This study was a cross-sectional analysis of data from ASHR2. The methods of ASHR2 have been described elsewhere [7]. In brief, ASHR2 is a national survey of 20,091 Australian residents aged 16-69 years. Data were collected in 2012 and 2013 via a computer-assisted telephone survey by trained interviewers. The study sample was selected using a modified random digit dialing sampling frame, which combined random digit dialing of landlines with that of cell phones. The overall participation rate among eligible people was 66.22%; the study population has been shown to be broadly representative of the Australian population, except for an overrepresentation of people with postgraduate degrees [7].

To allocate resources efficiently and gather more information from those with potentially higher HIV and sexually transmissible infection (STI) risk, we administered interviews in two forms [7]. All respondents who reported no sexual partners or >1 sexual partner in the previous year or any lifetime same-sex experience were given a long-form interview, as were a 20.00% random sample of survey respondents who had reported having 1 partner in the previous year and no same same-sex experience; the remaining 80.00% of one-partner respondents were given the short-form interview. Questions asked only in the long-form interview included those on meeting and having sex with a partner met on websites and smartphone apps.

The estimates of prevalence included sexually active and sexually inactive respondents. However, the predictor analysis of factors associated with meeting and having had sex with someone met on websites or smartphone apps was restricted to sexually active respondents because many sexual health outcomes were queried only of sexually active survey respondents. For this study, people were considered sexually active if they had ≥1 partners (for vaginal or anal intercourse or oral or manual sex) of the same or other sex in the previous 12 months. Respondents who reported no lifetime sexual experience were coded as not sexually active.

The primary outcome of this study was having sex with someone met on a website or a smartphone app in the past year. We also calculated the proportion of people looking for partners and meeting someone in person who they first met on a website or mobile app. These outcomes were ascertained using the following questions collected in the long-form questionnaire. The questions’ exact wording is shown in Textbox 1.

The proportion of people searching for, meeting, and having sex with partners on websites and smartphone apps were calculated separately using descriptive statistics. Data were weighted according to the Australian population and the probability of being selected for the long-form questionnaire. The characteristics of respondents who reported searching for partners using these technologies were compared with respondents who reported having sex with someone met on a website or a smartphone app. A chi-square test was used to compare differences in distributions between groups for a range of covariates.

Univariate and multivariate logistic regression, weighted in accordance with study procedures, were used to examine factors associated with having had sex with someone met on a website or a smartphone app in the last year. All data were analyzed using Stata statistical software version 14. Variables significant at the P<.1 level in the univariate analysis were included in multivariate logistic regression analyses. Backward elimination of variables was then used to determine the final adjusted model.

The demographic covariates included in the models were age group (16-29 years, and then 10-year age groups up to 69 years), language spoken at home (English or other), annual household income: very low or low (<Aus $52,000), middle (Aus $52,001- Aus $83,000), high (Aus $83,001-Aus $125,000), and very high (>Aus $125,000), and area of residence (urban or rural and remote) according to the Accessibility/Remoteness Index of Australia [9].

The behavioral covariates included in models were levels of alcohol consumption (high or not, with high alcohol consumption classified as >28 standard drinks per week for men and >14 standard drinks per week for women), injecting drug use in the last year (yes or no), smoking status (never and former, or current), sexual identity (heterosexual or gay, lesbian, bisexual, and other), condom use at last event (used or did not use condoms), STI history the last year (no STI test, STI test, or STI diagnosis), and sexual partner numbers in the previous year (1, 2-3, or >3). The numbers of sexual partners included both male and female partners. In relation to STIs, respondents were asked whether they had had an STI in the past year and whether they had an STI test; these two questions were combined to provide the composite variable. The STIs included were pubic lice, genital warts, chlamydia, genital herpes, gonorrhea, and syphilis; in addition, for women, warts virus on Pap smear, pelvic inflammatory disease, bacterial vaginosis or gardnerella, and trichomoniasis were included and for men, nonspecific urethritis and anal warts were included [10].

The study received La Trobe University’s (HEC 11-040) ethical approval, which was ratified by the ethics committees of the University of New South Wales, the University of Sydney, and the University of Sussex.

Overall, 12.09% (2346/19,398) of respondents reported ever searching for potential partners on websites and smartphone apps (13.52% men [1320/9761], 10.65% women [1026/9637]) and 5.40% (1048/19,398) of respondents (7.01% men [685/9637] and 3.77% women [364/9636]) reported doing so in the last year. Table 1 shows that searching for potential partners using smartphone apps and websites in the last year was most common among people aged 16-29 years (8.42%, 435/5169) and decreased with increasing age to 1.87% (52/2785) among people aged 60-69 years. Furthermore, 4.92% (815/16586) of sexually active respondents used websites and smartphone apps to look for potential partners in the last year, and 8.28% (233/2811) of sexually inactive respondents did. The activity was more common among gay, lesbian, and bisexual respondents (25.32%, 172/680) than among heterosexual respondents (4.68%, 876/19715).

Having met in person was reported by 2.98% (578/19,398) of survey respondents (3.70% men [363/9761], 2.19% women [214/9637]), whereas having had sex with someone first met on a website or a smartphone app was reported by 1.95% (378/19,398) of respondents (2.54% men [248/9761], 1.35% women [130/9637]). Having had sex with someone met on an internet site or a smartphone app in the last year was the highest among respondents who identified as gay, lesbian, or bisexual (15.80%, 107/680) and also more frequent among people aged 16-29 years and 30-39 years, as well as those who had had an STI test or STI diagnosis in the past year, those with a higher number of sexual partners, and those who used a condom at their last sexual event (Table 1). Results from Table 1 stratified by sex are presented in Multimedia Appendices 1 and 2.

When restricted to sexually active respondents,11.41% (1893/16586) had ever searched for potential partners online, 4.92% (815/16586) had done so in the last year, and 2.97% (493/16586) had met with someone in person in the last year (Table 1).

Table 2 displays the proportion of respondents who had sex with someone met on a website or a smartphone app, among those who reported using websites and smartphone apps to search for potential partners in the last year overall and according to selected characteristics. Overall, slightly over one-third (36.07%, 378/1048) of those who searched for partners also reported having had sex with someone they met using these technologies in the last year. There were no statistical differences by sex. Differences were observed for sexual identity, with gay, lesbian, and bisexual respondents being more likely also to report having sex with someone they met on a website or a smartphone app, than heterosexual respondents (62.41% [107/172] vs 30.89% [271/876], P<.001). Differences were also seen for STI testing history (no test: 35.89% [170/473], STI test: 53.77% [171/318], and STI diagnosis: 64.94% [33/51], P<.001), condom use at most recent event (used condoms: 48.43%[151/312] vs not used: 42.85% [122/285], P<.001), and numbers of sexual partners (1 partner: 15.36%, [83/541], 2-3 partners: 48.82% [150/307], and >3 partners 72.44% [145/200], P<.001).

Among the respondents who had had sex with someone they met on a website or a smartphone app, 62.88% (363/578) were male and 37.12% (214/578) were female. The majority of people who reported having had sex with someone met on a website or a smartphone app identified as heterosexual (77.75%, 129/578). In terms of STI testing, 41.91% (242/578) reported not having had an STI test in the previous year, 38.81% (224/578) reported having had an STI test, and 7.24% (42/578) reported being diagnosed with an STI in the last year (12.03%, 70/578) of respondents either refused to answer, could not recall, or were not asked). Regarding the numbers of sexual partners, 36.45% (211/578) had 1 sexual partner in the last year, 35.28% (204/578) had 2 or 3 sexual partners, and 28.27% (163/578) had >3 sexual partners.

Overall, our findings indicate that in 2012-2013, approximately 1 in 10 Australian adults aged from 16 to 69 years had ever looked for potential partners using websites or smartphone apps, of whom approximately half had done so in the in the last year. Among people who searched in the last year, over half had physically met with someone, and approximately two-thirds of these people had had sex with someone they met online in the last year, equating to 1.95% of the population.

These nationally representative estimates of looking for and having had sex with someone met on website and smartphone apps are lower than those of surveys focusing on specific subpopulations and using convenience-sampling frames, which have reported 6%-40% of the population meeting sexual partners using websites [2,5,11,12] and 18%-76% using dating applications and websites [3,13-18]. The difference is almost certainly related to different populations sampled and may also be related to the fact that ASHR2 asked specifically about having sex with someone, rather than only looking for partners or meeting in person. Age was strongly correlated with having had sex with someone met on a website or a smartphone app, especially since younger people tend to have higher levels of mobile phone and Internet access, but this study included a broader range of ages than many others. Just over a third (36.07%) of people who used internet dating and hookup applications in the last year reported having had sex with someone they met on a website or using a smartphone app in the last year. Certain populations were less likely to report having had sex with someone met using these technologies. For example, 30.89% of heterosexual respondents who reported that they had used the Internet or a mobile phone app to search for a potential partner in the last year reported having had sex with someone they met online, compared with 62.41% of gay, lesbian, and bisexual survey respondents (see Table 2). This suggests that other studies reporting on the use of internet dating and hookup apps as a proximal marker of having had sex with someone met online are likely to overestimate the practice’s prevalence among lower-risk segments of the population.

In this survey, 15.58% of gay, lesbian, and bisexual respondents reported meeting a sexual partner in the last year. Higher uptake of finding partners using websites among nonheterosexual respondents was also observed in a British population survey [4]. Yet, in our study, after adjusting for age and other characteristics, the strongest correlate of having had sex with someone met using a website or smartphone app was higher numbers of sexual partners in the last year, suggesting that these technologies are favored by the most sexually active respondents. This finding is consistent with other studies reporting that people who look for partners with these tools have increased sexual activity compared with nonusers, including younger age at first sex [2,5] and higher numbers of sexual partners [2,4,5,12,15,16]. Due to the cross-sectional nature of the survey, determining causality is not possible, so findings could mean platforms provide an efficient means for more sexually active individuals to connect with new partners, or alternatively, people who were already more sexually active were attracted to these sites and other ways to meet sexual partners.

In general, people who met partners using websites and apps and had sex with them were more likely to engage in higher-risk practices than those who did not—except for condom use at the last sexual event, which was higher among people who met partners online. There was attenuation of the condom use variable in the multivariate analysis, meaning that the association was not significant in the adjusted analysis after controlling for the numbers of sexual partners and other demographic factors. Higher levels of condom use at the last event could reflect condom use with newer and less established partners, with whom STI prevention is prioritized. This explanation seems highly plausible because online tools are often used to facilitate new sexual partnerships and those who report using websites and apps to find sexual partners also report higher numbers of recent sexual partners. However, the finding contrasts with many other studies that tend to find meeting or seeking partners online is linked to condomless sexual intercourse [4,5,14,15].

In relation to STI history, both STI testing and diagnoses were higher among people who reported meeting partners using websites and smartphone apps and having sex with them. Again, adjusted analysis showed attenuation of STI history for women after adjusting for the number of sexual partners. This implies that women who reported either an STI test or diagnosis in the last year were also more likely to have multiple sexual partners in the last year. The relationship between STI history and meeting partners using websites and apps remained significant for men, even after adjusting for partner numbers and other demographic factors. Interestingly, this pattern has also been seen in the other two population studies with a significant relationship between STIs and searching for or meeting partners on websites and smartphone apps among men, but not among women [4,5]. Findings from other studies have been mixed as to whether STI diagnoses are related to either finding or searching for sexual partners on internet sites and geosocial apps [1-4,16,19]. A number of studies have reported on HIV testing history among gay men and tend to find that men who use apps are more likely to have been tested for HIV [14,19]. Studies with gay men in Australia suggest that men who use a combination of mobile phone apps, internet websites, and offline places to meet partners appear to be at increased risk of STIs or HIV compared with men who use a narrower range of online and offline methods [20].

In women, low annual household income was associated with meeting partners on websites and apps. Socioeconomic deprivation has been linked with poor health outcomes, including STI acquisition [21-23], and other reports from ASHR2 found lower income related to multiple sexual partners [24]. Aside from age and income level for women, other sociodemographic factors, smoking, high alcohol consumption, and injecting drug use were not associated with having sex with someone met online.

To our knowledge, this is the first study to report the prevalence of having had sex with someone met on a website or a mobile phone app in the past year from a representative survey of the general adult population. One of this study’s strengths is the capacity to assess the proportion of people who used internet dating and apps, met in person, and had sex with someone met online within the same population. Nonetheless, our study also has several limitations to consider when interpreting findings. First, the study was conducted in 2012-2013, and since that time, the technology landscape and behaviors related to the uptake of technology have changed. Very likely, the uptake of dating and hookup apps has substantially increased since the survey was conducted. Second, the sample of homosexual and bisexual men was not sufficiently large to enable analyses focused on this group. Third, all outcomes including STI outcomes were based on self-report, which is susceptible to recall and other reporting biases. Several similar studies have used biological measurement to ascertain STI prevalence, a more robust measure [2,4,5]. However, when asked, 88.55% of participants reported answering all survey questions truthfully, and a further 9.89% reported that they had answered 90%-99% of the survey honestly. This equates to 98.44% of participants answering at least 90% of questions truthfully [7]. Furthermore, the clear majority (89.97%) of participants reported that they were either not embarrassed or only slightly embarrassed by the questions. Evidence of the relatively low embarrassment and discomfort with questions related to sexual practices is seen in the number of people declining to answer particular questions. The question with the highest rate of refusal involved annual income compared with much lower refusal rates for questions about sexual practices .

Understanding the number and characteristics of people most likely to use these technologies to meet new sexual partners assists organizations responsible for HIV and STI prevention programs in identifying places and populations wherein they can focus their health promotion and testing initiatives. Our study has also demonstrated that although the prevalence of having had sex with someone met on a website or a smartphone app was 3.03% in people aged 20-29 years, it remained at 2.25% in people aged 30-49 years, suggesting the need for promotional material to cover a broad range of ages, not just younger adults. Although STIs are most prevalent in people aged 16-29 years, recent studies have suggested an increased rate in people over 30 years [25]. Furthermore, despite the finding that people who met partners on dating websites and apps were more likely to have had an STI test in the last year, a substantial proportion of respondents who reported having sex with someone they met online (likely to be a new sexual partner) reported not having an STI test in the past year, suggesting an opportunity to raise awareness about STI testing further and to use targeted advertisements to direct people to easy access points such as new websites where pathology request forms can be downloaded without attending a clinic [26]. Some have also suggested that these platforms have the capacity to enable partner notification and data collection in relation to sexual health [27]. Notably, however, the owners of dating websites and apps have historically been concerned about associating their platforms with STIs and therefore are reluctant to promote public health initiatives [27].

Internet and smartphone technologies are a relatively common way of meeting new sexual partners among highly sexually active survey respondents, homosexual and bisexual men, and younger adults, suggesting that the use of in-app health promotion is a feasible approach to targeting these populations [28,29]. Future research could explore the potential of health promotion in dating websites and geosocial applications. The use of smartphone technologies to search for potential sexual partners may become a normative dating practice among Australian adults in their twenties and thirties, so repeat surveys would be important to document this prevalence over time.