Sexual behaviour during COVID-19: a repeated cross-sectional survey in Victoria, Australia

Keywords: Australasia, COVID-19, dating apps, pandemic, self-report, sexual behaviours, sexual practices, sexually transmissible infections, survey.

The COVID-19 pandemic and ensuing responses toward reducing virus transmission have affected communities across the world. During 2020, a total of 28 624 COVID-19 cases (113 per 100 000 population) were diagnosed in Australia.¹ The majority (71%) were in Victoria (Australia’s second most populous state)¹ where two COVID-19 waves occurred. Peaking in March, the first wave affected all of Australia and involved largely returned travellers and their close contacts.^2,3 Victoria’s second wave saw a total of 18 668 cases and was characterised by widespread community transmission and multiple outbreaks in health and aged care settings.^4–6

During the first wave, a range of measures were implemented toward reducing COVID-19 transmission. The Australian Federal Government implemented a national lockdown (termed ‘Stage 3 restrictions’) across the whole of Australia that involved limits on the size of gatherings, closure of non-essential business, and closure of international borders.^7,8 Australians were asked to stay home except for essential shopping, exercise, medical care, and work/education that could not occur at home. These national measures were gradually eased from May 2020.⁹ Domestic travel was also restricted by border closures implemented by states and territories at varying times throughout 2020.¹⁰ By July 2020, Victoria was experiencing a rapidly growing second wave and in response, the State Government in Victoria introduced ‘Stage 4 restrictions‘ (also termed the hard lockdown) that in addition to ‘Stage 3 restrictions‘ involved a nightly curfew, mandated facemasks, and a 5 km travel limit.^11,12 Victoria’s hard lockdown extended for several months and did not apply to the rest of Australia.

We conducted a repeated online survey at four key timepoints during 2020 to investigate the sexual and reproductive health of people living in Australia during the COVID-19 pandemic. While the survey was open to anyone throughout Australia, we limit the analysis here to those living in Victoria because our survey waves were aligned with two periods of lockdown and two periods of eased restrictions that followed the lockdowns in Victoria. Here, we examine the variation in sexual behaviour of people living in the state of Victoria during four waves of surveys.

We undertook a serial cross-sectional analysis of online survey data collected at four distinct time-periods during 2020 to investigate variation in sexual behaviours of Victorians during changing restrictions. Our four survey waves were implemented to correspond with two lockdown (LD1, LD2) and two reduced restriction (RR1, RR2) periods in Victoria (Box 1). Surveys 1–3 were open for approximately 2 weeks and Survey 4 for approximately 3 weeks due to a slower response. Although our survey was open to all Australians, this analysis is limited to people living in Victoria at the time of survey completion.

The methods for Survey 1 have been described previously¹³ and were repeated for subsequent surveys. For all four surveys, people aged 18 years or older and living in Australia at the time of the survey were eligible. Participants were recruited by circulating recruitment flyers via researcher’s networks, snowballing and social media including twitter and paid Facebook advertisements. Interested individuals could click a link to the survey where they were provided with an explanation of the survey, eligibility criteria, and how to consent if they chose. Each survey asked about sociodemographic information (including state of residence) and the respondents’ current (defined as 4 weeks prior to survey completion) sexual health experiences and practices (e.g. have you had sex in the past 4 weeks). Sex was defined as physical contact with other people for sexual pleasure. Respondents were also asked if they had changed the frequency (less often, the same amount, more often or stopped) they engaged in specific sexual practices (e.g. using sex toys alone, masturbating alone) compared with before the COVID-19 pandemic.

Binary outcomes were created for recent sexual partner type categorised as regular sex partners (spouse, partner, boyfriend, girlfriend), casual sex partners (hook-up, casual dating) and occasional sex partners (friend, ‘fuck buddy’, occasional sex partner). Binary outcomes for dating app use were defined as using apps for (1) chatting; (2) for physically distanced activities, which we included responses for sexting, sending pics or videos, or virtual dates; or (3) for face-to-face activities, which we included responses for dates or hook-ups. Because restrictions may limit how people interact with each other, we were interested if the frequency that people engaged in solo sex practices had changed and created binary outcomes for using sex toys alone more often and masturbating alone more often compared to before COVID-19. Categorical variables were created for age group (18–29 years, 30 years or above), gender (male, female, gender diverse), employment status (full time, part time, unemployed and looking for work, other), sexual identity [heterosexual, men who have sex with men (MSM), women who have sex with women (WSW)], relationship status (no steady relationship and no recent sex, no steady relationship and having sex, in a relationship and not cohabitating, in a relationship and cohabitating) and area of residence (metropolitan, rural/regional).

We examined variation in sexual behaviours over our four waves of data collection. Sociodemographic characteristics of respondents between surveys were compared using descriptive statistics. Logistic regression was conducted to see whether sexual partner type, dating app use and solo sex activities differed from our reference category of LD1 (Survey 1), adjusting for sociodemographic characteristics to account for changes in the participant profile over time. For solo sex activities, we investigated whether gender modified the association between survey period and our solo sex outcomes by comparing logistic regression models with and without an interaction term between gender and survey period using the likelihood ratio test, and report these if an effect modification was found. Odds ratios (OR) and 95% confidence intervals (95% CI) are reported. All statistical analyses were undertaken in Stata SE V.16 (StataCorp LLC).

The study was approved by the University of Melbourne Human Research Ethics Committee (ID 2056693).

Overall 2258 surveys were initiated by Victorian residents; of these 81.0% (n = 1828) completed the survey (Survey 1: 85.4%, n = 485, Survey 2: 75.2%, n = 521, Survey 3: 78.0%, n = 606, Survey 4: 98.2%, n = 216). Participant characteristics are provided in Table 1. The characteristics of participants varied over time. More females completed Surveys 2 (79.6%) and 4 (73.1%) than Surveys 1 (69.7%) and 3 (65.9%), more younger participants completed Surveys 2 (75.8%) and 3 (71.6%) than Surveys 1 (61.9%) and 4 (59.5%), and more participants in a relationship but not cohabitating completed Surveys 2 (30.0%) and 3 (27.0%) than other surveys. Most participants were Australian-born and lived in metropolitan areas.

Table 1.  Sociodemographic characteristics of participants, by survey.

The proportion of respondents reporting partnered sex in the past 4 weeks ranged from 54.9% during LD2 (Survey 3) to 70.2% during RR1 (Survey 2), with a median number of sexual partners of 1 (IQR 1–1) for each survey. Multivariable analysis (Table 2) found that compared with LD1 (Survey 1), respondents were more likely to report partnered sex in the past 4 weeks during RR1 (Survey 2) [adjusted odds ratio (aOR) 1.7; 95% CI 1.3–2.3], with no other differences observed.

Table 2.  Types of sexual partners, dating app use, and changes in sexual practices, by survey.

Among respondents reporting partnered sex, the most common sexual partners were regular partners (range 77.4–86.1%; Fig. 1). Compared with LD1 (Survey 1) multivariable analysis found that during periods of reduced restrictions (Surveys 2 and 4), participants were less likely to report partnered sex with regular partner(s) (RR1: aOR 0.5; 95% CI 0.3–0.8; RR2: aOR = 0.5; 95% CI 0.3–0.9), and, more likely to report partnered sex with casual partner(s) (RR1: aOR 2.0; 95% CI 1.1–3.7; RR2: aOR = 2.8; 95% CI 1.3–5.9). The proportion reporting partnered sex with occasional sex partner(s) did not differ by survey period.

Fig. 1.  Type of sexual partners reported at each survey period.

Between 25.4% and 29.6% of respondents reported using dating apps in the past 4 weeks. Multivariable analysis found that this did not vary between surveys (Table 2). Among dating app users, chatting was the most common reason for use (range 73.2–93.5%; Fig. 2). Multivariable analysis found that compared with LD1 (Survey 1) chatting was less likely in all other periods (Surveys 2–4; Table 2). Use of dating apps for face-to-face activities was lowest during LD1 (Survey 1) and LD2 (Survey 3) at 19.5% and 14.9%, respectively. Multivariable analysis found that compared with LD1 (Survey 1), dating app use for face-to-face activities was over four times more likely during periods of reduced restrictions (Surveys 2 and 4) (RR1: aOR 4.3; 95% CI 2.3–8.0; RR2: aOR 4.1; 95% CI 1.9–8.8). The proportion using dating apps for physically distanced activities was highest during LD1 (Survey 1) at 48.8% and lowest during RR2 (Survey 4) at 25.0%. Multivariable analysis found that compared with LD1 (Survey 1) that dating app use for physically distanced activities was less likely during RR2 (Survey 4; aOR 0.3; 95% CI 0.2–0.8).

Fig. 2.  Reasons for dating app use reported at each survey period.

While up to a half of respondents reported never using sex toys alone, 14.5–19.9% reported using them more often than before COVID-19. Between 43.6% and 53.2% of respondents reported that they had not changed how often they masturbated alone while 26.0–34.3% reported doing this more often than before COVID-19. Compared with LD1 (Survey 1), respondents during LD2 (Survey 3) were more likely to report more frequent use of sex toys alone (aOR 1.4; 95% CI 1.0–2.1) and masturbation alone (aOR 1.4; 95% CI 1.0–1.9).

This study found self-reported sexual activity and behaviours for a sample of Victorian Australians fluctuated during 2020 as COVID-19 restrictions were implemented and eased. We found that recent partnered sexual activity was reported by 55–57% of respondents during lockdowns and by 61–70% of respondents as restrictions were eased. Lockdowns were associated with more frequent physically distanced activities (e.g. virtual dates). Less restrictive periods saw respondents more likely to report casual sex partners and using dating apps to support face-to-face activities.

To date, there have been few if any studies investigating trends in sexual behaviours since the COVID-19 pandemic began. Aligning our survey waves with four distinct timepoints in the COVID-19 response provided us a unique opportunity to capture the experiences of Australians living in Victoria as they were subject to tightening and easing of restrictions. Our study is strengthened by our multivariable analysis where we were able to adjust for sociodemographic changes in participants across surveys, which may have confounded our findings. However, there are several limitations to consider. First, our recruitment methods provided a convenience sample that is unlikely to be representative of Victoria’s diverse population groups. When compared with population data, our participants were more likely to be female, Australian-born (80% vs 65%),¹⁴ university educated (67% vs 18%), and 19% identified as WSW (vs 3.4%).¹⁵ Victoria’s second COVID-19 wave disproportionately impacted those who were born outside of Australia, lived in more disadvantaged areas and worked in lower paid employment.¹⁶ Our findings need to be considered in the context of our sample. Second, the characteristics of our sample varied over time which despite adjustment in our multivariable analysis may have contributed to our findings. Third, survey responses were based on recall and may have been subject to recall bias, particularly questions about activities in 2019. Finally, our fourth survey received a lower number of responses and a higher completion rate than the preceding three surveys. Although we do not know the exact reasons for these differences, it is possible that when restrictions were eased after Victoria’s extended lockdown, many potential respondents were enjoying freedom to socialise rather than completing an online survey. Also, a higher proportion of respondents to Survey 4 were university educated compared to earlier surveys, potentially contributing to the higher completion rate.

The changes in self-reported sexual behaviours observed in this study are consistent with a growing evidence base of the impacts of COVID-19 on sexual behaviours. During Australia’s first nationwide lockdown (March–May 2020), responses to our first survey indicated that many people reported reduced partnered sexual activity and increases in solo activities compared to before COVID-19.¹³ Also during this period, substantial reductions in sex with casual partners were reported by gay and bisexual men in Australia.¹⁷ In the United Kingdom, a large quasi-representative survey found after 3 months lockdown that half of participants reported no change in partnered sex frequency and those in non-cohabitating relationships were more likely to report a decrease in partnered activity and an increase in non-partnered activities.¹⁸ Declines in partnered activities and increases in masturbation alone during COVID-19 restrictions have also been reported in several German speaking countries, China and the United States,^19–21 in particular for males.²¹ Our observed changes over time did not vary by gender identity.

Notably, in this study we found that patterns of sexual activity fluctuated over time as lockdowns were implemented or subsequently eased. Regular sexual partners were most commonly reported across all surveys but less frequently during reduced restrictions periods. Conversely, casual sexual partners were reported more often when restrictions were eased compared with the first lockdown. These and other reported changes in sexual behaviours¹⁷ could potentially alter patterns in diagnosis of sexually tranmissible infections (STIs). A large metropolitan Victorian sexual health clinic reported marked reductions in attendance and diagnosis of mild conditions (e.g. nongonococcal urethritis) during the first lockdown compared to before and after the lockdown. However, no difference in diagnosis rates for symptomatic STIs were found when comparing pre, during and post-lockdown periods, suggesting that attendance was prioritised based on urgency of need.²² Delays in attending sexual health services since the start of the pandemic have also been reported in Canada, with users of online services less likely to report unmet sexual health needs.²³ Continued availability of sexual health services is vital during the evolving pandemic.

We also observed some small differences in self-reported behaviours between Victoria’s first and second lockdown. Respondents were more likely to report that they engaged in solo activities, particularly masturbation during the second extended lockdown than the first, which could suggest that the duration of lockdowns affected sexual activity differently. However, given the substantial variation in our sample over time it is possible this finding reflects differences in the characteristics of our respondents. Other reports indicate that Victoria’s extended lockdown placed a significant toll on the Victorian population²⁴ including their mental health.²⁵

While overall dating app use remained constant between surveys, the way that respondents interacted with dating apps varied over time. During lockdowns, dating apps were more likely to be used for physically distanced activities such as virtual dates. However, when restrictions were eased, use of dating apps to support face-to-face activities such as organising dates or hook-ups was more common. As the COVID-19 pandemic unfolded during 2020, dating app companies reported an increase in engagement and a shift in how dating apps were used, in particular increased use of video dating functions.^26,27 Dating app companies also responded to the COVID-19 pandemic by including health messages (e.g. advice about physical distancing), promoting virtual dating, and self-care to address loneliness or isolation.²⁸ It is uncertain whether additions to dating apps impacted the findings in this study.

The COVID-19 pandemic and the public health response it prompted was associated with fluctuations in sexual activity and behaviours in this convenience sample of Victorian residents. While this repeated survey provided a valuable means to investigate the experiences of Victorians at key points during the COVID-19 response, it is possible that variation in our sample between surveys contributed to our findings. However, we did find changes in self-reported partnered and solo sexual activity during lockdown and reduced restriction periods. While dating apps can provide a mechanism for virtual connections, this may only be temporary until physical connections are possible. The importance of intimate connections for wellbeing should be considered within public health measures toward reducing COVID-19 transmission.

The data that support the findings of this study are not publicly available due to information that could compromise the privacy of research participants. Please contact the authors to request a copy of the questionnaire.

JSH is an Editor of Sexual Health but was blinded from the peer-review process for this paper. JSH is supported by a National Health and Medical Research Council Fellowship (1136117). All other authors declare that they have no conflicts of interest.

This research did not receive any specific funding.

All authors contributed to the design and development of the survey. JC was responsible for administering the survey. JLG, HB and JSH conducted the analysis. JLG, JC, HB and JSH interpreted the results and drafted the manuscript. All authors contributed to the revision of draft iterations of the manuscript prior to submission.