Background Globally, external HIV-related stigma is a major threat to all HIV prevention, care and treatment interventions including the recently launched Universal Test and Treat (UTT) strategy in South Africa and the 90-90-90 targets set by UNAIDS for the global response by 2020. The 90-90-90 targets are put in place to track the progression from HIV testing to durable viral load suppression among people living with HIV. The targets guide HIV programmes to achieve 90% known HIV status, to access 90% antiretroviral therapy and to suppress 90% viral loads (UNAIDS, 2017). Achievement of the 90-90-90 targets has since become a part of South Africa's National Strategic Plan for HIV, TB and STIs 2017-2022. External HIV-related stigma in this study was defined as the presence of one or more of the following attitudes and behaviours: rejection, avoidance, intolerance, stereotyping, discrimination, and physical violence towards people living or perceived as having HIV. There have not been many efforts to attenuate HIV-related stigma in South Africa, as it continues to exist. There has been a scale up of other HIV responses, such as HIV Counselling and Testing (HCT) and treatment, with the argument made that in scaling up these biomedical approaches, stigma would disappear. Furthermore, its exact magnitude, trends over the years and correlates have not been explored fully at national level, hence the present study known as Stigma Assessment Study in South Africa (SASSA). SASSA is based on the Institutional Social Construction theory framework. It explores the external HIV-related stigma magnitude and its trends between 2005 and 2012, as well as the associated factors which influence its prevalence in South Africa at a national level. The study further explores the mediating and moderating factors of external HIV-related stigma and tries to explore external HIV-related stigma by viewing individuals with HIV living in families, societies and structures, with the hope of contributing to the development of new systematic HIV-related stigma interventions in South Africa as well as, strengthening existing ones. Methodology The project used secondary data obtained from three South African national population HIV surveys which were conducted in 2005, 2008 and 2012 by a research consortium led by the Human Sciences Research Council. Sub-samples of the original surveys consisting of respondents aged 15 years and older who had responded to the stigma questions in the three surveys were extracted and included in the SASSA analysis. A nationally representative sample of a total of 16 140 individual respondents from the 2005 survey, 13 134 from the 2008 survey and 30 748 from the 2012 survey was used in the study. Two different measures of external HIV stigma are used in this analysis, one is a summary measure from the latest survey data, i.e. 2012, which was used to do a regression analysis. The summary measure was regarded as reliable to use for the regression analysis as it provides crude effect of the exposure factors on external HIV stigma. However, this summary measure was not included in the previous surveys, i.e. 2005 and 2008, and therefore could not be used for trends analysis. We therefore used 4 individual stigma items for the trends analysis as these were included in all 3 surveys. The use of 4 individual stigma items was important because when data from a variety of sources or categories have been joined together, the meaning of the data can be difficult to see. It was therefore considered ideal to assess the performance of the individual constructs on their individual contribution to the impact on HIV external stigma. Furthermore, analysis using individual constructs provided an opportunity to see specific patterns which could have remained obscure in crude analysis. The first measure of external HIV-related stigma used in the regression analysis was measured by five individual items which elicited attitudes towards people living with HIV (PLHIV). The five items were based on a 9-item scale that was originally developed and tested in a South African population, and the 9-item scale was found to be internally consistent (alpha = 0.75) and reliable (r = 0.67). The 5 items were (1) People who have AIDS are dirty; (2) People who have AIDS are cursed; (3) People who have AIDS should be ashamed; (4) People with AIDS must expect some restrictions on their freedom; (5) A person with AIDS must have done something wrong and deserves to be punished. As explained above, the specific external HIV stigma patterns second measure, which was utilized to explore stigma trends over the years 2005, 2008 and 2012, consists of four individual stigma items which elicited attitudes towards PLHIV. The four individual stigma items included: (1) If you knew that a shopkeeper or food seller had HIV, would you buy food from them? (2) Would you be willing to care for a family member with AIDS? (3) Is it a waste of money to train or give a promotion to someone with HIV/AIDS? (4) Would you want to keep the HIV positive status of a family member a secret? Findings Overall, external HIV-related stigma was found to exist among 38.3% of adult South Africans in 2012. Multiple regression analysis showed that predictors of external HIV-related stigma were race, sex, education level, self-perceived risk of HIV infection and HIV knowledge (p< 0.01). Females were less likely to report external HIV-stigma than males (AOR = 0.9, P< 0.001). Those who perceived themselves to be at high risk of HIV infection were less likely to display some stigma than those who believed they were at low risk (AOR = 0.89, p < 0.01). The study did not find any significant associations between HIV testing or awareness of HIV status, with external HIV-related stigma in this study. Looking at the individual external HIV stigma items used to measure trends, the study reveals a slight decrease in the reporting of stigma over the three time periods (2005 vs 2008 vs 2012) on responses for two of the stigma items (Q1: If you knew that a shopkeeper or food seller had HIV, would you buy food from them, and Q2: Would you be willing to care for a family member with AIDS). While an increase was observed in the reporting of stigma over the three years on responses for two of the stigma items (Q3: Is it a waste of money to train or give a promotion to someone with HIV/AIDS, and Q4: Would you want to keep the HIV positive status of a family member a secret). The structural equation modelling (SEM) showed likelihood ratio test results with a p-value greater than 0.05, a root mean square error of approximation (RMSE) of 0.008 and Tucker–Lewis index (TLI) value of 0.985. The model fit assessment results allow us to accept that an hypothesized model of the study is not far from a perfect model. The SEM results also showed a direct effect of sex on HIV knowledge statistically significant at p < 0.001, with race having an effect of 3.3% and education a direct effect of 9.5%, and both of these showed a statistically significant effect (p < 0.001) respectively. HIV knowledge showed to have a statistically significant inverse relationship on external HIV stigma of -10.4% (95% CI: -12.3-0.09) p< 0.001. Awareness of HIV status had the highest positive direct effect on external stigma of 10% (95% CI: 4.41-15.67%) p<0.001. With regard to indirect effects, sex, race, and education had minimal negative indirect effects on external stigma, which was statistically significant for all the three covariates. With the said effects of external HIV-related stigma, it was found that HIV knowledge independently mediates the relationship between Level of Education, Awareness of HIV status, Race, HIV testing history, and Sex, with External HIV-related stigma. Conclusion External HIV-related stigma still exists in South Africa despite previous success in massive ART rollout, HTC campaigns, and most recently test-and-treat programmes, which were arguably thought to have a parallel effect in the decrease of HIV related stigma. The focus on individualistic health structural approaches that do not generally have stigma-reduction as a specific aim, as discussed, is likely to undermine the successes achieved in the fight against HIV thus far. There is a need to develop innovative holistic interventions which are specifically intended for HIV stigma reduction. These should be inclusive of both social institutional elements and health structural elements to address the challenge of external HIV-related stigma.