The impact of the COVID-19 lockdown on HIV care in 65 South African primary care clinics: an interrupted time series analysis

BACKGROUND: The effect of the COVID-19 pandemic on HIV outcomes in low-income and middle-income countries is poorly described. We aimed to measure the impact of the 2020 national COVID-19 lockdown on HIV testing and treatment in KwaZulu-Natal, South Africa, where 1·7 million people are living with HIV. METHODS: In this interrupted time series analysis, we analysed anonymised programmatic data from 65 primary care clinics in KwaZulu-Natal province, South Africa. We included data from people testing for HIV, initiating antiretroviral therapy (ART), and collecting ART at participating clinics during the study period, with no age restrictions. We used descriptive statistics to summarise demographic and clinical data, and present crude summaries of the main outcomes of numbers of HIV tests per month, ART initiations per week, and ART collection visits per week, before and after the national lockdown that began on March 27, 2020. We used Poisson segmented regression models to estimate the immediate impact of the lockdown on these outcomes, as well as post-lockdown trends. FINDINGS: Between Jan 1, 2018, and July 31, 2020, we recorded 1 315 439 HIV tests. Between Jan 1, 2018, and June 15, 2020, we recorded 71 142 ART initiations and 2 319 992 ART collection visits. We recorded a median of 41 926 HIV tests per month before lockdown (January, 2018, to March, 2020; IQR 37 838–51 069) and a median of 38 911 HIV tests per month after lockdown (April, 2020, to July, 2020; IQR 32 699–42 756). In the Poisson regression model, taking into account long-term trends, lockdown was associated with an estimated 47·6% decrease in HIV testing in April, 2020 (incidence rate ratio [IRR] 0·524, 95% CI 0·446–0·615). ART initiations decreased from a median of 571 per week before lockdown (IQR 498–678), to 375 per week after lockdown (331–399), with an estimated 46·2% decrease in the Poisson regression model in the first week of lockdown (March 30, 2020, to April 5, 2020; IRR 0·538, 0·459–0·630). There was no marked change in the number of ART collection visits (median 18 519 visits per week before lockdown [IQR 17 074–19 922] vs 17 863 visits per week after lockdown [17 509–18 995]; estimated effect in the first week of lockdown IRR 0·932, 95% CI 0·794–1·093). As restrictions eased, HIV testing and ART initiations gradually improved towards pre-lockdown levels (slope change 1·183/month, 95% CI 1·113–1·256 for HIV testing; 1·156/month, 1·085–1·230 for ART initiations). INTERPRETATION: ART provision was generally maintained during the 2020 COVID-19 lockdown, but HIV testing and ART initiations were heavily impacted. Strategies to increase testing and treatment initiation should be implemented. FUNDING: Wellcome Trust, Africa Oxford Initiative.