Geographical variation in HIV testing in South Africa: Evidence from the 2017 national household HIV survey

Background Identification of the geographical areas with low uptake of HIV testing could assist in spatial targeting of interventions to improve the uptake of HIV testing. Objectives The objective of this research study was to map the uptake of HIV testing at the district level in South Africa. Method The secondary analysis used data from the Human Sciences Research Council’s 2017 National HIV Prevalence, Incidence, Behaviour and Communication Survey, where data were collected using a multistage stratified random cluster sampling approach. Descriptive spatial methods were used to assess disparities in the proportion of those ever tested for HIV at the district level in South Africa. Results The districts with the highest overall coverage of people ever having tested for HIV (> 85%) include West Rand in Gauteng, Lejweleputswa and Thabo Mofutsanyane in Free State, and Ngaka Modiri Molema in North-West. These provinces also had the least variation in HIV testing coverage between their districts. Districts in KwaZulu-Natal had the widest variation in coverage of HIV testing. The districts with the lowest uptake of HIV testing were uMkhanyakude (54.7%) and Ugu (61.4%) in KwaZulu-Natal and Vhembe (61.0%) in Limpopo. Most districts had a higher uptake of HIV testing amongst female than male participants. Conclusion The uptake of HIV testing across various districts in South Africa seems to be unequal. Intervention programmes must improve the overall uptake of HIV testing, especially in uMkhanyakude and Ugu in KwaZulu-Natal and Vhembe in Limpopo. Interventions must also focus on enhancing uptake of HIV testing amongst male participants in most districts. Strategies that would improve the uptake of HIV testing include HIV self-testing and community HIV testing, specifically home-based testing.

http://www.sajhivmed.org.za Open Access revised 2030 targets of 95-95-95, which are set out to be achieved by 2030. 3 South Africa has made progress towards the UNAIDS 90-90-90 targets, especially regarding HIV testing and viral load suppression. 4 Over the past decade, the country had made excellent progress in involving more people to test and become aware of their HIV status, after the launch of two national HIV testing initiatives: firstly, the national HIV testing and counselling (HTC) campaign that took place in 2010, and secondly, the HTC revitalisation strategy in 2013. 5 As a result of these campaigns and other similar campaigns, more than 10 million people in South Africa test for HIV every year. 5 In scaling up efforts around HTS interventions, civil society organisations continue to work with government departments in South Africa. The South African National AIDS Council continues to provide a platform for engagement between the civil society and government to work together on the HIV response. 6 Although South Africa has made steady progress towards reaching the UNAIDS targets, many people affected with HIV are still unaware of their HIV status. 7 Despite the availability of HTS, research studies have revealed that only a fraction of South Africans who are at risk get tested for HIV. 8 Evidence shows that access to HTS may be limited geographically because of the inadequacy and heterogeneous distribution of available services. 9,10 Achieving high coverage of HIV testing is critical for linking HIV-positive people to care across the country. Therefore, equitable geographical distribution of HTS is vital for achieving optimal coverage for HIV testing. 10 This highlights the importance of conducting and collecting population-based HIV testing covereage data at the sub-national level needed for decision making.
In South Africa, gathering spatial data on HIV and mapping its distribution have been carried out in selected microgeographical areas, limiting the generalisability of the findings to the country. 11 The main source of estimating the number of people who tested for HIV in the country comes from the District Health Information System and from modelling. 12 Both sources have limitations and rely on healthcare facility and programme data from districts to produce estimates. 13,14 The current study used large-scale nationally representative population-based household survey data to describe the spatial coverage in the uptake of HIV testing amongst youth and adults 15 years and older. The aim of this research study was to identify the spatial gap in the uptake of testing in people who had ever tested for HIV at the district level in South Africa.

Study design and sampling
The data used in the secondary analysis were obtained from the National HIV Prevalence, Incidence, Behaviour and Communication Survey conducted in 2017. 15 The survey used a multistage stratified, cluster randomised, cross-sectional design. The survey chose a systematic probability sample of 15 households randomly from 1000 small area layers (SALs), selected from 84 907 SALs released by Statistics South Africa in 2015. 16 The sampling of SALs was stratified by province and locality type (urban formal, urban informal, rural formal and rural informal localities

Measures
The primary outcome measure 'ever testing for HIV' was obtained from individuals who responded to the original survey question 'have you ever been tested for HIV?' The response was dichotomised into a binary outcome (yes = 1 and no = 0).

Ethical considerations
The

Statistical analysis
Statistical analysis was carried out in STATA 15.0 (Stata Corporation, College Station, TX, United States [US]) software.
Descriptive statistics were used to summarise the sample characteristics. Multilevel mixed-effects logistic regression models were used to estimate the excess probability of prior testing for HIV after adjusting for the effect of age and sex. District-level random effects predicted from the model, including age and sex were used to estimate the excess probability of prior testing. Results are shown with 95% confidence intervals (CI), and p-values < 0.05 were reported for all statistically significant associations. The proportion of the population, aged 15 years and older, that have ever been tested for HIV were geo-located using the South African district-level boundaries. The maps were generated in QGIS, version 3.14.
10. An adjusted weight, benchmarked to the general population by age and sex at the national level, was computed to facilitate this analysis.  Figure 1 illustrates the geographical distribution of people who have ever been tested for HIV in the 52 districts of South Africa (Table 1-   Adjusted coverage of ever being tested for HIV Figure 3 illustrates the geographical coverage of the excess probability of ever having tested for HIV in the 52 districts of South Africa after adjusting for age and sex (

Discussion
HIV testing is a crucial component of the national HIV response in South Africa. 17 This study presents the first geographic analysis of youth and adults (≥ 15 years) who  have ever been tested for HIV in South Africa using simple GIS mapping and data obtained from a cross-sectional nationally representative population-based survey.
The mapping results revealed that the uptake of HIV testing varied across the various districts in South Africa. The age and sex distribution across the districts were different.   Studies have revealed that age and sex are crucial factors in HIV testing. 18,19 The estimates from a multilevel model with district-level random effects showed that excess probability of ever having tested for HIV was different among the districts after adjusting for age and sex. Variations in the quality of healthcare services, health promotion activities, easier access to healthcare facilities and socio-economic status could have an impact on the uptake of HIV testing in districts.
The overall proportion of people who had ever tested for HIV at the district level in South Africa ranged from 54.7% to 86.1%. uMkhanyakude and Ugu districts in KwaZulu-Natal and Vhembe district in Limpopo had the lowest overall testing coverage of < 62%. Ngaka Modiri Molema district in North West, and Lejwelepuswa and Thabo Mofutsanyane districts both in Free State reported the highest coverage for HIV testing. None of the districts in the Eastern Cape or Limpopo had an overall coverage of higher than 80%. These districts are characterised as being predominately rural.
Other studies have also found that people living in rural informal or tribal areas were significantly less likely to test for HIV when compared with those from urban areas. 20,21 The finding that uptake of HIV testing was less likely amongst those in rural areas could be linked to limited resources and structural barriers to healthcare in terms of geographical and financial accessibility. 22,23 Additional barriers included fear, discrimination and stigmatising attitudes, as well as lack of education and awareness. 24 Another factor playing a major role in the higher coverage districts included the epidemic control plans implemented by the President's Emergency Plan for AIDS Relief (PEPFAR), which aims to achieve maximum impact and reach in areas with the highest burden of disease (COP19). This is informed by population-based surveillance. The PEPFAR country operational plan (COP) for 2017, in 27 districts with an estimated number of people living with HIV of 5.6 million, which account for 79% of number of people living with HIV in South Africa (COP19), identified 1969 sites for intensified support as part of the country's district-level implementation plan (DIP). 25 According to the National Strategic Plan for 2012-2016, 26 6 Therefore, achieving high coverage of HIV testing amongst men is critical in the fight against HIV in the country. However, data elsewhere suggest that boys and men are lagging. 27,28 Men were found to have lower levels of participation in HIV testing. 29 Some of these reasons include fear of damaging reputations, losing their masculine pride, fearing both community rejection and a loss of emotional control because of the psychological burden of knowing one was HIV positive. 29 HIV testing programmes, therefore, need to carefully review who is being reached by their services and implement interventions specifically tailored to engage people who might be missed.
There are various settings in which HTS can be provided to the public and expanded further, for instance in healthcare facilities, such as hospitals, clinics and mobile clinics, and at community sites, be these stand-alone or even home-based services, where testing services are provided within the community. 2 There is also an option for HIV self-testing (HIVST), which is carried out by an individual who wants to know his or her HIV status and is carried out privately by the individual alone. 5 HIV self-testing provides an opportunity for testing to be carried out discreetly and at one's convenience, which could increase the uptake of HIV testing amongst those unable or unwilling to access other healthcare services. 2 Concerns raised regarding HIVST include lack of HIV counselling, 30 instructions are difficult to follow 31 and there should be more of a focus on linkage to care. 32 This research study has a few limitations. 'Ever testing' for HIV is self-reported, and therefore, prone to biases related to social desirability, recall and under-reporting. Nevertheless, the results of the nationally representative population-based survey can be generalised to adults aged 15 years and above who tested for HIV in South Africa. There may be a high degree of within-district heterogeneity. In future, work will include examining the sub-district level estimates applying the robust methodology of small area estimation, which involves using auxiliary predictors to improve the precision of imprecise district-level estimates.

Conclusion
This study demonstrated the utility of visually displaying spatial inequities in HIV testing using nationally representative data by presenting simple maps for targeted priority setting. The findings suggest that provinces and districts with low testing coverage, especially amongst male participants, should prioritise tailored interventions to improve uptake of HIV testing. The strategies for HTS should include scaling up of HIVST and community HIV testing, specifically home-based testing to improve the uptake of HIV testing in those districts that are lagging behind in order to ensure equity in the geographical coverage of HIV testing.

Disclaimer
The findings and conclusions of this research study are those of the authors and do not necessarily represent the official position of any affiliated agency of the authors or the funding agencies.