Cape Province of South Africa through triangulation of context-appropriate population level routine monitoring and surveillance systems

Osler, Margaret

Assessing the effectiveness of the ART programme in the Western Cape Province of South Africa through triangulation of context-appropriate population level routine monitoring and surveillance systems

Thesis / Dissertation

2023

Abstract

Background After a decade of free antiretroviral therapy (ART) provision, countries continued facing challenges both in trying to meet the ever-increasing pool of eligible people needing HIV treatment, and efficiently monitoring programme effectiveness to improve patient care and service delivery. Concerns about the feasibility of further treatment expansion were being debated with trials showing benefit, but with ongoing uncertainty as to whether those benefits would be realised in resource-limited settings with fragile health systems. Key questions underpinning this thesis were how to robustly develop, implement, monitor and use routine health information systems to explore pertinent epidemiological questions, including real-world effectiveness of the ART programme, determinants of ongoing morbidity and mortality, and the impact of guideline changes. Methods The thesis includes a health systems review of the implementation of person-level information systems for HIV care, followed by a number of cohort analyses based on the public-sector health services in the Western Cape, South Africa. The study population consists of people living with HIV, who had at least one CD4 test or HIV-care visit, and who were ≥16 years of age. The cohort analyses utilized a population-wide linked dataset containing all available digital data from fixed health facilities, laboratory, pharmacy, and death registry systems. The first analysis described temporal trends in the CD4 distributions over 10 years, with longitudinal categorization of ART status of people with extremely advanced HIV disease (AHD). Two analyses used a regression discontinuity design to consider the causal impact of guideline changes, while the last two analyses explored important longer-duration determinants of morbidity and mortality in a survival analysis cohort framework, including through the use of flexible parametric survival models. Results Developing a tiered suite of interoperable information solutions enabled each health facility to independently evolve from paper to offline and then hybrid/online electronic registers when dependencies such as electricity, stable networks and resources allowed them to. The largest proportion of people with severe AHD (CD4 <50 cells/µL) came from those already on ART in more recent years, in comparison to people first presenting or not eligible for ART. Of those on ART with severe AHD, more than three-quarters had a confirmed treatment interruption (>3 months) and/or viraemia within the previous year. The biggest benefits (based on24-month survival) when increasing eligibility thresholds for ART were seen during earlier guideline changes which expanded access at lower CD4 count thresholds (from CD4 <200/µL to CD4 <350/µL); however, at a patient level, benefits from ART were seen at all three eligibility threshold changes (200, 350 and 500 cells/µl). Deferring treatment for people ineligible lead to, on average, >2 years delay prior to starting ART, increasing risk of AHD and death. The greatest increases in ART initiations and decreases in mortality happened between guideline changes, reflecting large increases in ART access prior to formal expansion of access through guideline changes. As the ART programme has matured, men living with HIV continue to have poorer access to ART, a greater risk of TB, are more likely to interrupt treatment and have inferior clinical outcomes compared to woman, especially between diagnosis and the first five years on ART. Women, however, carried the larger absolute mortality burden, due to the greater numbers living with HIV. Almost two-thirds of the ART cohort interrupted treatment for >4 months at least once, increasing their risk of death by 27%, compared to people who had never interrupted ART. Each additional interruption was associated with further increases in mortality. Conclusions Pragmatic interoperable offline/hybrid/online health information systems can be successfully implemented at scale in lower resource settings to improve patient care, provide information on interventions and inform policy and resource allocation. Programmatic ART outcomes did improve during expansion of ART eligibility including into the time-period of the treat-all policy when CD4 count criteria were removed. More people accessed ART over time, independent of guideline changes, improving population HIV outcomes. The guidelines changes were nonetheless shown to be effective at an individual level. The number of people with AHD has not decreased however due to ART experienced patients returning to care after interruptions with considerable immune deterioration. Recommendations focused on improving systems for retention, re-engagement, and AHD and are most likely applicable to similar public-sector settings in Southern Africa.

Keywords

Public Health

Reference:

Collections

PhD / Doctoral

Full item page