Patient losses from HIV care across HIV treatment guideline periods between 2016 and 2020 in six provinces of South Africa

By  Dr Dorina Onoya  Dr Khumbo Shumba  Nelly Jinga  Cornelius Nattey  Dr. Matthew Fox  William MacLeod  Dr. Jacob Bor  |  | 

Background: South Africa’s move to universal HIV test and treat (UTT) and same-day initiation (SDI) of antiretroviral therapy (ART) improved treatment uptake with hopes of also reducing patient losses from care.

Methods: We included all patients diagnosed with HIV at public-sector health facilities in 2015-2019 and represented on the TIER.net clinical database in participating provinces (six of nine) in South Africa. Patients were followed up from the date of HIV diagnosis to 12 months. We computed the share of patients lost to follow up (LTFU) at 12 months in HIV care, defined as not having visited the clinic between months six and 12 post-HIV diagnosis. We also assessed trends in patients’ retention time in the first 12 months on ART, stratifying by province.

Results: We analyzed data on 2,392,668 patients entering care. Figure 1 shows the time trend in LTFU patients at 12 months after diagnosis and the time spent in care during the study period. The proportion of patients becoming LTFU remained steady from 37.1 % in 2016 to 38.8% in 2020, ranging from 30.7% in Mpumalanga to 42.1% in Limpopo. The North-West province (NW) experienced an increase from 35.0% in 2016 to 42.9% in 2020. Gauteng and Limpopo provinces had LTFU rates consistently over 40% during the study period. Over 90% of patients in the TIER.Net database had started ART by 12 months, making these on-ART losses. However, time to LTFU gradually rose from a median of 30.5 days (IQR: 30.5-148.5) at the
2016 endpoint to 92 days (IQR: 7-158.0) in 2020, with some variability across provinces.

Discussion: Patient losses from care has remained high despite treatment guideline changes, but patients are remaining in care longer under the new treatment guidelines. These data call for further examination of the key determinants of this change to inform program improvements.

Publication details

PDF