In 2016, it was estimated that 30% of people living with HIV worldwide were unaware of their infection [1]. This gap represents a major challenge to achieving the targets of the Joint United Nations Programme on HIV/AIDS (UNAIDS) “90-90-90,” which aims to ensure that 90% of all people living with HIV are aware of their status by 2020 [2].

Unknown HIV status is associated with late entry into care and increased mortality rates and is a key driver of the epidemic [3-5]. An HIV test result is the first step in engaging in the HIV treatment continuum for those who test positive or in the HIV prevention continuum for those who test negative. Nonetheless, many barriers to expand HIV testing coverage persist, such as stigma, difficulty in accessing testing facilities, and inconvenient clinic hours [6,7].

HIV self-testing (HIVST) allows individuals to perform an HIV test and interpret their own results [8]. Different strategies for HIVST distribution worldwide have been proposed, such as through peers [9], nongovernment organizations (NGOs) [10], and mobile/social media apps [11]. Randomized controlled trials have shown that HIVST increases uptake and frequency of HIV testing overall, without adverse events, social harm, or increased risk behaviors [12]. Nevertheless, a review of HIVST implementation studies by Estem et al [13] concluded that despite the many potential benefits of HIVST, dissemination strategies need to be improved and the high cost of HIVST kits needs to be reduced to scale-up this strategy.

Key populations, including gay and other men who have sex with men (MSM), are greatly affected by HIV in Latin America and other regions [14,15]. Such populations have continued high HIV incidence and encounter multiple structural barriers, including stigma and discrimination, which increase their risk of HIV exposure and acquisition and inhibit access to evidence-based HIV prevention and treatment services [16]. In Brazil, the prevalence of HIV among MSM, estimated at 14% in 2009 [17] and 17.5% in 2016 [18], requires sustained prevention and treatment efforts. Furthermore, despite an increased proportion of MSM reporting annual testing (from 21.2% in 2009 to 43.3% in 2016) [19], innovations are needed to increase MSM testing coverage in the country. The convenience and discreet approach of self-testing would be particularly useful to increase first-time HIV testing and frequency among this vulnerable population.

Internet-based and mobile phone technologies are promising tools to overcome HIVST delivery barriers among MSM [20]. In the United Kingdom and United States, both online and social media–based promotion and delivery of free HIVST kits effectively increased the frequency of testing and reached MSM who were not previously tested [21,22]. In accordance with international data [23,24], a previous Brazilian study suggested that HIVST would be well accepted among MSM [25]. Thus, providing HIVST using these technologies may represent a step forward in increasing testing access in Brazil, where free HIV testing is already provided by the Brazilian Public Health System (“Sistema Unico de Saúde”).

Considering the limited access to and uptake of HIV testing among MSM in Brazil and the evidence supporting its acceptability, this paper reports on the development and feasibility of an internet-based HIVST (electronic testing [e-testing]) approach. E-testing was designed to promote HIV prevention by providing free anonymous HIVST and to enhance linkage to HIV care for those with a confirmed HIV positive status. We describe the results from the e-testing program’s first year, when HIVST was not available in Brazil outside the research studies.

The detailed frequencies and data sources of the e-testing study during its first year of implementation are provided in Figure 4.

We describe an internet-based strategy to provide HIV prevention information and HIVST to MSM in Curitiba, Brazil. In the first year, we delivered 2526 HIVST packages, substantially more than predicted in the study design (1000 packages). Our findings are similar to reports from previous studies that have shown good uptake and use of websites containing information on HIV prevention, self-assessment of risk, and HIVST among MSM [35,36]. Such high uptake is likely due to the significant number of MSM who use social media and the internet to look for sexual partners [37]. A previous Web-based survey in 10 Brazilian cities [38] found that 1798 (35.6%) and 678 (13.4%) of responding MSM reported daily and weekend use of websites and smart phone apps to seek sexual partners.

The use of internet and social media approaches for health promotion and behavioral interventions has become more routine according to international literature and already includes smoking cessation [39] and heart disease prevention programs [40]. The internet and social media are particularly promising ways to offer testing to MSM at high ongoing risk of HIV [41], as those using social media and the internet to meet sexual partners also have high-risk behavior and increased HIV risk [42,43]. Thus, internet-based health initiatives can help overcome health disparities to access prevention and care, particularly among MSM, and can help facilitate uptake of HIV testing, prevention, and treatment services.

E-testing appeared to appeal to never-tested MSM who reported a fear of having a positive result, shame (which is probably related to stigma), and unfriendly health facilities. These results are consistent with reports showing that HIVST is highly acceptable to MSM [25,44] and can facilitate uptake and frequency of testing. This demand was likely driven by the privacy of e-testing, which systematic reviews have highlighted as a priority for MSM [23]. Similarly, the large number of HIVST packages delivered by mail, compared to packages retrieved in the public pharmacy, was likely related to introducing delivery models that addressed challenges such as travel distances, wait times at facilities, and confidentiality issues [45,46], all of which have been previously cited as barriers to testing.

Of the individuals who answered the entire online questionnaire, most identified as young, white, well-educated, and without a steady partner. Although these characteristics were similar to those reported in other studies among MSM in the country [38,47], there were some differences as compared to a recent household probability survey conducted among men from Curitiba [48]: In our study, participants were younger, more likely to be white (77% vs 64%), less likely to perceive high risk for HIV infection (8% vs 52%), and less likely to have ever had an HIV test (62.5% vs 75.7%) than found in the survey [48].

However, the latest Brazilian estimates have shown an increase in the HIV prevalence among young MSM [18], who also showed riskier sexual behavior than older MSM [19]. This new evidence reinforces the importance of strategies to successfully reach this key population. Our findings suggest that, at least for young MSM, important subpopulations are open to new HIV testing strategies and understand the need for frequent or risk-based HIV testing.

In our study, HIVST was provided for free but without the provision of any incentives (the Brazilian Ethic regulations do not allow the provision of incentives for research participants or patients). We received 21.4% of HIVST results (18.2% through the website and 3.2% by mail). This was a substantially lower proportion than presented in China by Zhong et al [49], who reported 97% of self-testers returning results when HIVST kit costs were refunded upon return of results. We believe that this difference is related to the provision of incentives, and not providing incentives was one of the reasons why we estimated a low return rate when designing this feasibility study. However, the World Health Organization recommends self-testing as an additional HIV testing tool [8], not as an HIV surveillance tool, meaning that a low return rate may not represent a major drawback. On the other hand, confirmatory testing and rapid linkage to care are essential after a reactive HIVST result. A study conducted in Brazil reported that of 131 MSM who tested positive, only 95 (72.5%) were linked to care [50]. Because of the anonymity of the e-testing strategy, the proportion of study participants who sought confirmatory testing is uncertain. Furthermore, although users were instructed to bring their PIN to the COA when seeking confirmatory testing, no one accessing services at COA provided a PIN. Therefore, we could not match the HIVST results and confirmatory results. Other programs seeking ways to track users’ confirmatory testing and linkage to HIV services following HIVST should be aware of this limitation and the challenges with monitoring HIVST implementation. Other potential caveats of self-testing discussed in the literature are psychological crises and social harm. In our study, neither of these events were reported, which is consistent with the international literature [12,23].

E-testing was largely successful during its first year: E-testing achieved most of the previously defined study outcomes, and demand for HIVST was twice that initially anticipated. However, our study had limitations. The anonymous, self-reported nature of data collection and the low proportion of users who reported their HIVST results prevented us from validating individual HIVST results with confirmatory tests and from estimating the HIV prevalence (due to the inability to exclude multiple answers from the same individuals). Additionally, because we did not track the second test distributed in the package, it is possible that the test requestor gave the second test to a friend or partner, which would increase the number of individuals tested during the study period. We likely also missed MSM who may be fearful of HIV testing and those who did not have access to the internet, which limits the generalizability of our findings. Despite these limitations, we achieved our primary goal, which was to reach individuals who do not access traditional HIV testing services.

In conclusion, our findings show that an internet-based strategy to provide free, anonymous HIVST kits was acceptable and feasible in Brazil. Lessons learned from this study may also be applied to other countries in Latin America with similar cultural and epidemiological contexts. The discretion and convenience of HIVST may increase testing among MSM who are not reached through existing services. Future investigation is needed, however, to identify strategies to support linkage to prevention and treatment services and to determine the potential public health impact and cost-effectiveness of HIVST in Brazil.