The prevalence of, and factors associated with, paying for sex among men resident in Britain: findings from the third National Survey of Sexual Attitudes and Lifestyles (Natsal-3)

Objectives Men who pay for sex (MPS) are considered a bridging population for sexually transmitted infections (STI). However, the extent, characteristics and role of MPS in transmission is poorly understood. We investigate these questions using data from Britain's third National Survey of Sexual Attitudes and Lifestyles (Natsal-3). Methods We performed complex survey analyses of data from 6293 men aged 16–74 years resident in Britain who completed Natsal-3, a probability sample survey undertaken during 2010–2012, using computer-assisted personal interviewing and computer-assisted self-interview. Results 11.0% (95% CI10.1% to 11.9%) of all men reported ever paying for sex. Among MPS, 18.4% (95% CI 18.2% to 18.7%) of their lifetime sexual partners were paid. 3.6% (95% CI 3.1% to 4.2%) of men had paid for sex in the past 5 years. Partners of MPS constitute 14.7% of all reported partners and MPS report 15.6% of all reported STI diagnoses in the past 5 years. Paying for sex in the past 5 years was strongly associated with reporting larger numbers of sexual partners (adjusted OR, AOR for 5+ partners, past 5 years, 31.50, 95% CI 18.69 to 53.09). After adjusting for partner numbers, paying for sex remained strongly associated with reporting new foreign partners outside the UK (AOR 7.96; 95% CI 4.97 to 12.73) and STI diagnosis/es (AOR 2.34; 95% CI 1.44 to 3.81), all in the past 5 years. Among men ever paying for sex, 62.6% (95% CI 58.3% to 66.8%) reported paying for sex outside the UK, most often in Europe and Asia. Conclusions MPS in Britain remain at greater risk of STI acquisition and onward transmission than men who do not. They report high numbers of partners, but the minority are paid partners. They are an important core group in STI transmission.


INTRODUCTION
Men who pay for sex (MPS) are considered to be a bridging population for sexually transmitted infections (STI), as their paid partners are often individuals at high STI risk, whose risk is conferred to the unpaid partners of MPS, with whom condom use is less likely. [1][2][3][4] MPS may thus be at higher risk of STIs than men who do not pay for sex. 1 5-7 A study of Scottish genitourinary medicine (GUM) clinic attendees found that over half of paid sex occurred abroad, primarily in Europe and Asia, 8 potentially driven by known 'hotspot areas' for sex tourism, such as Amsterdam or Bangkok. 9 As MPS are a hard-to-reach population for research and intervention, most studies have used convenience sampling to access them, 10 such that their data may not be representative of the general population. Britain's third National Survey of Sexual Attitudes and Lifestyles (Natsal-3) by contrast, uses probability sampling of the general population. Initial results from Natsal-3 show that 4% of men aged 16-44 years reported paying for sex in the 5 years prior to interview in 2010-2012, and that prevalence is similar to a decade ago. 11 However, the role of this minority of the population in the transmission dynamics of STI/HIV depends on the prevalence of paid partners and on the number of paid and unpaid partners. Furthermore, as STI/HIV prevalence varies globally, 12 it is also important to know in which global regions men pay for sex.
We examine variations in the prevalence of paying for sex and the recency of last paid sex by age among men in Britain, and estimate the proportion of paid sex partners among British men's sexual partners. We also explore how paying for sex varies by key sociodemographic, sexual behaviour, and health-seeking variables, including selfreported STI diagnosis/es, and how it is statistically associated with these variables. Last, we identify the global geographical regions where men pay for sex. Given that the prevalence of women paying for sex among  year olds in the past 5 years is 0.1%, 11 the analyses presented in this paper focus exclusively on men.

METHODS Study design and participants
Natsal-3 is a stratified probability sample survey of 15 162 men and women (6293 men) aged 16-74 years, resident in Britain, undertaken between September 2010 and August 2012. Details of the sampling methodology and data collection are published elsewhere. 11 13 Briefly, participants were interviewed using a combination of face-to-face, computer-assisted, personal interviewing (CAPI) and computer-assisted self-interviewing (CASI).
The more sensitive questions, including those on paying for sex and sex while outside the UK, were asked in the CASI. 13 The response rate was 57.7% and the cooperation rate (ie, of all eligible addresses contacted) was 65.8%.

Interviewing and data collection
Men were asked if they had ever paid for sex with a man and/or a woman. Those who reported they had paid for sex were then asked when they last paid for sex, how many sexual partners they had ever paid, and whether they had ever paid for sex while outside the UK. Men who had paid for sex while outside the UK were asked a further question about the geographical region in which they had paid for sex. 13 Men were also asked to indicate whether or not they had included their paid partners in their total reported partners at the time of interview (referred to hereon as 'lifetime partners' for brevity). Demographic variables include age and relationship status at interview, individual socioeconomic status according to the National Statistics Socio-Economic Classification (NS-SEC), 14 and area-level deprivation, using quintiles of the Index of Multiple Deprivation (IMD), a multidimensional measure combining income, employment, health, education, access to housing and services, crime and living environment. 15 Measures of health, sexual health, and health-seeking behaviour include frequency of binge drinking (consuming more than eight units of alcohol on one occasion), 16 recent drug use, low sexual function, 17 18 STI diagnoses, and sexual health clinic attendance. Finally, measures of sexual risk behaviour include number of sexual partners, reporting new foreign sexual partner(s) while outside the UK, 19 same-sex experience, condom use, reporting concurrent partners, and finding sexual partners on the internet.

Statistical analysis
Prevalence of paying for sex and recency of the occasions of the experience (ever, past 5 years, and past year) was examined by age group. Prevalence of same-sex paid sex, 0.2% (95% CI 0.1% to 0.4%) of the male population, was too low to conduct any separate analyses. The number of paid partners ever reported was added to the number of lifetime partners reported if the participant had indicated that these partners had not been included in their partner counts. Mean lifetime partner numbers, mean paid partner numbers, and the proportion of paid partners were calculated and examined across sociodemographic, general health, sexual behaviour, and sexual health-related variables for the population and for all MPS. The proportion of paid partners was calculated by dividing total reported paid partners by total lifetime partners for the population and within the categories of each variable. CIs were bootstrapped, as the resulting estimates were calculated from summary statistics and needed to have SEs estimated. Unadjusted logistic regression models were used to explore associations between reporting paying for sex in the 5 years prior to interview and sociodemographic, general health, and sexual behaviour variables. Multivariable logistic regressions adjusting for the potentially confounding effect of key sociodemographic variables were used to assess the associations of general health and sexual behaviour variables with the outcome of paying for sex in the past 5 years. Sexual behaviour variables were additionally adjusted for the confounding effect of the reported number of sexual partners in the past 5 years. Additional univariable and multivariable logistic regressions adjusting for sociodemographic variables and number of sexual partners in the past 5 years assessed the associations between paying for sex and key sexual health outcomes. Finally, we drew comparisons between the geographical regions where men had ever paid for sex when outside the UK, and the geographical origin of unpaid new sex partners while outside the UK, in the past 5 years. Those reporting unpaid sex partners outside the UK were restricted to those who had never paid for sex abroad to give a distribution of their partners' origins without including any potential paid partners. No formal statistical comparison between the regional distributions for paying for sex abroad and having new sex partners abroad are made because of inconsistent time periods and subtle changes in wording in the questions of comparative interest.
All analysis used Stata V.12.1, 20 and accounted for the sample weighting, clustering and stratification within the Natsal-3 sample. The data were first weighted to correct for unequal selection probabilities and then to match the demographic profile of the British population figures, in terms of gender, age and Government Office Region, according to the 2011 UK census. 13

Variation in prevalence by age
A total of 6108 men answered the question on paying for sex (97.1% of all male participants in Natsal-3), and of these, 11.0% (95% CI 10.1% to 11.9%) reported ever paying for sex; 3.6% (95% CI 3.1% to 4.2%) of all men reported paying for sex in the past 5 years, and 1.1% (95% CI 0.9% to 1.5%) reported doing so in the past year (table 1). Reporting ever paying for sex was lowest in men aged 16-24 years and three times as high in 25-34-year-olds. Lifetime prevalence was highest in men aged 55-64 years, while men aged 25-34 years had the highest prevalence in the past 5 years. Paid partners accounted for 4.6% (95% CI 4.5% to 4.7%) of all reported partners (table 1). This percentage ranged from 1.4% (95% CI 1.3% to 1.4%) of all partners reported by men aged 16-24 years to 7.5% (95% CI 7.3% to 7.8%) of all partners reported by men aged 55-64 years.
Proportion of all partnerships among MPS that are paid On average, MPS reported more than twice as many lifetime partners as the male population average (means of 31.6 and 13.6, respectively) which holds true across all age groups (table 2). Among men who had ever paid for sex, one in five of their partners were paid, increasing with age from 12.8% of partners in men aged 16-24 years to 26.1% of partners in men aged 65-74 years. This proportion was higher among men who reported being in bad or very bad health, driven by a low denominator of mean lifetime partner numbers, and not the numerator of mean paid partner numbers. The proportion of paid partners was lower among men who had used drugs other than cannabis in the past year, driven by high mean lifetime partner numbers. While the prevalence of men paying for sex in the past 5 years was higher in men who reported partners outside the UK, same sex contact, finding partners on the internet, and concurrent partners, these groups did not report proportionally more paid partners than other men who paid for sex.

Factors associated with paying for sex
After adjustment for age, relationship status, individual socioeconomic status, residence and area-level deprivation, paying for sex in the past 5 years was most commonly reported by men aged 25-34 years (table 3). Men aged 16-24 years and 64-74 years were least likely to report paying for sex in this time period. Compared with men who lived with a partner, men with no current steady partner were more likely to have paid for sex in the past 5 years (adjusted OR (AOR) 2.58 (95% CI 1.84 to 3.62)). Men in a managerial or professional occupation were more likely to have paid for sex than men in other occupations (AOR 1.88 (95% CI 1.27 to 2.79)). Men who reported binge drinking once a week or more, or using drugs other than cannabis in the past year, were more likely to report paying for sex in the past 5 years than those who did not report binge drinking or drug use (AOR 1.84 (95% CI 1.30 to 2.61) and AOR 5.01 (95% CI 3.38 to 7.42), respectively) (table 3). The odds of reporting paying for sex increased with the number of sexual partners reported in the past 5 years (AOR for 5+ partners; 31.50 (95% CI 18.69 to 53.09) relative to reporting 0-2 partners). After additional adjustment for partner numbers, men who reported new foreign sexual partners while outside the UK, sexual partners found online, and concurrent partners, were also more likely to have paid for sex (AORs of 7.96 (95% CI 4.97 to 12.73), 1.97 (95% CI 1.28 to 3.03), and 2.47 (95% CI 1.60 to 3.80), respectively), while there was no association with using condoms specifically for HIV/STI prevention (AOR 1.40 (95% CI 0.93 to 2.10)).

Sexual health outcomes associated with paying for sex
In the past 5 years, partners of MPS accounted for 14.7% of all reported sexual partners, and MPS accounted for 15.6% of all men reporting STI diagnoses (data not shown). After adjusting for the number of sexual partners, MPS were still more likely to report the sexual health outcomes listed in table 4 than those who did not pay for sex; they were more likely to report STI diagnosis/es in the past 5 years (AOR 2.34 (95% CI 1.44 to 3.81)) and to have lower sexual function in the past year (AOR 1.80 (95% CI 1.25 to 2.69)). 17 These men were also more likely to report having attended a sexual health clinic (AOR 1.69 (95% CI 1.10 to 2.58)) and testing for HIV (AOR 2.20 (95% CI 1.53 to 3.17)) in the past 5 years.

Where men pay for sex
Among men who reported ever paying for sex, 62.6% (95% CI 58.3% to 66.8%) had, at some time, paid for sex outside the UK (data not shown). The most commonly reported geographic regions where men paid for sex were Europe (64.5% (95% CI 58.9% to 69.6%)) and Asia (25.4% (95% CI 20.9% to 30.5%)), with other regions accounting for much smaller proportions (see web appendix). Among men who had not paid for sex, Europe (52.7% (95% CI 45.9% to 59.4%)) and Asia (13.9% (95% CI 9.8% to 19.2%)) were also the most commonly reported geographic regions in terms of the origins of their new foreign (unpaid) partners. By contrast with men who had paid for sex, large proportions of men who had not paid for sex also reported having sex abroad with people who usually lived in North America (22.1% (95% CI 16.9% to 28.5%)) and Australasia (12.4% (95% CI 8.5% to 17.7%)).

DISCUSSION
Around one in 10 men in Britain report having ever paid for sex. These men account for a disproportionate number of sexual partners reported by all men, of which the minority (18.4%) are paid. They are also more likely to report STI diagnoses even when accounting for their disproportionately larger number of sexual partners, which is frequently considered the most important behavioural variable associated with STI diagnoses. 21 This evidence strongly supports the idea that these men are a bridge for disassortative sexual mixing and for the spread of STIs. MPS are most likely to be aged between 25 and 34 years, single, in managerial or professional occupations, and have high partner numbers. After adjusting for the key risk behaviour of sexual partner numbers, these men still report many other sexual behaviours, such as having new foreign partners while outside the UK, and with less pronounced differences in sexual health-related behaviour, such as STI clinic attendance or condom use, exposing an increased vulnerability without an  Behaviour increase in precautionary behaviour. However, this increase in precautionary behaviour, whether it be a consequence of other sexual behaviour or not, may present an opportunity for targeted interventions for these men. Despite lifetime prevalence of paid sex varying little between men aged 25-64 years, there is a steady increase with age in the mean number of paid partners, suggesting either generational changes in paid sex, or that a proportion of MPS continue to pay for sex as they age. The varying proportions of partners paid among MPS is largely driven by lifetime partner numbers rather than numbers of paid partners, such that a lower proportion of paid partners actually reflects a high number of lifetime partners. However, for some reported behaviours, such as sex partners outside the UK, same sex contact, sex partners found online, and concurrent partners, total and paid partner numbers increase. This suggests that MPS exhibiting these behaviours have higher lifetime partner numbers than other MPS as well as higher paid partner numbers, putting them at a higher risk for STIs than other MPS.
The theory of Situational Sexual Behaviour may provide some explanation for the strong relationships observed between reporting paying for sex and new sex partners while outside the UK. The theory posits that sexual behaviour which is different to an individual's normal sexual behaviour, like paying for sex, is sometimes determined by the set of circumstances in which individuals find themselves. [22][23][24] For example, as a place removed from the day-to-day activities of most men, overseas travel facilitates the opportunity for paid sex. These opportunities may also be facilitated by an ease of access in terms of proximity or cost of travel to an area with an established sex-work industry, and/or a legislative framework which serves to increase the safety of the encounter, as in The Netherlands. 25 As paying for sex in the past 5 years was far more commonly reported by men who had new sexual partners while outside the UK (29.9%), it is then not surprising that the majority of men who had paid for sex reported doing so outside the UK, with Europe and Asia the most commonly reported geographical regions. It is interesting to consider these geographical patterns in the context of Britons' international travel patterns with nearly 80% of British residents visiting Europe when travelling abroad in 2011, and around half of all paid sex outside the UK occurring in Europe. By contrast, a fifth of all paid sex outside the UK occurred in Asia, but this is the destination for only 1 in 20 Britons travelling abroad. 26 This suggests that travel to Asia results in disproportionately more sex tourism than travel to Europe. Although paid sex in South America and Australasia constitutes around 5% of reported paid sex, only 1% of travel is to these destinations, suggesting a similar imbalance between sex tourism and travel. These findings concur with those of earlier studies, 11 12 and have implications for targeting sexual health interventions and health promotion messages to those travelling to these sex tourism areas, especially given the higher STI/HIV prevalence in these regions. 10 As with all cross-sectional studies, causality cannot be determined. This is especially problematic with data relating to sexual risk and health-seeking behaviour, as we can only assume that increased sexual risk resulted in increased health-seeking behaviour. Questions referring to an entire sexual history, such  13 may produce less reliable estimates for the population of MPS. §CI is estimated using 100 bootstrapped replications of each estimate. **Index of Multiple Deprivation (IMD) is a multidimensional measure of area (neighbourhood)-level deprivation based on the participant's postcode: IMD scores for England, Scotland and Wales were adjusted before being combined and assigned to quintiles, using a method by Payne and Abel. 15 † †More than 8 units at one time. 16 ‡ ‡Excludes partners from the UK while outside of the UK. § §Sexual function score calculated using the Natsal-SF 17 18 only for the population who were sexually active the past year. ¶ ¶STIs include genital warts, trichomonas, gonorrhoea, chlamydia, syphilis, non-specific or non-gonococcal urethritis, and pelvic inflammatory disease. MPS, men who pay for sex; NS-SEC, National Statistics Socio-Economic Classification; STI, sexually transmitted infection. Behaviour as lifetime partners, rather than more recent timeframes, may be subject to recall biases. There are also likely to be different types of MPS, and it is possible that our data may be skewed by particular types of MPS. As a population-based survey of a wide range of sexual behaviours, Natsal-3 asked a limited number of questions about paying for sex. As a result, it was not possible to determine whether men paid for sex in the UK only, outside the UK only, or in and outside the UK. This also means that we are not able to measure the extent of overlap between reporting paying for sex and reporting new sex partners outside the UK.
Additionally, formal statistical comparisons across the geographical regions for paid and unpaid partners are not possible because of differences in timeframes and wording of the questions of comparative interest, that is, origin of unpaid partners abroad and region where paid partners were found. Last, the geographical regions listed in the questionnaire where paid sex occurred are very large, and future surveys should consider asking men for specific locations which could subsequently be categorised. Despite these limitations, data from Natsal-3 enable us to present the first ever population-based estimates of where  British men pay for sex outside the UK, and update our understanding of the factors associated with paying for sex using data that are broadly representative of the British population. Our data show that MPS continue to be at increased risk of STI acquisition and onward transmission. Further research is needed to establish why men may choose to pay for sex in their own country, while abroad, or both, and whether these groups of men and their behaviours are heterogeneous and, if so, how.

Key messages
▸ Men who pay for sex are at higher risk of acquiring STIs even after considering other key STI risk behaviours. ▸ Having paid partners may not directly increase men's STI risk, but may rather be a marker for men who are likely to partake in other sexual risk behaviours. ▸ Men who pay for sex should be considered a core-group for sexual health interventions and services.
Handling editor Jackie A Cassell  1 Excludes 121 (140 weighted) men who report paying for sex abroad to avoid counting any paid partners and biasing the data towards the regional distribution of paid partners. 2 Columns do not add to 100% as men can report more than one region

British men who pay for sex have many (unpaid) sexual partners
Often young professionals, with a taste for binge drinking and recreational drugs And those who paid for sex within the past five years are often young, single professionals, with a taste for binge drinking and recreational drugs, the findings show.
The researchers base their findings on a detailed analysis of data from the third National Survey of Sexual Attitudes and Lifestyles (Natsal-3), carried out between 2010 and 2012.
Natsal is run every 10 years, and includes a representative sample of men and women resident in Britain and aged between 16 and 74.
The current study focused on the responses of 6108 men to questions about paid-for sex, to find out more about the behaviour and profile of those who pay for sex and their risk of, and potential role in, the transmission of sexually transmitted infections.
More than one in 10 (11%) of the men said they had paid for sex at some point in their adult lives; 3.6% said they had done so within the previous five years; and just over 1% said they had done so within the preceding 12 months.
Although 55-64 year olds were most likely to say they had ever paid for sex, those aged between 25 and 34 were most likely to have done so within the past five years, and they were three times more likely to have done so than 16 to 24 year olds.
Paying for sex was associated with reporting twice as many lifetime sexual partners as the general population average across all age groups, but relatively few of these were paid for.
Paying for sex within the past five years was associated with being single, having a managerial/professional job, binge drinking and using recreational drugs other than cannabis.
And men who reported at least five sexual partners in the past five years were much more likely to have paid for sex (16% of them) than men who reported a maximum of two sexual partners during this time (1%).
Similarly, men who said they had had paid sex outside the UK, found sexual partners online, or who had concurrent or overlapping partners, were more likely to have paid for sex.
Almost two thirds of those (62.6%) who had paid for sex had done so outside the UK, with Europe and Asia the most popular destinations. By contrast, large numbers of men who had not paid for sex also had sex abroad-but predominantly in North America and Australasia.
Compared with other men, those who had paid for sex within the past five years were around twice as likely to say they had been diagnosed with a sexually transmitted infection, visited a sexual health clinic, or been tested for HIV.
"The evidence strongly supports the idea that these men are a bridge for disassortative [dissimilar] sexual mixing and for the spread of [sexually transmitted infections]," write the researchers.
In an accompanying podcast, lead author, Mr Kyle Jones, makes the point that while paying for sex may not itself be a risk factor for sexually transmitted infections, it may none the less be a marker for sexually risky behaviour.
In a linked editorial, Professor Sonia Dias, of the Institute of Hygiene and Tropical Medicine at the University of Lisbon in Portugal, emphasises: "The evidence that men paying for sex report high numbers of partners, including foreign partners outside the UK, and are more likely to report STI diagnoses, supports that they are a key population to be targeted."