Article Text
Abstract
Objectives The objectives are to analyse social determinants of sexual health behaviour (condom use and HIV testing) among young, internet-active, cis men who have sex with men (MSM) in a high-income country. The aspects of sexual health behaviour analysed here are condomless anal intercourse with one or more new or casual partner(s), condomless anal intercourse during the most recent sex with a man and HIV testing.
Methods A randomised sample of men active on Sweden’s main online community for Lesbian, Gay, Bisexual and Trans people responded to an online survey (response rate 19%). A subsample of young people, aged 15–29, was analysed (effective sample 597–669) using multivariable logistic regression with respect to factors associated with condomless anal intercourse with one or more new or casual partner(s), condomless anal intercourse at most recent sex and not having had a test for HIV.
Results Low education, being single and living in a metropolitan area were found to be independently associated with condomless anal intercourse with new or casual partner(s). Sex with a steady partner was associated with condomless anal intercourse during the most recent sex. Knowledge of where to get tested, high education, being born outside Sweden and condomless anal intercourse with new or casual sex partner(s) were independently associated with having been tested for HIV.
Conclusions The factors associated with sexual health behaviour among young MSM are complex, and preventive messages need to be tailored accordingly.
- adolescent
- condoms
- gay men
- HIV testing
- public health
Statistics from Altmetric.com
Introduction
In high-income countries, HIV incidence and prevalence remains high and is in some cases increasing among men who have sex with men (MSM), while decreasing in the general population.1 One explanation is that the biological risk of HIV transmission is considerably higher for anal intercourse than for vaginal intercourse, when no protection is used.1 Thus, compared with the general population, MSM have a higher risk of meeting an HIV-positive sex partner (due to the higher prevalence), and a higher risk of HIV transmission at intercourse with a HIV-positive male sex partner, unless protection is used.1
Two important aspects of sexual health behaviour thus are using protection for intercourse, and regular HIV testing to ensure prompt treatment.2 Though nowadays, antiretro virals and pre-exposure prophylaxis (PrEP) are viable ways of protection against HIV infection,3 the focus of this article is condom use for anal intercourse.
Research indicates that, similar to young people in general, adolescent and young adult MSM are more likely to have condomless intercourse4 and less likely to have had an HIV test,5 6 compared with adult MSM. Thus, it is highly relevant to explore determinants of these health behaviours among MSM in the younger age range.
Moreover, it is important to analyse social determinants of health so as to prevent social inequity in health.7 Socioeconomic factors are a relevant predictor of health behaviour among young people, but less consistently so than among adults.8 However, socioeconomic determinants of behaviour have not been fully explored among young MSM in relation to sexual health.
Though educational level has been shown to be related to condom use among adult MSM,9 10 we could identify only one study on this among young MSM, which showed that educational level was not related to condom use among young MSM.11 ,i
Intercourse with a steady partner is less likely to be with a condom,4 12 but condomless intercourse with a casual partner is generally considered to be more risky with regard to HIV transmission, per encounter.1
Factors associated with never having been tested for HIV, as indicated by previous studies among MSM, can be younger age, lower education, low knowledge of HIV and living in a small town or rural area.4–6 13 14 These factors are correlated and could confound each other, and thus need to be analysed in relation to each other.
Country of birth is also relevant to analyse both in relation to condom use and HIV testing, since immigrants in general, and immigrant MSM, tend to be over-represented among newly diagnosed cases of HIV, in Sweden15 16 and across Europe.17
Given this background, the present study aims to analyse social determinants of sexual health behaviour, that is, condomless anal intercourse with any new or casual partner(s) or at most recent sex with a man, and HIV testing, among young, internet-active, cis MSM in a high-income country, Sweden.
The context
Sweden (with a population of approximately 10 million) has a low general HIV prevalence (~0.07%).15 On average, about 125 new HIV cases in Sweden are annually reported as being due to sexual transmission between men; of which approximately 70 cases where transmission was estimated to have occurred abroad, either before migration to Sweden or on vacation abroad (personal communication from the Swedish Public Health Agency, 2018-01-15). MSM born outside of Sweden are over-represented among new cases in Sweden.16 In the 2013 MSM survey, about 4% of respondents reported being HIV positive; increasing with age from 0.8% among the youngest to 6.8% among the oldest.16 However, the percentage who were uncertain of their HIV status was highest among the youngest respondents.16
Adolescents and young adults can get counselling and HIV testing at youth friendly clinics called Youth Clinics (Ungdomsmottagning), but estimates show that 85%–90% of the patients visiting these clinics are girls.18 The metropolitan areas (Stockholm, Gothenburg, Malmö) have special MSM health clinics, which are very popular in all age groups. The largest MSM clinic, in Stockholm, receives about 8.000 visits per year (all ages) (personal communication from the Swedish Public Health Agency, 2018-01-15).
Same-sex marriage has been legal in Sweden since 2009 and same-sex civil union since 1995. Social acceptance for sexual and gender minorities is comparatively high.
Methods
Participants and data collection
The 2013 MSM survey (MSM-enkäten) was conducted by the Public Health Agency of Sweden in collaboration with the Department of Public Health Sciences, Karolinska Institutet, among a randomised sample of members of the online social community Qruiser QX, which is the largest platform for Swedish Lesbian, Gay, Bisexual and Trans people (LGBT) (www.qruiser.com). Based on population surveys, the Public Health Agency of Sweden estimates that there are 80 000–120 000 MSM aged 15+ years in Sweden.16 Qruiser has about 50 000 registered profiles that are men aged 15+ years.
The following inclusion criteria were used: profiles with age registered 15+ years, not female, living in Sweden. From the eligible sample base, duplicate profiles were removed, resulting in 52 979 Qruiser profiles. Of these, 14 514 profiles were selected through stratified random sampling, based on age groups and regional counties, as registered in their profile. Although the sampling algorithm was prepared by a statistician at the Public Health Agency, the profiles were selected by staff at Qruiser. In other words, for confidentiality reasons, the researchers were not provided direct access to the membership register.
The selected profiles received an invitation via their Qruiser inbox, which included information and a link to the online survey. The response rate was approximately 19% both in the whole sample and in the stratum aged 15–24 years. The low response rate may be due to the fact that around 40% of the selected individuals did not log into Qruiser during the month the survey was conducted (October 2013). After excluding certain responses (see figure 1), the response rate was approximately 16%. The survey has been described in detail elsewhere.4 16 19–22
In the current study, 763 respondents between ages 15 and 29 years were included. The age of the respondents was obtained either from their survey response (n=760), or if missing, from their Qruiser profile age (n=3) as reported by Qruiser staff. The age cut-off was based on the advice from practitioners working with young MSM in Sweden.
Due to listwise deletion of missing cases, the effective sample for multivariable analyses varied between 597 and 699 (for drop-out analyses, see below).
Ethics
The survey was approved by the Regional Ethics Review Board in Stockholm, including inclusion of respondents below age 18 but above age 15 years (Dnr 2013/3:3). The introduction page of the survey clearly stated the following, twice: ‘Participation is voluntary, and you may withdraw from the survey at any time.’ Contact information to researchers was given at the beginning and end of the survey. At the end, a list of contacts to healthcare and social care for further support was provided. The respondents cannot be traced back to their Qruiser profile (as all profile management was performed by Qruiser staff).
Measures
Sexual health behaviour
The respondents were asked to list the number of new or casual sex partners with whom they had had unprotected anal intercourse within the previous 12 months (in Sweden or abroad). The word ‘unprotected’ was used in the questionnaire, but it was implied to mean ‘condomless’ (this was before availability of PrEP in Sweden). Thus ‘condomless’ will be used for the rest of the article, rather than ‘unprotected’. The variable was dichotomised into zero versus one or more partners. Internal drop-out was high, but did not differ in terms of age, geographical area or self-reported frequency of logging in to Qruiser. However, the internal drop-out was significantly related to HIV testing. Therefore, in order to avoid skewed results due to biased drop-out, the response alternative ‘Don’t know or can’t remember’ (n=18) and the category of respondents who did not respond to this question (n=82) were included as categories when using the measure as an independent variable in relation to HIV testing.
With respect to the most recent sex with a man (whether in Sweden or abroad), condom use was dichotomised to denote condomless anal intercourse versus no condomless anal intercourse. The former could be anal intercourse either completely without a condom or using condom only part of the time (including condom failure); while the latter meant either anal intercourse with condom for the whole time or sex activities other than anal intercourse.
The young respondents had generally either tested recently or not at all, hence HIV testing (see table 1) was dichotomised to ‘Recent HIV test’ or ‘No HIV-test within the previous 12 months’. The non-responders to this question (n=68) were younger (mean age 22.1 years vs 23.2 years among responders, p=0.025) but did not differ in terms of condomless intercourse with new or casual partner(s), number of sex partners in general, geography or frequency of Qruiser login. These non-responders were excluded from analyses.
Independent variables
The survey had 10 questions on knowledge and skills regarding sexual health and relationships, based on questions used in previous Swedish studies16 23 and developed in collaboration between practitioners, public health planners and LGBT advocacy groups. The responses were rated on a 5-point scale from ‘Very good’ (=5) to ‘Very bad’ (=1). Five items (see table 2) were chosen for the analysis, following principal component analysis, which revealed these five items to form one factor (Cronbach’s alpha 0.812). These items concerned knowledge of HIV/STI, while the excluded items mainly concerned skills regarding relationships and communication. These five items were summarised using the mean score for participants with ≥2 non-missing items and were used as a linear variable.
Being born outside of Sweden was dichotomised such that international adoptees were categorised together with Swedish-born individuals. Educational level was dichotomised into higher than high school education versus high school or lower. Other included variables are presented in table 1.
Analyses
Descriptive statistics are presented in table 1.
Multivariable binomial logistic regression was used to analyse educational level, age group, living in a metropolitan area, relationship status, HIV status and born outside of Sweden, in relation to the outcome condomless anal intercourse with new or casual partner(s) (table 3) and in relation to the outcome condomless anal intercourse during the most recent sex (not in table). In the analysis of condomless anal intercourse during the most recent sex, a variable describing the relationship to that specific sex partner was included, to control whether associations were explained by general relationship status or the relationship to the specific sex partner. Education, age group, metropolitan area, condomless anal intercourse with new or casual partner(s), born outside of Sweden, knowledge of where to get tested and HIV-related knowledge were analysed in relation to not having had an HIV test within the previous 12 months (table 3).
Weights were available to correct for the sample design and drop-out per stratum.19 The regression analyses were unweighted; our subsample of young respondents had several strata containing few or single respondents, which we considered a problem for weights constructed for the whole sample. However, we also present the results from the weighted analyses (descriptive weighted analyses in brackets in table 1, multivariable weighted analyses described in footnotes to tables) using STATA’s svy command, treating single units as certainty units.24
Results
As shown in table 1, 48% reported no condomless anal intercourse with new or casual partner(s) in the previous 12 months. With respect to the most recent sex with a man, 37% reported condomless anal intercourse (receptive or insertive anal intercourse without condom or not using condom the whole time), 22% reported anal intercourse using a condom the whole time and 39% reported sex activities other than anal intercourse. Forty-two per cent of respondents had gotten tested for HIV within the previous 12 months (28% in the last 6 months).
Lower education, being unsure of one’s HIV status, living in metropolitan areas and being single were independently associated with having reported condomless anal intercourse with new or casual sex partner(s) in the past 12 months (table 3).
At the most recent sex occasion with a man (data not in table, n=699), singles were significantly less likely to have had condomless intercourse compared with respondents in relationship with a man (adjusted OR 0.53, p=0.001). However, after controlling for type of sex partner at that specific occasion, the relationship status was found to be no longer significant (adjusted OR 0.79, p=0.376). Instead, what mattered was whether the sex partner at that time was a boyfriend or husband (adjusted OR 2.10, p=0.004) or a steady lover (adjusted OR 1.93, p<0.001) compared with a non-steady sex partner. Age, education, country of birth or living in a metropolitan area were not related to condom use at most recent sex in the multivariable analysis. Weighted analyses gave the same conclusions.
Having low education, being born in Sweden and being unsure or only quite sure about where to get tested for HIV were independently associated with not having gotten tested for HIV (table 3). Living in a metropolitan area and self-reported high knowledge had bivariate associations with HIV testing, but were explained by the other variables in the full model.
Reporting condomless anal intercourse with at least one new or casual male partner in the past 12 months or not remembering how many partners generated lower odds of not having gotten tested for HIV (ie, higher likelihood of having gotten tested), while complete non-response was associated with not having gotten tested.
Discussion
In casual sexual encounters, young MSM with low education or living in metropolitan areas were more likely to have had condomless anal intercourse. However, with respect to the most recent sex encounter, only stability of the relationship to the partner was related to condomless intercourse. For HIV testing, the level of education, knowledge of where to get tested and the awareness that one had had condomless intercourse independently contributed to the tendency to have had an HIV test.
It is notable that the most common response regarding most recent sex with a man was that there was no intercourse but only other sex activities. This response should be considered in research and practice regarding strategies for safe sex among young MSM. A French survey conducted in 1985–2011 indicated that the proportion of anal intercourse decreased among MSM during the 1980s but subsequently increased again.25 Unfortunately, we have no similar data for Sweden to make a trend analysis.
Based on previous research,9 10 we assumed that higher educated young MSM would be more likely to use condoms compared with those with low education. This was also observed for intercourse with new or casual partner(s) but not for intercourse at most recent sex, where there was no difference by education. This finding is in contrast with one previous study (among adult MSM), where education was related to condom use both in ‘main’ and ‘non-main’ relationships.9
Though singles reported more condomless anal intercourse with new or casual partner(s) compared with non-singles, sex with a boyfriend, husband or steady lover was significantly more likely to involve condomless anal intercourse. This finding is consistent with the findings of previous research.4 9 12 Our study did not differentiate between failed or inconsistent condom use versus no condom use. D’Anna et al 9 showed that both condom use and breakage or slippage of condoms were more common in ‘non-main’ relationships.
In steady relationships, the partners are more likely to know the HIV status of each other and, if serodiscordant, they may be protected through medication instead of condoms.3 Some studies have indicated that MSM may be more likely to get HIV from a steady partner in the long term, due to higher frequency of sex and lower condom use. However, it should be noted that these studies were conducted before the availability of PrEP,12 26 which was also the case for the current study.
The geographical pattern observed in the results—that respondents living outside the metropolitan areas were less likely to have gotten tested for HIV, but also less likely to report any condomless anal intercourse with new or casual sex partner(s)—could be due to those respondents being younger, less educated and having less access to testing facilities.4 For HIV testing, the geographical patterns were explained by the other variables; a relevant question for further analysis is thus to what extent the MSM clinics in metropolitan areas contribute to the higher testing rate there. One of the open responses in the survey was from a man in a rural county who reported feeling questioned when he asked for an HIV test at the primary healthcare centre there. This is only anecdotal of course, but merits further investigation.16 For casual condomless anal intercourse, geographical differences remained after controlling for age and education. Local networks of MSM outside metropolitan areas could be smaller, which could imply fewer potential new or casual sex partners.
Young MSM born outside of Sweden were more likely to have gotten tested for HIV compared with Swedish born. This could partly be explained by the fact that asylum seekers are offered HIV testing as part of the health check-ups required by European Union (EU) policies for migration.27 A study among the respondents born outside of Sweden in this sample (all ages) revealed that having arrived in Sweden within the past 5 years was a significant predictor of having gotten tested in the past 12 months.22 It is also possible that this sample of Qruiser-active young MSM born outside of Sweden was a special sample, with for instance strong LGBT social networks and information that is transferred between peers about testing. A previous study on the same survey material, with weighted analyses on all ages, found no associations between country of birth and HIV testing.19
Discussion of method
Random samples among MSM are rare since there are no clear sample frames. The random sample taken from Qruiser is a strength, compared with self-selected or venue-based surveys, which are more vulnerable to selection bias. Since Sweden is a relatively small country with one large LGBT social website and high internet usage, it was possible to reach a large sample representing a majority of MSM in Sweden. Young men might be more active online than older men,28 and the sample may thus be more representative for the younger population of MSM in Sweden. However, the response rate was similar among the young respondents years compared with the whole sample. It is also possible that young MSM use other online platforms that we are not aware of. Singles may be more active on social media,28 and MSM with stable relationships may be under-represented. People who are less sexually active may be less comfortable in answering intimate questions about sex. The questionnaire was pilot tested with Sweden’s largest LGBT organisation RFSL (The Swedish Federation for Lesbian, Gay, Bisexual, Transgender and Queer Rights) and is based on previous MSM surveys. Open comments at the end of the questionnaire generally expressed positive opinions about the survey.
Due to the cross-sectional nature of this study, causality can only be indicated, not tested.
HIV risk at condomless intercourse is greatly reduced if the HIV-negative partner is taking PrEP and/or the HIV-positive partner has an undetectable viral load.3 The survey was performed in 2013, and the first PrEP was approved in the EU in 2016,29 so PrEP was not an option for these respondents. Data on viral load was not available and could not be controlled for.
Conclusions
Patterns of factors associated with sexual health behaviour among young MSM are complex, and depend both on personal and structural factors. Prevention should be tailored accordingly, taking into account relationship status, educational level, geographic location, age and accessibility to testing facilities.
The implications for practice and policy include developing existing youth clinics, sex education and general primary healthcare to be inclusive of young LGBT people, especially young MSM, and increasing awareness among healthcare staff and planners, regarding the facts that young MSM would benefit from needs-based preventive interventions and access to testing and counselling.
Key messages
Condom use and HIV testing were analysed in a random sample of young men who have sex with men (MSM, ages 15–29) from a Lesbian, Gay, Bisexual and Trans people online community.
For casual sex, low education or urbanicity was associated with condomless anal intercourse; but at most recent sex, only being in a steady relationship was.
Geographical differences in HIV testing among young MSM were explained by age, education and access to a testing facility.
Prevention could benefit from a needs-based approach, taking into account relationship status, educational level, geographic location, age and accessibility of testing facility.
References
Footnotes
↵ i The referenced study, however, did find an association between educational level and condom use among young respondents that the authors identified as transgender (based on how the respondents presented and the venue where they were recruited). But the authors gave no information on how these respondents themselves identified (transgender or not, and if so, trans women, trans men, or other categories), and thus those results were hard to interpret.
Handling editor Stefan Baral
Contributors KJ conceived the article, did analyses and drafted the manuscript. KIP participated in data collection and formulation of research question, and critically reviewed the manuscript. CD participated in formulation of research question, and critically reviewed the manuscript. ZEK conceived the article, did analyses and participated in drafting the manuscript. All authors read and approved the final manuscript.
Funding The study was funded by the Public Health Agency of Sweden (grant number 05434/2014), and the data collection was performed by the Public Health Agency in collaboration with Karolinska Institutet and Qruiser QX.
Competing interests The MSM survey was designed and conducted by the Public Health Agency of Sweden, while the analyses and writing of this substudy were done at Umeå University, funded by the Public Health Agency. The funding agency had insight into but no decision over the choice of research question, the analyses, the writing or decision of where to submit. KJ also has ongoing consultancies with the Public Health Agency of Sweden on sexual health issues, unrelated to this study, but her work on this study was done through employment at Umeå University, funded by the Public Health Agency. KIP and CD have employments at the Public Health Agency of Sweden, but participated in this study as part of their affiliations with Karolinska Institutet. Since the Public Health Agency has no vested interest in biasing the results or the research questions in any direction, the close connection with the funding agency should not constitute a problematic conflict of interest. ZEK has nothing to disclose.
Ethics approval Regional Ethics Review Board in Stockholm (Dnr 2013/3:3).
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The data from the MSM suvey 2013 are under the responsibility of the Public Health Agency of Sweden.