Of 13,320 male respondents, 442 (3.0%) reported sex with a man in the past year, of whom 215 (46.3%) had met a partner online. Table 1 presents the demographic characteristics of MSM overall and by whether they had OSP in the past year. Overall, 33.0% of MSM were aged 15 to 24 years, 55.2% were non-Hispanic white, 61.6% had more than a high school diploma or GED, 84.8% had never married nor cohabited, and 73.1% lived above the poverty level. About 60% of MSM identified as gay, 27.8% identified as bisexual, and 12.2% identified as straight. Almost half of MSM reported an OSP in the past year (46.3%). Men who have sex with men with OSP did not differ significantly from those without by age, race/ethnicity, marital status, poverty level, or education. However, a higher percentage of men who had OSP identified as gay than those without (68.4% vs 49.5%, P = 0.0124).

Table 2 shows the distribution of behaviors by whether MSM had OSP in the past year. The mean number of male sexual partners (oral or anal sex) in the past year was higher among MSM who had OSP (M = 3.04; SD = 0.18) than those who had not (M = 1.60; SD = 0.11; P = 0.0000). A higher percentage of MSM with OSP reported 2 or more oral sex partners (65.1% vs 28.0%, P < 0.0001), receptive anal sex partners (33.2% and 15.4%, P = 0.0055), and insertive anal sex partners (46.3% and 8.5%, P < 0.0001) in the past year than those without.

Regarding sexual risk behaviors, no differences were seen in condom use at last receptive or insertive anal sex by whether MSM had OSP or not. However, serosorting, or limiting partners to those with the same HIV status, was more prevalent among MSM with OSP (88.3% vs 70.7%; P = 0.0001). A higher percentage of MSM with OSP reported sex with a male partner who had other sex partners (72.9% vs 49.6%, P = 0.0002) or sex in exchange for money or drugs (10.5% vs 2.8%, P < 0.0001) than those who had not.

A higher percentage of MSM with OSP reported sexual health care behaviors in the past year than MSM without OSP (Table 3), including receiving a sexual risk assessment from a doctor or other medical care provider (56.0% vs 40.4%; P = 0.0129), and testing for HIV (48.4% vs 31.3%; P = 0.0134) or other STI (57.4% vs 35.3%; P = 0.0002). Also, higher percentages of MSM with OSP reported treatment for any STI (17.8% vs 8.7%; P = 0.0152) or were told by a doctor or other medical care provider that they had chlamydia or gonorrhea in the past year (13.2% vs 2.3%; P < 0.0001), genital herpes (10.0% vs 2.7%; P = 0.0020), or syphilis (9.8% vs 2.6%; P = 0.0225) in their lifetime. The percentage of MSM who reported extragenital STI screening (pharynx or rectum) in the past year was higher for those with OSP than those without, but the difference was not statistically significant (63.5% vs 49.2; P = 0.2321).

Almost half of MSM in this sample reported OSP in the past year. We did not find sociodemographic differences among MSM by OSP except that those who met partners online were more likely to identify as gay. When considering behaviors associated with increased risk of STI/HIV, MSM with OSP reported more partners and a higher percentage reported exchanging sex for money or drugs and having sex with partners who had other partners. However, having OSP was not correlated with more drug use or less condom use. Behaviors typically considered to be protective from STI/HIV were more likely to be reported by MSM with OSP, including serosorting and higher use of sexual health services. These men, although more likely to report having had an STI, were also more likely to indicate getting tested or screened for an STI.

The considerable proportion of MSM reporting an OSP may reflect the increasing availability and accessibility of these sites over the past 2 decades, effectively the normalization of online dating and sex seeking. Men who have sex with men with OSP were more likely to identify as gay, which may suggest that using online sites and dating apps are a common or accepted activity among MSM who disclose their sexual identity. Although MSM with OSP reported more sex partners, in this study, we were not able to determine if use of online websites and apps was the reason for having more partners, or if these men would have more partners regardless of these technologies. Because using technology may be correlated with the number of sex partners, it is difficult to discern if using the Internet or dating apps to meet sex partners is riskier than any other venue. Because of survey question limitations, we were only able to measure if any sex partner in the past year was met online, rather than ascertain the number of partners met online or frequency of sex seeking behavior. Additionally, because all data are reported from the perspective of the respondent, it is not possible to infer information about sexual networks. And although MSM with OSP reported a greater likelihood of having sex with a partner who had other sex partners, this may be a function of the many online websites and apps whose goal is to facilitate the meeting of many men regardless of relationship or partnership status. Moreover, an OSP does not necessarily mean risker sex. As we found in this study, there were no differences in condom or drug use by whether MSM had OSP. A higher percentage of MSM with OSP reported exchanging money or drugs for sex, but the total number of men reporting this, although significant, was low (n = 24).

Men who have sex with men with OSP were more likely to report engaging in protective behaviors, such as serosorting and receiving sexual health services. Identifying partners with the same HIV status may be easier online because many sites allow users to disclose their status or indicate status preference in partners when establishing a profile. It should be noted that although more men, regardless of OSP, reported condom use at last anal sex (insertive or receptive), there was still a sizable proportion of men (38.3%–45.3%) who did not use a condom. A proportion of condom nonuse among MSM in this sample may be because of use of preexposure prophylaxis (PrEP) to prevent HIV exposure; however, we were unable to determine use of PrEP by MSM in NSFG.

Engagement in sexual health care behaviors could be influenced by many factors, including symptoms of or exposure to an STI or exposure to STI/HIV health messages, which are common online and in apps. If men with OSP are more likely to seek sexual health care, they are more likely to receive STI/HV testing, which may result in higher rates of STIs. Many of the men in this sample did not receive STI or HIV testing in the past 12 months and a low proportion of men reported STI testing in the throat or rectum, as recommended by the Centers for Disease Control and Prevention,⁹ despite high numbers of oral sex partners, suggesting opportunities to increase receipt of sexual health services.

Results from this study confirm the findings from many previous, less rigorous studies that men who use online venues including dating apps engage in behaviors that may increase their risk of acquiring an STI but are also more likely to engage in behaviors that may protect against STI/HIV.^1-3 A strength of this study is that the estimates are relatively comprehensive, incorporating MSM who are not encountered in venue-based assessments or via online surveys in which people respond to a general solicitation. Such respondents might have different risk and vulnerability profiles and provided a basis of comparison absent in many studies. However, the cross-sectional design of the NSFG prevents us from making causal inferences regarding Internet/mobile app use and STI risk-related behaviors among MSM. Nor does NSFG collect STI biomarkers, so we were not able to directly examine the relationship between Internet use and STI acquisition. Other limitations include recall biases, which could also affect estimates, and our inability to control for some variables that could affect sexual risk, such as alcohol and drug use during sex or use of PREP, or generalize findings to men older than 44 years.

These data are some of the first to explore OSP among a probability sample of MSM and offer a baseline for understanding OSP app use in this population. Future research describing sexual behaviors among MSM who meet sex partners online will help us understand the role of the Internet, dating apps, and other future emerging technologies that may increase STI/HIV risk. Future research is also needed to capture trends in online sex seeking over time to clarify if, and when, online venues impose an increased risk for STI/HIV and when that risk may be mitigated. Lastly, more research on individual level motivation for using online venues for sex seeking, and a better understanding of men's awareness and mitigation of risks associated with online/dating app use may help inform STI/HIV prevention efforts.