Brazil has the largest population of individuals living with HIV and acquired immune deficiency syndrome (AIDS) in Latin America [1], representing a concentrated epidemic with an estimated HIV prevalence of 0.6% in the general population (0.4% among women and 0.8% among men), and a 14.2% prevalence among men who have sex with men (MSM) [2]. Geographic differences have been reported in HIV prevalence among MSM (ranging from 5.2% [Recife] to 23.7% [Brasília]) in a respondent-driven sample survey conducted in 10 cities [2]. In 2016, approximately 60% of reported HIV infections were attributed to male-to-male sexual contact, although MSM represent only 3.5% of the Brazilian population [3,4]. New infections in this population continue to rise, especially among young people (24 years-old or less) [3].

Preexposure prophylaxis (PrEP) with tenofovir/emtricitabine is now a key component of prevention packages for MSM. The efficacy of treatment for this population has been shown with both once daily and on-demand use in clinical trials and open-label studies [5-9], and demonstration studies have been successfully implemented in different settings [10-13]. Results show that willingness to use PrEP among MSM may vary across different geographic locations, ranging from 32% to 92% [14-28].

With expanded access to the Internet via mobile phones, geosocial networking (GSN) smartphone apps are becoming the main venue for casual sexual encounters [29-32]. MSM report logging into GSN apps at least three times per day, with an average of 12 minutes spent per log-in [33,34]. These new technologies have generated quicker and easier ways for MSM to meet potential partners, and may facilitate the users’ ability to have multiple concurrent partners, thereby increasing their risk for acquiring HIV [32,35-37]. Men who use the Internet to meet other men may present a different behavioral profile than men who meet men in physical venues. For instance, MSM who do not identify as such would be excluded from venue-sampling [38]. Therefore, a better understanding of the profile of MSM who use apps for sexual encounters is needed in order to design tailored, combined prevention interventions. This study aimed to describe the characteristics of Brazilian MSM using two GSN apps for sexual encounters, including risk behavior for HIV infection and their awareness of prevention strategies. In addition, factors associated with daily oral PrEP willingness were assessed.

During the 30 days of the online survey, 8885 individuals provided informed consent. Of these, 23.05% (2048/8885) were ineligible, 6837 (6837/8885, 76.94%) initiated the questionnaire, and 5065 (5065/8885, 57.00%) completed the questionnaire and were included in the present analysis (Figure 1). Differences among those who did not complete the questionnaire (n=1772) and those who completed the questionnaire (n=5065) are presented in Table 1.

Among all individuals who accessed the questionnaire and completed questions regarding HIV serostatus (n=6664), HIV prevalence was 12.27% (818/6664). Considering ineligible respondents (n=2048), 2.29% (47/2048) self-declared as cisgender women, 1.56% (32/2048) as transgender women or transvestites, 1.95% (40/2048) as transgender men, and 9.33% (191/2048) as other genders.

Among those included in this analysis (n=5065), most participants were from southeast Brazil (3532/5065, 69.73%) and accessed the survey through the Hornet app (2800/5065, 55.28%), and a total of 4618 (4618/5054, 91.37%) respondents used apps for sexual encounters. The median age of the cohort was 30 years old (interquartile range: 25-36), 63.06% (3194/5065) were white, 68.13% (3541/5065) reported middle-to-high income, and 61.36% (3106/5062) reported more than 12 years of schooling. Most respondents identified themselves as MSM (4991/5065, 98.54%) and reported having friends with the same sexual orientation (3368/5065, 66.50%). Only 20.48% (1029/5024) of respondents reported a male steady partner and 3.97% (198/4986) reported a female steady partner (Table 1).

Binge drinking and tobacco use in the last six months were reported by 71.79% (3636/5050) and 32.60% (1651/5065) of individuals, respectively (Table 2). Marijuana (or hash) was the most frequent illicit substance reported (1679/5065, 33.15%), followed by stimulants (1177/5065, 23.24%), hallucinogens (519/5065, 10.25%) and GHB (222/5065, 4.38%). A total of 2305 (2305/5065, 45.51%) participants reported no substance use in the last six months.

In the previous 6 months, only 236 (236/5065, 4.66%) respondents reported having no sexual partners. Condomless receptive anal sex prevalence was high (2121/5065, 41.88%). Approximately 10% (480/5065, 9.48%) of respondents reported having had sex with one HIV-positive partner, 2.19% (111/5064) with more than one partner, and 18.62% (943/5064) reported that they did not know how many HIV-positive partners they had sex in the prior six months. Regarding the number of times they were the insertive partner without a condom with an HIV-positive partner, 18.90% (957/5064) of participants reported one to four times and 8.87% (449/5064) reported five times or more. Reported prevalence of STIs (syphilis, gonorrhea, or rectal chlamydia) in the previous 6 months was 12.06% (604/5010; Table 2). Most of the participants (3363/5064, 66.41%) scored >10 points in The HIV Incidence Risk for MSM Scale (high HIV risk) and fall into the category of individuals who should undergo evaluation for PrEP use. Conversely, only 21.39% (1083/5064) of participants had high HIV risk perception and 9.16% (464/5065) reported nPEP use in the last 12 months. Among those that used PEP, 81.68% (379/464), 12.72% (59/464), 2.59% (12/464), and 3.02% (14/464) reported nPEP use once, twice, three times, and more than three times in the past 12 months (respectively).

A total of 738 (738/5065, 14.57%) respondents had never performed an HIV test and one of the main reasons was the fear of having an HIV-positive result (263/738, 35.64%). Compared to those who had been tested before, these respondents were younger (24 years old, interquartile range: 21-30), had less years of schooling (<12 years: 495/736, 67.26%), had lower income (less than 4 minimum wage: 359/738, 48.64%), and almost half self-reported as white (370/738, 50.14%). Almost half of the respondents believed that the best testing venue is at home (2400/5065, 47.38%) and most would be comfortable with picking up the HIVST somewhere else (2867/5065, 56.60%; Table 3).

PrEP, nPEP and HIVST awareness was reported by 57.89% (2932/5065), 57.39% (2907/5065), and 26.57% (1346/5065) of respondents, respectively. Willingness to use different HIV prevention methods is depicted in Figure 2. Willingness to use daily oral PrEP and injected PrEP was similar (2653/5065, 52.38% vs 2408/5065, 47.48%), while PrEP on demand was lower (1751/5065, 35.57%) and PrEP during short periods or vacations was much higher (4652/5065, 91.85%). In addition, 51.08% (2587/5065) of respondents would use PrEP if available commercially and 4.72% (239/5065) would never use PrEP.

In the final multivariate model (Multimedia Appendix 1), variables independently associated with daily oral PrEP willingness were: high number of male sexual partners (>10) versus 0-5 partners (AOR 1.26, 95% CI 1.09-1.47), condomless receptive anal intercourse (AOR 1.27, 95% CI 1.12-1.44), sex with HIV-positive partner versus no HIV-positive partner (one HIV-positive partner: AOR 1.36, 95% CI 1.11-1.67), daily use of apps for sexual encounters versus never use (AOR 1.48, 95% CI 1.17-1.87), high and unknown perceived likelihood of getting HIV in the next year (AOR 1.72, 95% CI 1.47-2.02 and AOR 1.39, 95% CI 1.13-1.70), STI diagnosis (AOR 1.25, 95% CI 1.03-1.51), stimulant use (AOR 1.24, 95% CI 1.07-1.43), PrEP awareness (AOR 1.48, 95% CI 1.30-1.70), and unwillingness to use condoms (AOR 1.16, 95% CI 1.00-1.33).

Our findings provide important insights about the characteristics of Brazilian MSM based on two popular GSN apps for sexual encounters, and their preferences amidst the HIV prevention strategies. This information can potentially guide and support national prevention programs. Over half of the respondents would be willing to use daily oral PrEP if available through SUS, which is lower than previously reported in Brazil [21]. This finding is probably related to the fact that the previous studies also included individuals who were actively seeking PrEP, and thus possibly overestimated PrEP willingness. Nevertheless, our finding is consistent with reports from other settings [14,20-28,34] and represents a reassuring result, given that daily PrEP will soon be provided by SUS at no cost for MSM at higher risk of contracting HIV [48].

Our study showed that, in addition to daily oral PrEP, a high proportion of respondents were also interested in other PrEP formulations, such as injectable PrEP (2408/5065, 47.48%). Long-acting injectable PrEP can be very beneficial among individuals for whom adherence to a daily oral regimen is challenging [22,49,50]. Cabotegravir, an integrase inhibitor that can be administered through intramuscular injection, is one of the products currently under clinical development for PrEP formulation [51]. Individuals’ personal preferences and priorities may have a significant impact on acceptability, uptake, and retention of different PrEP modalities and the ability to choose among multiple contraceptive methods, which was shown to be associated with increased population coverage [52].

Although nPEP has been available at no cost through SUS since 2009, awareness, willingness, and use were lower than expected. This problem is highly concerning given that almost half of our sample reported unprotected anal sex. Similar results have been described from other settings with MSM-concentrated epidemics [53,54].

Few of the respondents had heard about HIVST; however, half of them would be willing to use it and this strategy, which could increase serostatus awareness and facilitate integration into HIV care [55]. The acceptability of HIVST ranged from 21-98% in studies among MSM [56], including a pilot study using blood-based HIVST in Brazil and Peru [57], and an Internet-based feasibility study using oral fluid, which was conducted in Brazil [58]. As social networking advertisements on HIVST were shown to be effective at increasing HIVST awareness and uptake [59,60], feasibility of HIVST distribution through apps for sexual encounters should also be further evaluated.

Although smartphones became the main devices used for Internet access, especially in lower income families [61], it is hard to estimate the proportion and possible selection bias of online studies. Most of the MSM included in this study were white and aged 25-36 years, and reported middle to higher income with more schooling years, which may not reflect the Brazilian MSM population. Some of these characteristics are in accordance with a systematic review, which showed that those using apps were younger, presented higher educational levels, reported higher incomes, had higher proportions of risky sexual behaviors, and were more likely to have tested for HIV in their lifetime compared to nonapp-using MSM [62]. In addition, our sample was similar to the one recruited on Grindr for a study in Los Angeles [63] and another from a social networking site for MSM in Latin America, Spain, and Portugal [38]. Indeed, studies have shown that online samples tend to be biased toward a lower median age, as younger MSM are over-represented on GSN apps [64]. Conversely, younger MSM aged 24 years or less was not the majority of this sample, perhaps because the content of the survey or the advertisement strategy was not attractive to this population.

Considering The HIV Incidence Risk for MSM Scale, most of the sample should be further evaluated for PrEP use, but perceived likelihood of getting HIV in next year was low. Moreover, rates of unprotected anal intercourse reported by study participants was high, which has also been found to be high among MSM who use Grindr in studies conducted in the United States [35,65] and among MSM from a website for sexual encounters in Latin America [38]. The CDC recommends that sexually active MSM should be tested annually, and clinicians should consider the potential benefits of more frequent HIV screening (eg, every three or six months) for some asymptomatic sexually active MSM, based on their individual risk factors, local HIV epidemiology, and local policies [66,67]. Almost 15% of the sample (738/5065) had never performed an HIV test and 15% (762/5065) were only tested once in their lifetime, which is almost two-fold higher than that observed in a US study using a GSN app [68], but much lower than that found in other Brazilian studies [2,69]. Knowledge of HIV status enables individuals to make decisions about behavioral strategies to reduce HIV transmission risks, such as serosorting [70-72], using condoms with partners who do not share the same HIV status, or restricting behaviors to partners who are HIV-positive and have undetectable viral loads (UVLs) or HIV-negative partners that take PrEP [73]. Disclosure of PrEP use and UVL is not uncommon among MSM using apps in the United States, and the majority of the respondents have engaged in condomless sex at least once based on this status [73]. This behavior is unknown in Brazil, as PrEP is still not available and information and knowledge about the relationship of UVL and HIV transmission is not widely spread. Hopefully this reality will change in the foreseeable future, and studies to evaluate this behavior shift are needed.

Binge drinking and tobacco and substance use were high in the study population, compared with the general population [74], which is consistent with previous reports [75,76]. In an online survey among MSM through Facebook in seven countries, including Brazil, it was observed that social networks and minority stressors can have significant effects on drug use and sex while drunk or high [77]. Binge drinking and drug use (eg, marijuana, amphetamines, poppers) in general have been associated with condomless intercourse [78-80]. A recent study in young lesbian, gay, and bisexual individuals showed that the use of marijuana is associated with sex with multiple partners [81]. Conversely, a lower prevalence of binge drinking and substance use was observed during the PrEP Brasil Study [10]. Nevertheless, public policies on HIV prevention still need to acknowledge and address the relationship between substance use and risk behaviors.

Moreover, our results provide evidence that MSM reporting higher risk behaviors were more willing to use daily PrEP, as observed in other studies [21,82-87]. Accordingly, our results show that those reporting higher or unknown perceived HIV risk and STI diagnoses in the last six months were also more willing to use daily PrEP. This is an important finding since PrEP is recommended to high-risk MSM. In addition, those not willing to use condoms were more willing to use PrEP. In a study comparing preferences for PrEP, condoms, and both PrEP and condoms, MSM reporting recent risk behaviors were more likely to prefer PrEP compared with condoms only, and less likely to prefer both methods compared with condoms only [88].

Unequivocally, this study has limitations. First, online studies are not probabilistic sampling strategies, thus precluding the generalizability of the findings. Given the cross-sectional nature of the data, causality and the direction of association may not be inferred. All collected data were self-reported by participants and may be subjected to bias, including social desirability bias. Our data were also subjected to recall bias due to 6-month or 12-month recall periods. There is also a concern about participants taking the survey multiple times. To avoid this issue, the first question of the survey was, “Are you answering this survey for the first time?” (4% of participants answered “no” and were excluded from the study). Finally, we have measured intention to use PrEP, nPEP, condoms, and HIVST as a proxy of willingness. There are different methods for accessing PrEP willingness, as reviewed by Young and McDaid [89], and as such our results should be interpreted with care.

In summary, our observed high HIV risk scores suggest that most MSM would be eligible for PrEP, and that those who present risky sexual behaviors were more willing to use it. Notwithstanding, only 58% (2932/5065) of individuals were aware of this prevention strategy. Additionally, awareness, willingness, and use of nPEP—which has been available in Brazil since 2009—were low. Efforts to increase awareness of new prevention strategies are urgently needed to create demand among those at the highest risk for HIV infection. Mobile health tools are a promising strategy to reach high risk MSM in Brazil.