WHAT IS ALREADY KNOWN ON THIS TOPIC

Approximately half of Japanese men aged 20–49 years have paid for sex, including call services, erotic massages and oral sex.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICYIntroduction

Commercial sex work (CSW) has a long history in Japan: government-sanctioned red-light districts were established in Yoshiwara, Edo (present-day Tokyo), Shimabara, Kyoto and Shinmachi, Osaka during the 17th century.1 Modern-day neighbourhoods resembling red-light districts continue to thrive in most major Japanese cities,2 most famously Kabukicho, Tokyo. CSW remains strikingly visible in Japan, with kiosks offering sexual services (muryōannaijo) and love hotels (hotels that can be rented for short time periods that cater to couples looking for a space to have sex) conspicuously located between restaurants and convenience stores. The use of CSW services, especially by men, is associated with low stigma in Japan compared with Western nations, and in some cases may even be a pastime openly discussed with friends (contrastingly, employment as a sex worker remains highly stigmatised). Many types of non-penetrative sexual services are legal in Japan; though vaginal intercourse is not technically legal, a blind eye is often turned to the world of CSW.3 4 In total, sex work comprises a multitrillion yen industry in Japan and up to 1% of the country’s gross national product.2 5

Over the past few years, bacterial sexually transmitted infection (STI) notifications in Japan have been increasing6; in particular, the number of syphilis cases has been rising since 2019 and is now at its highest in the past 20 years.7 The reason for the sudden increase in STI notifications remains unknown, but many have posited that increased use of sexual services and/or decreased condom use with commercial sex workers may be to blame.8 However, there are little data concerning rates of CSW use in Japan, and the research that exists tends to focus on high rates of STIs among the commercial sex workers themselves.9 10 Those who use CSW are also an important target of public health research, as they constitute bridges between sex workers and the community,11 12 and a considerable proportion of CSW is performed without the use of condoms (condom use rates of 26.3–80.5% depending on the type of service).13

A prior descriptive analysis of the inaugural National Inventory of Japanese Sexual Behavior (NInJaS), conducted in 2022, found that approximately 48.3% of men aged 20–49 years reported having used CSW services in their lifetime.14 This high figure may be explained by the easy accessibility of CSW in Japan, the overall low stigma associated with the use of sex work and the diversity in the types of services offered (summarised in table 1).11 14 15

Table 1

Description of major types of commercial sex services used in Japan

In this study, we use NInJaS data to determine the factors associated with lifetime history of CSW use among men aged 20–49 years in Japan. Our findings constitute the first step to illuminate the characteristics of CSW users who comprise critical targets in public health efforts to promote safer sexual practices, improved sexual wellness and increased STI testing.

MethodsData sources

Data for the current study were taken from the NInJaS survey, for which the methods are explained in detail elsewhere.14 Briefly, the survey was administered via Cross Marketing, an online survey company, in July 2022. The survey was made available to Cross Marketing panel members aged 20–49 years who were able to respond to surveys in Japanese. Participation in the survey was voluntary, and in return for participating, respondents receive ‘points’ that can be redeemed for goods/services from partner companies.

The target sample size was set at 8000, which was to be split evenly by gender and 10-year age groups (20–29, 30–39 and 40–49 years). Respondents participated on a first-come first-served basis until the quota for a certain subgroup was reached, at which point the survey period would end for that particular gender/age stratification.

Because of the sensitive nature of the questionnaire, participants were required to pledge that they would answer the remaining survey questions honestly at the start of the survey; this method has been shown to improve response quality.16 Similarly, respondents were required to provide consent for participating in the survey; they were welcome to withdraw from survey participation at any time. However, those who did participate in the survey were required to answer all questions such that no missing values were generated.

Measures

An English translation of the survey, as well as a detailed description of survey design and contents, is provided elsewhere.14 Briefly, the survey assessed sociodemographic variables, anthropometric variables, sexual attitudes and a variety of sexual behaviours over the lifetime and past year, including number of sexual partners, frequency of sexual intercourse and use of commercial sex workers, among others.

Study population

The target population of the current study was men (as reported on birth certificates), aged between 20 and 49 years; in total, 4000 men, evenly split between ages 20–29, 30–39 and 40–49, were included in the analysis. Sample weights based on the 2020 Population Census of Japan were calculated and applied to the respondents in order to make the sample representative of the national population with respect to 5-year age groups (20–24, 25–29, 30–34, 35–39, 40–44, 45–49), marital status (married and unmarried) and region (Hokkaido/Tohoku, Kanto, Chubu, Kansai, Chugoku/Shikoku and Kyushu/Okinawa). After application of weights, the sample largely reflected the distribution of the national population.14

Outcome measures

The primary outcome of the current analysis was lifetime use of CSW services, based on the following question: ‘In your lifetime, have you ever used any of the following sexual services? Please select all that apply. For the purposes of this question, hostess/cabaret clubs do not apply.’ Options included: ‘soaplands’, ‘erotic massage’, ‘call services’, ‘image clubs’, ‘pink salons’, ‘urisen’, ‘sexy cabaret’, ‘other’, ‘I have used sexual services before, but I do not know what kind’, and ‘I have never used any sexual services’ (the final answer choice was not mutually compatible with any of the other options). Lifetime use of CSW services was binarised to either 0 (never used) or 1 (at least one use).

Descriptive analysis

First, we calculated the median number of paid sexual partners over the lifetime and in the last year among those who have paid for sex during their lifetime and in the past year, respectively (values for all respondents are reported elsewhere).14 ‘Sexual intercourse’ was explicitly defined in the text of the survey questions as vaginal, anal and/or oral intercourse, consistent with previous surveys.17 We described the sociodemographic, anthropometric and attitudinal characteristics of respondents, stratified by 10-year age groups and lifetime use of CSW services. Sociodemographic variables include age, marital status, relationship status, sexual orientation, occupational status, income, education, region of residence and urbanity of residence. Anthropometric variables include height, weight, body mass index (BMI) (calculated based on height and weight) and self-rated attractiveness (further discussed in the online supplemental file 1). Attitudinal variables include sexual satisfaction, desire to change frequency of sexual intercourse and the importance of sex. We hypothesised that these variables could be associated with sexual behaviours and thus they were selected, a priori, at the time of survey creation. All descriptive analyses were performed with the application of sample weights in Stata V.17.0 (StataCorp, College Town, Texas, USA).

Regression analysis

We analysed the association between sociodemographic, anthropometric and attitudinal variables and lifetime use of sexual services using logistic regression, with lifetime use of CSW services as the dependent variable (binarised as described above). Regressions were performed for all respondents, as well as stratified by 10-year age groups, as use rates of sexual services differ considerably by age.14 As has been done in previous studies of Japanese sexual behaviour,18 we conducted univariate regressions for each variable and adjusted for age to calculate age-adjusted ORs (aORs) as our primary analysis, consistent with prior studies on the social determinants of sexual health both within and outside of Japan.17–20 Furthermore, we also conducted multivariate regressions using all independent variables in a single model to determine which variables retained significance when controlling for other associations. An additional analysis was performed with ‘CSW use in the past year’ as the dependent variable. Sample weights were applied to each regression. Results were considered statistically significant if the 95% CI of the aOR did not cross 1.

Data availability

The dataset analysed in the current study is not publicly available due to ethical considerations but is available from the corresponding author on reasonable request.

Role of the funding source

The funder of the study had no role in study design, data collection, data analysis, data interpretation or the writing of this report.

ResultsDescriptive analysis

Among those who reported sexual intercourse with at least one paid sexual partner over the lifetime, there was a median of six paid partners (IQR 3–17) among all men (aged 20–49 years), three paid partners (IQR 2–9) among men aged 20–29 years, five paid partners (IQR 3–12) among men aged 30–39 years and eight paid partners (IQR 3–20) among men aged 40–49 years (full distribution shown in online supplemental table 4). Among those who reported sexual intercourse with at least one paid sexual partner in the last year, there was a median of two paid partners (IQR 1–4) for all age groups and in total. The weighted results of the descriptive analysis, stratified by lifetime use of CSW services, are shown in table 2 and elaborated upon in the online supplemental table 1 stratified by age group.

Table 2

Sociodemographic, anthropometric and attitudinal characteristics of male NInJaS respondents aged 20–49 years, stratified by lifetime use of sexual services

Regression analysis

The weighted results of the univariate regression analyses are shown in table 3 and online supplemental table 2. Increasing age was statistically significantly associated with lifetime use of CSW services (aOR 1.07, 95% CI 1.06 to 1.08), with this association being most robust among men aged 20–29 years. Never-married men were significantly less likely than married men to have used CSW services (0.42, 0.36 to 0.48). Similarly, those who were unmarried but in a relationship (0.76, 0.58 to 0.98) or not in a relationship (0.39, 0.34 to 0.46) were significantly less likely than married persons to have used CSW services. Compared with heterosexual respondents, bisexual men (2.42, 1.63 to 3.60) were significantly more likely to have used CSW services, while asexual men (0.61, 0.45 to 0.81) were much less likely to have used CSW services. In all age groups, increasing income was monotonically associated with a significantly higher likelihood of having used CSW services. In all age groups other than men aged 20–29 years, undergraduate education was significantly associated with a higher likelihood of having used CSW services compared with those with high school education or less (1.36, 1.15 to 1.61). Region of residence and urbanity of residence were not significantly associated with lifetime use of CSW services, except among men aged 40–49 years living in Kyushu/Okinawa, who were significantly more likely to have used CSW services compared with those living in Hokkaido/Tohoku (2.16, 1.18 to 3.93).

Table 3

Sociodemographic, anthropometric and attitudinal characteristics associated with lifetime and past year use of sexual services among male NInJaS respondents aged 20–49 years (univariate age-adjusted-only ORs (aORs))

Anthropometric characteristics such as height, weight and BMI were not found to be significantly associated with lifetime use of CSW services in any age group with the exception of height among men aged 30–39 years, which was borderline significantly associated with a higher likelihood of using CSW services (1.02, 1.00 to 1.04). However, self-rated attractiveness of ‘about average’ (0.39, 0.30 to 0.51) or ‘below average’ (0.26, 0.20 to 0.35) was significantly associated with a lower likelihood of using CSW services compared with those who responded ‘above average’ in all age groups.

Compared with those who reported a neutral degree of satisfaction with their sex lives, men who reported satisfaction (strongly agree: 2.11, 1.46 to 3.05; agree: 1.80, 1.47 to 2.20) and dissatisfaction (strongly disagree: 1.33, 1.07 to 1.67; disagree: 1.82, 1.48 to 2.25) were both significantly more likely to have used CSW services in their lifetime. Those who reported that they would like to increase their frequency of sexual intercourse (increase a lot: 2.94, 2.41 to 3.60; increase a little: 2.02, 1.70 to 2.41) also were significantly more likely to have used CSW services compared with those who reported that their current frequency was amenable. Contrastingly, those who reported that they would like to considerably decrease their frequency of sexual intercourse were significantly less likely to have used CSW services (0.45, 0.29 to 0.72). Those who rated sex as being ‘very important’ (3.12, 2.51 to 3.89) or ‘somewhat important’ (2.29, 1.91 to 2.75) were significantly more likely to have used CSW services compared with those who rated sex as ‘neither important nor unimportant’; additionally, those who rated sex as ‘not important at all’ were much less likely to have used CSW services (0.48, 0.33 to 0.68).

The weighted results of the multivariate regression analyses are shown in table 4 and online supplemental table 3. As demonstrated with the univariate analyses, age, bisexuality, income, self-rated attractiveness, the desire to increase frequency of sexual intercourse and the self-rated importance of sex remained significantly associated with lifetime use of CSW across age groups. Occupational status and sexual satisfaction were not consistently significantly associated with lifetime use of CSW, unlike in the univariate analyses.

Table 4

Sociodemographic, anthropometric and attitudinal characteristics associated with lifetime and past year use of sexual services among male NInJaS respondents aged 20–49 years (multivariate ORs including all independent variables in a single model)

Discussion

The use of CSW services has significant implications for public health with respect to the spread of STIs. In Japan, rates of CSW service use are high among men, and these individuals constitute an important bridge between commercial sex workers and the general population.11 12 Using data from the NInJaS conducted in 2022, we present some of the first findings describing the sociodemographic, anthropometric and attitudinal characteristics of Japanese men aged 20–49 years who have used CSW services in their lifetime.

Almost half of men aged 20–49 years reported lifetime use of CSW services, which is a considerably larger proportion than those reported in previous studies conducted in Japan that were limited to specific subpopulations, such as patients with STI and college students.21 22 Prior research in Asia suggests that the use of CSW in Japan is roughly similar to that of South Korea and may even be surpassed by that of Taiwan (approximately 70% among men in 2001).23 Household-based surveys conducted in Asia suggest that approximately 5–10% of men purchased sex in the preceding year, including up to 11% of men in Hong Kong and 15% of men in Cambodia.24 Japan far outpaces other high-income, Western nations such as Sweden (men aged 16–84 years, lifetime: ~10%, 2017) and Germany (men aged 18–75 years, lifetime: ~27%, 2018/2019).25 26

The large proportion of Japanese men reporting CSW use is likely multifactorial. First, sex work is easily accessible and visible in Japan. Participation in the unofficial sex industry—though stigmatised—is also high, especially among young women who need a safety net income; in this survey, 11.8% of female respondents aged 18–29 years reported receiving money for sex via channels other than official sexual entertainment services.27 Second, the use of CSW by men is associated with an overall low stigma in Japan.22 28 In fact, the use of sex work is exceedingly common among Japanese men engaging in business trips23; male clientele crossing prefectural borders have been shown to exert a significant influence on CSW network structure in Japan.11 This is in part due to the long history of the government-sanctioned sex industry in Japan, which dates back to the Edo Period (1600s).1 Third, there is a multitude of different services available, appealing to a broad population of men who may harbour some interest in using CSW services.3 These various services also have a relatively long history, dating back to the implementation of the Prostitution Prevention Law, enacted in 1958, which explicitly banned penile–vaginal intercourse but paradoxically led to the proliferation of non-penetrative services.3 Notably, police regulation of the sex work industry is lax, even with respect to penetrative intercourse, though anal intercourse is not strictly forbidden from a legal standpoint.15

We found that socioeconomic variables were significantly associated with a lifetime history of having used CSW services. First, we found a robust and positive dose-dependent relationship between income and history of CSW use. Purchasing CSW services can be very expensive: using call services or soaplands can average US$100–200 for 60–90 min, while at the less expensive end, pink salons may cost approximately $25–50 for 30-minute sessions.29 Thus, a higher disposable income would permit the use of these expensive services, and similar associations have been found previously with respect to CSW use23 24 and virginity.18 A similar explanation is likely responsible for the relationship between stable employment and undergraduate education with a lifetime history of CSW use, as both full-time workers and highly educated men may have more financial resources at their disposal. Furthermore, married and cohabiting men were more likely to have used sexual services than their single or non-cohabiting counterparts, as has been observed in Australia.30 The associations with these socioeconomic variables (excluding education) remained significant even in the multivariate analysis controlling for all variables. Thus, in clinical settings, even partnered men should be asked about a history of sex work use when taking thorough sexual histories. With respect to sexual orientation, bisexual men were most likely to have used sex work services; importantly, this population constitutes an important bridge between men who have sex with men, women who have sex with men and sex workers. Notably, we did not note any consistent patterns in regional or urban versus rural use of sex work, though prior research has suggested it is more commonly used in urban areas.31 Of note, many of these same socioeconomic associations were found in the supplemental analysis using CSW use in the past year as the dependent variable.

Anthropometric data such as height, weight and BMI were not significantly associated with lifetime use of CSW services; however, high self-rated attractiveness was consistently significantly associated with a higher likelihood of CSW use. These associations persisted in the multivariate regression analysis, suggesting that factors other than income and occupation (eg, physical appearance) were considered when rating one’s own attractiveness (as the question intended). The reasons behind this association are unclear and merit further study.

All attitudinal variables were found to be associated with lifetime use of CSW services. Both men who reported satisfaction and dissatisfaction with their sex lives were more likely than those who reported neutrality to have used CSW services. As this analysis cannot delineate cause from effect, it is unclear whether men were satisfied because they used CSW or vice versa (the same applies to dissatisfaction). However, that the relationship is U-shaped merits further attention. Men who reported that they would like to increase their frequency of sexual intercourse and who rated sex as being important to them were also more likely to use CSW services. These associations are in line with the notion that the use of CSW may be intended to address self-identified deficiencies in one’s sex life, or alternatively may be akin to a hobby that one prioritises because sex is important to them irrespective of partnership status. Notably, the significance of the association with sex life satisfaction was lost in the multivariate regression, as there may be some collinearity with the other attitudinal questions. The concept of sexual satisfaction and well-being remains underdiscussed in the Japanese context, and additional research is needed to characterise these parameters and the mechanisms by which they are related to the use of sex work.

Limitations

This study has limitations. First, as the study is based on the results of an online survey, the results are subject to selection bias: only those with access to the internet who volunteered to take the survey in return for points were able to participate. However, we attempted to minimise this bias by calculating sample weights to make the sample population representative of the national population. Ultimately, the weighted sample was more educated and had slightly higher income compared with Japan’s National Fertility Survey, a nationally representative survey conducted in Japan in 2015.14 We suspect that a higher proportion of educated and high-income respondents may have inflated the amount of sexual activity (including CSW use) reported in the study, but it should have a limited effect on logistic regression analyses for which these variables are controlled. Second, the sensitive nature of the questions, as well as the need to consider lifetime practices, raise concern for social desirability bias and recall bias, respectively.31 The anonymous nature of the survey,32 as well as the relatively low stigma associated with sex work use in Japan, may have ameliorated the effects of social desirability bias when answering questions regarding use of commercial sex workers. Third, the survey was implemented during the COVID-19 pandemic, which was associated with social distancing and decreased sexual activity33; though this likely deflated the use of CSW services in the past year analyses, there should be a minimal effect on questions regarding the lifetime.

Conclusions

We characterise Japanese men aged 20–49 years who have a lifetime history of using CSW services and find that socioeconomic and attitudinal variables are strongly associated with having used these services before. Given the large proportion of Japanese men who have used sexual services and the importance of these men in bridging commercial sex workers and the general population, they constitute an important target of public health interventions. Moving forward, additional research is needed to delineate the STI transmission dynamics in this population, identify ideal testing and treatment strategies, and test whether CSW use is the primary driver for recent increases in bacterial STI notification rates. Furthermore, community-level outreach with CSW institutions is paramount to collect real-world data and enact meaningful change with respect to safe sex practices.

Abstract translationThis web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.Data availability statement

Data are available upon reasonable request. The dataset analysed in the current study is not publicly available due to ethical considerations but is available from the corresponding author on reasonable request.

Ethics statementsPatient consent for publication

Not applicable.

Ethics approval

This study involves human participants and was approved by the Regional Ethics Committee at the University of Tokyo, Japan (2019305NI-(2)). Respondents were required to provide consent before beginning the survey.