Objectives: South Africa has made remarkable progress in increasing the coverage of antiretroviral therapy (ART) among pregnant women; however, viral suppression among pregnant women receiving ART is reported to be low. Access to routine viral load testing is crucial to identify women with unsuppressed viral load early in pregnancy and to provide timely intervention to improve viral suppression. This study aimed to determine the coverage of maternal viral load monitoring nationally, focusing on viral load testing, documentation of viral load test results, and viral suppression (viral load 4000) primary healthcare facilities in South Africa provide antenatal care services. Prevention of vertical HIV transmission services, including viral load testing, are provided in all antenatal care facilities. At the time of this survey (October to November 2019), the South African viral load monitoring guidelines recommended that all women on ART should have viral load testing at their first antenatal care visit or upon confirmation of pregnancy [11]. For newly diagnosed women initiating ART during pregnancy, viral load testing was recommended at 3 months after ART initiation. For all HIV‐positive pregnant women, repeat viral load testing was recommended at delivery and 6 months after delivery. For women who were not virally suppressed, depending on their viral load level, viral load testing was to be repeated in 8–10 weeks (for viral load 50–999 copies/mL) or 4–6 weeks (for viral load ≥ 1000 copies/mL) after the initial unsuppressed viral load result. Although viral load testing services have been available in South Africa since 2004 [7], the coverage of viral load testing increased in recent years following WHO’s endorsement of viral load monitoring as the primary method for monitoring response to ART [6]. At the time of this survey, viral load testing for public health facilities in South Africa was provided at 17 laboratories using both the Roche and Abbott assays [12]. Results are returned via hard copy, short message system (SMS) printers or accessed through web portals [13]. At the time of this study, the first‐line regimen for women initiating ART during pregnancy was a non‐nucleoside reverse transcriptase (NNRTI)‐based ‐ most often efavirenz (EFV) ‐based ‐ regimen. The antenatal survey is a cross‐sectional survey conducted biennially in South Africa to primarily monitor HIV prevalence among pregnant women but also to evaluate the performance of the prevention of vertical HIV transmission programme. The survey aimed to enrol 36 015 pregnant women (regardless of HIV status) from 1589 public health facilities selected from each of the 52 districts in South Africa, biennially. For this sub‐analysis, only HIV‐positive pregnant women who either initiated ART before pregnancy or were taking ART for ≥ 3 months at enrolment were included in the analysis. Health facilities that took part in the 2019 antenatal survey were selected using a stratified probability proportional to size (PPS) sampling method from each district. The 2019 survey was conducted between 1 October and 15 November 2019. During this period, consenting pregnant women aged 15–49 years, attending the antenatal clinic for the first time or for follow‐up visits during their current pregnancy were consecutively enrolled (regardless of HIV or ART status) until either the required sample size or the end of the study was reached. Health workers providing antenatal care services in the selected facilities collected demographic and clinical data (including maternal education, relationship with the father of the child, and gravidity) through interviews. Data were extracted from medical records, which included: age of the woman, gestational age at booking, gestational age today, HIV status (per rapid test performed at the clinic at the time of first antenatal care visit, during follow‐up visit, or test done before pregnancy if participants were already on ART at the time of pregnancy), the timing of ART initiation as a categorical response (i.e. initiated before pregnancy, at the first, second or third trimester), viral load test done during pregnancy, whether viral load test result was documented, and latest viral load test result as a categorical response (i.e. 1000 copies/mL). If more than one viral load test was done during the pregnancy, only information on the most recent viral load test was extracted. On the day of the survey, blood specimens were collected from each participant, regardless of prior knowledge of HIV status, and tested for HIV at regional laboratories using two serial imunoassays (IAs) following the standard guideline [14]. Detailed descriptions of the methodology of the survey have been published previously [15, 16 The coverage of the following three viral load cascade indicators was estimated: viral load testing; viral load test result documentation; and viral suppression (defined as viral load < 50 copies/mL). This indicator measured the proportion of HIV‐positive women eligible for viral load testing who received a viral load test in their current pregnancy (if there was more than one viral load test, the most recent viral load test was assessed). Participants were considered as ‘eligible for viral load testing’, according to the national guideline, if they initiated ART before pregnancy, or if they were newly initiated on ART during pregnancy but had received ART for at least 3 months. For the latter group, as the timing of ART initiation was reported by trimester (i.e. as first, second or third trimester), duration on ART was calculated by subtracting the mid‐week of the trimester that ART was initiated from the participants’ gestational age at survey enrolment. For participants initiated on ART in the first trimester, the gestational age/week of their first antenatal care visit was considered as the time of ART initiation, as almost all participants started antenatal care after the mid‐week of the first trimester; and where gestational age at the first visit is not reported, time of ART initiation was set at 10 weeks as most women do not attend antenatal care before 10 weeks. Participants whose ART status was not reported, those who reported not initiating ART, and those whose rapid test or IA test was HIV‐negative were excluded from this analysis, regardless of their response to the viral load monitoring questions. This study also excluded participants whose reported timing of ART initiation was less than 3 months from their gestational age at enrolment in the survey. This indicator measured the percentage of participants who had undergone a viral load test with a result (from a test provided during pregnancy) that was documented on the medical record – if there was more than one viral load test, documentation of the most recent viral load test result was assessed (the denominator for this indicator was the number of HIV‐positive women with viral load test done). This indicator measured the percentage of participants with a documented viral load test result of < 50 copies/mL (using participants with documented viral load test result as a denominator). In the case where participants had more than one viral load test result during pregnancy, the most recent result was extracted from the medical record. The three viral load cascade indicators were analysed at the national level and in stratified groups. χ2 test was used to assess significant associations. A sensitivity analysis was conducted to validate the viral load data extracted from the medical record using laboratory data (these data are presented under Supplementary Box 1). We fitted a multivariable survey logistic regression model to assess factors associated with not receiving a viral load test. All variables significant at a P‐value cut‐off point of 0.2 in a bivariable analysis were included in multivariable analysis. Variables significant at a P‐value cut‐off point of 0.05 and other variables that have ≥ 10% effect on other significant variables were kept in the final model. Adjusted odds ratios (AORs) and 95% confidence intervals (CIs) were reported. All analyses took into account the survey design (i.e. all analyses adjusted for the different sampling stages: stratification and clustering within primary sampling units, and for the finite number of primary sampling units). All analyses were also weighted for sample size realization and the Statistics South Africa 2019 midyear population size of women of reproductive age (15–49 years) to adjust for differential population size across provinces and for different sample size achievement at district level [17] Participation in the survey was voluntary, requiring written informed consent. Ethical approval was obtained from the University of the Witwatersrand Human Research Ethics Committee (Medical) (ethics clearance number: M170556), and the nine provincial health research ethics committees. The study protocol was reviewed following the Centers for Disease Control and Prevention, United States (CDC‐US) human research protection procedures.