Article Text

Original article
Men seeking sex online practise riskier sexual behaviours than men frequenting brothels: survey findings from Singapore
1. Mee Lian Wong1,
2. Ting Ting Koh1,
4. Mandy Govender2
1. 1Saw Swee Hock School of Public Health, National University of Singapore, Singapore, Singapore
2. 2Health Promotion Board, Singapore, Singapore
1. Correspondence to Dr Mee Lian Wong, Associate Professor, Saw Swee Hock School of Public Health (MD 3), National University of Singapore, 16, Medical Drive, Singapore 117597; ephwml{at}nus.edu.sg

## Abstract

Objective To compare the prevalence of condom use and its associated factors within commercial and casual sex relationships between heterosexual men who seek sex online and those who frequent brothels in Singapore.

Methods A survey was conducted between 2011 and 2012 among Singaporean heterosexual men aged 21–70 years who reported engaging in sexual intercourse with commercial or casual partners in the past year. Men seeking sex online (n=304) were recruited from the two most frequented local internet sex forums. Men frequenting brothels (n=297, 97% response rate) were recruited from a stratified cluster random sample of licensed brothels.

Results Compared with men frequenting brothels, men seeking sex online reported significantly higher rates of oral and anal sex and sexually transmitted diseases but lower condom use for oral sex with commercial (16.5% vs 56.6%) and casual partners (21.2% vs 35.3%). In multivariate analyses, the online group showed that partner request for condom use was strongly associated with condom use for vaginal sex (adjusted ORs >3.0) but not with condom use for oral sex with both casual and commercial sex partners. The perceived risk of getting HIV from a partner was also associated with condom use for vaginal sex with commercial and casual partners; and for oral sex with casual partners. In contrast, condom use for oral and vaginal sex with both partners in the brothel group was only associated with partner request for condom use.

Conclusions Men seeking sex online should receive sexually transmitted infection (STI)/HIV prevention interventions that deal with the factors associated with condom use.

• Condoms
• Public Health
• Sexual Behaviour

## Introduction

The wide use of the internet has changed the sources of seeking commercial and casual sex partners among heterosexual men. More men are going online rather than to traditional settings, such as brothels, massage parlours, and bars, to seek commercial and casual sex partners.1–3 Previous studies reported that between 4.4%4 and 45.5%5 of heterosexual men went online to meet sexual partners.4 ,5 The reasons for choosing the internet were the accessibility, affordability and anonymity.1 ,3 ,6 ,7

In many countries, heterosexual men engaging in unprotected sex with multiple casual and commercial partners make up large populations who are at high risk of acquiring and spreading sexually transmitted infections (STIs) and HIV to their wives and female partners,8 ,9 yet, few condom promotion programmes are available for them, although there are many programmes targeting female sex workers. It is important to determine the prevalence of unprotected sex among the increasing numbers of heterosexual men who go online to meet sex partners. However, few studies have been conducted on their risky sexual behaviours and they showed mixed findings. One study found that internet sex seekers reported more sexual partners, more anal sex and more exposure to partners known to be HIV positive in the past 12 months than those not seeking sex online,1 whereas another study showed no increase in unprotected vaginal or anal sex at first sexual experience with an online than with an offline partner.10

Asia has over 1.1 billion internet users, representing about half of the world's total internet user population of 2.3 billion.11 Studies in Asia have shown that the internet is becoming a rapidly growing environment for seeking sex.12–14 Although a number of studies12 ,13 have been conducted on online sex seeking among men who have sex with men, little is known about the sexual behaviours of heterosexual men who go online to seek sex partners. An understanding of their sexual risk behaviours will inform the design of interventions for this rapidly increasing group in Asia. In Singapore, there are 4 015 121 internet users,15 representing 85% of households.16 We compared the prevalence of condom use and its associated factors within commercial and casual sex relationships between men who seek sex online and those who frequent brothels in Singapore. We chose the latter as a comparison group because brothels for female sex workers in Singapore, unlike the internet, are regulated and have an established condom promotion programme.

## Methods

The cross-sectional study was conducted between December 2011 and October 2012 among heterosexual men in Singapore. Eligibility criteria were male residents aged 21–70 years who reported engaging in vaginal, oral or anal intercourse with a female sex worker or/and casual partner in the last 12 months preceding the survey. A casual partner was defined as a partner who was not paid for sex and who was not a regular partner or in a longstanding sexual relationship. Condom use was defined as always (consistently) using condoms in the past year. The men were recruited from two types of sex venues—brothels and online sex forums. Based on the level of condom use of 13% for casual sex from the most recent local national behavioural surveillance survey in 2007,17 we estimated that a sample size of 300 per group would yield sufficient numbers (n=39) of consistent condom users for meaningful analysis.

The sample of men patronising brothels was obtained using proportional stratified cluster random sampling of brothels, followed by systematic sampling of men who stepped out from the brothels. We obtained a sampling frame of brothels (n=93) from the department of STI control, from which 61 brothels were randomly selected, taking into account non-response. Next, five clients who met the inclusion criteria from each of the sampled brothels were selected by systematic sampling at different times of the day from 12 noon to 12 midnight and over weekdays and weekends over a 4-month period to reduce selection and seasonal bias. Interviewers approached 310 men, of whom 300 (96.8%) consented and 297 were eligible for the survey. The sampling frame of men seeking sex online was obtained from two online sex forums (sammyboy and laksaboy forums) most frequented by heterosexual men. The two combined websites were found to have 58 000 members. Survey Monkey, a web-based survey tool, was used to create the online version of the questionnaire. We started with a pilot study in which invitations were sent to 50 men through the forums’ private messaging function to participate in the online survey. Only two men responded after 3 weeks. Because of the low response, we sent out invitations to all members of the two online sex forums. We used a web link collector function to permit only one response for each computer. A list of the internet protocol addresses of all responses was also generated to check for duplicate responses. The online survey was closed after 3 months, when the number of responders was approaching 300. Five men who concurrently submitted the completed online questionnaires were included in the sample, bringing the final sample size to 304.

### Data collection

Interviewers, working in pairs for safety reasons, waited outside the randomly selected brothels and requested interviews from men who stepped out from these places. We attempted to reduce self-reporting bias on their sexual behaviours by taking the following measures. First, as interviewers we engaged polytechnic and university students because students were perceived to be less threatening than official staff. Interviewers were trained to approach the men in a polite, non-threatening and non-judgemental manner and explained that they were university students conducting a lifestyle survey among men in Singapore to help plan a programme to improve their health. The interviewers emphasised the brevity of the questionnaire and assured the men that they would not be asking for their names or other personal identifiers. They also tried to put them at ease by starting with informal conversations before asking non-sensitive screening questions such as whether they were Singaporean or permanent resident and between the ages of 21 and 70 years. The men were then asked to read the next question (also translated into Chinese and Malay) and tick the types of sex partners they had had sex with in the past 12 months. If they ticked ‘wife’ and/or ‘girlfriends’ only, the interviewer stopped the interview, thanked them and gave them a small gift (tissue paper packs with health messages and a canned 100 plus drink). If they ticked ‘casual sex partners’ (defined as above) or ‘prostitutes’, the interviewer asked them to complete a short questionnaire after showing them the participant information sheet and obtaining oral consent. Interviewees were given Singapore $15 on completing the questionnaire. For the online survey, the participant information sheet and the questionnaire were sent through a private messaging function in the survey Monkey web-based tool to each member of the online sex forums. Participants who completed the survey were asked to leave an email address for claiming the$15 reimbursement for online shopping with an independent online vendor.

The short questionnaire included sociodemographic information and questions on the number of sexual encounters, types of sex worker (freelance or brothel based) visited and frequency (never, sometimes, always) of condom use with them in the past year. They were also asked on their chance of getting HIV/AIDS/STIs from not using condoms; whether they had had any STD symptoms such as genital discharge, ulcer or sore, genital growth and painful urination and whether they had been screened for STIs/HIV in the past 12 months. The study was approved by the institutional review board of the National University of Singapore.

### Statistical analysis

We compared categorical variables with χ2 test, continuous variables with independent samples t test and ordinal variables with Wilcoxon rank-sum test. Separate multiple logistic regression analyses (Statistical Package for Social Science (SPSS), V.20.0 (SPSS, Chicago, Illinois, USA), using a backwards stepwise elimination procedure, were performed to identify independent factors significantly associated with each outcome: consistent condom use in the past year according to type of partner and intercourse in the brothel and online groups, respectively.

## Results

Table 1 compares the sociodemographic characteristics of men from the two groups. Men who went online to seek sex were younger and more likely to be married. They were also higher educated with almost half (45.4%) having completed university education compared with 6.6% in the brothel group. Online users were also more likely than the brothel group to patronise massage parlours (50.8% vs 7.6%), social escort services (48.1% vs 5.5%) and entertainment establishments (32.7% vs 5.5%) (not in the table). On comparing their sexual behaviours in the past year (table 2), internet sex seekers were more likely than men patronising brothels to report casual sex (56.3% vs 26.6%). The online group were more likely than the brothel group to engage in both casual (77.6% vs 19.0%, p<0.001) and commercial sex (56.2% vs 33.7%, p<0.001) while travelling overseas (not in the table). Both groups reported a similar median number of paid and casual partners. The online group also reported higher rates of oral and anal intercourse with commercial, casual and regular partners. While both groups reported similar rates of consistent condom use for vaginal intercourse with sex workers (76%, 72%) and casual partners (51%, 54%) in the past year, the online group reported lower rates of condom use for oral sex with both commercial (16.5% vs 56.6%) and casual partners (21.2% vs 35.3%). Condom use varied by type of partner for both groups. Consistent condom use for vaginal intercourse in both groups was highest with commercial partners, decreased with casual partners and was lowest with wives/girlfriends. On comparing condom use by the type of sexual intercourse, both groups reported the lowest rates of condom use for oral sex for all partner types. The online group, however, reported higher rates of condom use for anal than vaginal intercourse with casual partners and wives/girlfriends. STD symptoms in the past year were more often reported by the online than the brothel group.

Table 1

Comparison of sociodemographic characteristics of men seeking sex online and from brothels

Table 2

Comparison of sexual behaviours of men seeking sex online and from brothels

For men seeking sex online, consistent condom use for vaginal sex was significantly higher among the younger than older age groups during commercial sex. It also increased with perceived risk of getting HIV for both casual and commercial sex but it decreased with higher educational level for casual sex. These associations were not found in the brothel group. Both groups reported some similarities—an increase in condom use with both casual and commercial partners’ request for condom use; and a decrease in condom use with alcohol consumption before having sex with casual partners in the online group and with commercial partners in the brothel group (table 3).

Table 3

Consistent condom use for vaginal sex by partner type and sociodemographic characteristics

As oral sex was widely practised, we examined factors associated with consistent condom use for oral sex with commercial and casual partners (table 4). The significant associations in the online group were found in sexual relationships with casual partners only. Condom use for oral sex decreased with alcohol consumption before having sex and increased with perceived high risk of getting HIV from casual partners. However, it was not associated with partner request for condom use. In contrast, condom use for oral sex with both partner types in the brothel group increased significantly with partner request for condom use.

Table 4

Consistent condom use for oral sex by partner type and sociodemographic characteristics

For multiple logistic regression analysis, all independent variables with p<0.1 in the bivariate analyses were entered into each of the respective logistic regression models. We did not find notable interaction or collinearity between sociodemographic factors and partner factors. Consistent condom use for vaginal sex with commercial partners in the online group showed a strong association (adjusted OR (AOR)=3.54, 95% CI 1.99 to 6.29, p<0.001) with request by commercial partners to use condoms; and a borderline association (AOR=1.80, 95% CI 0.98 to 3.33, p=0.06) with perceived high risk of getting HIV from commercial partners. Condom use with casual partners, was associated with perceived risk of getting HIV from casual partners (AOR=4.50, 95% CI 1.58 to 12.87, p=0.004), alcohol consumption (AOR=2.32, 95% CI 1.09 to 4.94, p=0.03) and request for condom use by casual partners (AOR=9.30, 95% CI 3.26 to 26.56, p<0.001). For the brothel group, only partner request for condom use was associated with condom use for vaginal sex with both commercial (AOR=3.18, 95% CI 1.81 to 5.59, p=0.004) and casual partners (AOR=5.33, 95% CI 1.37 to 20.73, p=0.016), respectively.

On examining oral sex in the online group, condom use with casual partners was negatively associated with alcohol consumption (adjusted AOR=0.27, 95% CI 0.10 to 0.72) and positively associated with perceived high risk of getting HIV from casual partners (adjusted AOR=3.13; 95% CI 1.20 to 8.13). For the brothel group, condom use for oral sex with casual and commercial partners, respectively, was associated with partner request to use condoms but not with perceived risk of getting HIV from their partners. Alcohol consumption showed a borderline association with condom use with commercial partners only (AOR=0.57, 95% CI 0.32 to 1.01, p=0.06).

## Discussion

The online group, in comparison with the brothel group, reported higher rates of oral and anal intercourse with both commercial and casual partners, but significantly lower rates of consistent condom use for oral sex in the past 12 months. These findings are consistent with the higher rates of self-reported sexually transmitted disease (STD) symptoms in the online group. Our findings are similar to those found by a study in the USA, which found that online seekers report higher rates of oral and anal intercourse and higher rates of self-reported STDs than offline clients.1

The higher prevalence of self-reported STD symptoms in the online group is unlikely to be due to its higher screening rates as only 34% had been screened for STIs compared with 44% among men frequenting brothels. We found notable differences in factors associated with condom use between the two groups that might be explained by the differing influence of the ‘institutional’ environment. Condom use for vaginal sex with both commercial and casual partners; and condom use for oral sex with casual partners in the online group were significantly associated with their perceived risk of getting HIV/STIs from non-condom usage. Self-perception of risk is a logical ‘reasoning’ process which operates at the individual level. It was found in our study to influence the behaviour of using condoms in the online group. In contrast, condom use with both commercial and casual partners among men patronising brothels was not associated with their perceived risk of getting HIV/STIs from their partners. A possible explanation is that the brothel's 100% condom use policy has superseded and played a more important role than individual level reasoning in influencing condom usage with commercial partners for this group. The internalisation of the behaviour of consistent condom use acquired from the ‘brothel's 100% condom use policy’ environment might have been transferred to casual sex relationships. This might explain the lack of association of condom use with self-perceived risk in casual sex relationships in the brothel group.

The strong association of condom use for vaginal sex with partner request for condom use in both groups is encouraging. However, this association was not seen for oral sex in the online group. Possible explanations, which need to be confirmed with further research, include the low awareness of the oral route of STI transmission and the low self-perceived risk of acquiring STIs from oral sex. Casual sex with opposite-sex partners in the past year has increased markedly from 1.1% in 1989 to 17.4% in 2007 among heterosexuals in Singapore, with the majority (84%) practising unprotected sex.17 We need to increase awareness of the risk of acquiring infections from oral sex and empower mainstream women to initiate condom use with their male heterosexual partners.

This study's limitations include the low response from online users. The two groups may not be mutually exclusive because 10% of the online group also patronised brothels. None in the brothel group, however, ventured online to seek commercial partners, although 10% did so to seek casual partners. These small percentages are unlikely to introduce substantial bias in the comparison of risky sexual behaviours between the two groups. Moreover, restricting our analysis to men who bought sex online only shows similar findings. The questionnaire was kept short to encourage good response because the majority wanted to rush away after leaving the brothels. Finally, clients might over-report condom use, but the low rates of self-reported condom use show that this was unlikely. Moreover, we tried to facilitate recall by approaching men who had just stepped out from the brothels.

Despite the limitations, our study has several strengths. The samples of men visiting the two venues were obtained from a sampling frame of brothels and the most popular internet sex forums in Singapore. Most previous studies have recruited online sex seekers from attendees of STI clinics.1 ,18 The high response rate and time location sampling also support the generalisability of the findings of the brothel group to all licensed brothels in Singapore. Finally, this study, to our knowledge, is the first to compare risky sexual behaviours of heterosexual men using the internet with the behavior of those frequenting brothels.

In conclusion, heterosexual men seeking sex online were more likely than men frequenting brothels to report unprotected oral intercourse with both commercial and casual partners and STD symptoms in the past year. They were also more likely than the brothel group to report having unprotected sex with their wives and girlfriends. Clearly, STI/HIV prevention programmes should be prioritised for this online group at greater risk. Their differences from the brothel group in sociodemographic characteristics and factors associated with condom use suggest a need to design separate safer sex promotion programmes for them. One major challenge that might impede the effectiveness of interventions targeting internet sex forums is the inability to employ institutional policies to mandate condom use. More attention should be given to online health education programmes to promote consistent condom use with both commercial and casual partners, and STI testing. Innovative messages on prevention of HIV/STIs, such as real accounts of how men have acquired HIV/STIs, should be designed to make the messages more convincing. Given that condom use among the online men increased with self-perceived risk, messages on the high risk of HIV/STIs via oral sex and with casual partners should be disseminated. Other online resources such as websites and help lines about prevention of HIV/STIs should also be provided.

### Key messages

• Men seeking sex online reported riskier sexual behaviours and higher rates of sexually transmitted diseases than men patronising brothels.

• For the online group, commercial and casual partner request for condom use was associated with condom use for vaginal sex but not for oral sex.

• Condom use in the online group was significantly associated with their perceived risk of getting HIV/STIs from their partners.

• The non-association of perceived partner's risk with condom use in the brothel group suggests that institutional policy has superseded individual reasoning in influencing condom use.

## Acknowledgments

We thank Dr Howie Lim Sin How, Mr George Qiongze and Mr Terry Lim for their invaluable contributions to this study. We also thank all the interviewers for their painstaking efforts in conducting the interviews and coordinating the survey.

## Footnotes

• Handling editor Jackie A Cassell

• Contributors MLW developed the original protocol, study design, conducted the statistical analysis and drafted the manuscript. MG initiated the study and contributed to the study protocol and write up. TTK and ST contributed to data collection, statistical analysis and write-up.

• Funding This study was funded and supported by the Health Promotion Board (R-186-000-135-490) and Saw Swee Hock School of Public Health, National University of Singapore.

• Competing interests None.

• Ethics approval Institutional review board of the National University of Singapore.

• Provenance and peer review Not commissioned; externally peer reviewed.

## Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.