Background/Objectives Commercial sex work is a primary context for heterosexual HIV/AIDS transmission. Violence victimisation is considered to compromise women's ability to protect against HIV and other sexually transmitted infections (STI); little research has investigated violence as it relates to sexual risk and STI among female sex workers (FSW). This study sought to compare sexual risk and STI symptoms among FSW based on recent violence exposure.
Methods Data from 815 FSW in Thailand were used to assess the prevalence of physical or sexual violence within the context of sex work, and associations of victimisation with sexual risk and STI symptoms.
Results Approximately one in seven FSW (14.6%) had experienced violence in the week before the survey. Compared with their unexposed counterparts, FSW exposed to violence demonstrated a greater risk of condom failure (19.6% vs 12.3%, ARR 1.92, 95% CI 1.24 to 2.95) and client condom refusal (85.7% vs 69.0%, ARR 1.24, 95% CI 1.14 to 1.35). In analyses adjusted for sexual risk, violence related to STI symptoms collectively (ARR 1.11, 95% CI 1.02 to 1.21) and genital lesions as an individual STI symptom (ARR 1.78, 95% CI 1.20 to 2.66).
Conclusion Physical and sexual violence against FSW in Thailand appears to be common, with women experiencing such violence demonstrating diminished capacity for STI/HIV harm reduction and greater prevalence of STI symptoms. Efforts to reduce violence towards this vulnerable population must be prioritised, as a means of protecting the health and wellbeing of FSW, and as a key component of STI/HIV prevention and control.
- sex work
- sexual assault
- sexual risk
- STI symptoms
Statistics from Altmetric.com
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.
Sex work is widely recognised as a primary context for the heterosexual transmission of HIV/AIDS across the globe, including the majority of Asia.1 2 Within Southeast Asia, Thailand faces a disproportionate HIV burden; Thailand leads the region in the total number of persons living with HIV and demonstrates a population prevalence of 1.7%.3 The 100% Condom Use Campaign initiated in the 1990s within Thailand and across much of Southeast Asia is renowned for its success in promoting condom use among female sex workers (FSW) as well as reducing HIV.2 4 Despite these efforts, sexually transmitted infections (STI)/HIV persists among Thai FSW.3 5 Moreover, the persistence of inconsistent condom use, condom failure, pervasive client condom refusal, client pressure and coercion into unprotected sex, and other dimensions of FSW difficulty in negotiating condom use in Thailand6 7 and elsewhere1 8–11 indicate the critical need to identify and understand continued threats to FSW-initiated HIV prevention in order to reduce both incident STI/HIV infection and secondary transmission.
Violence is considered a likely threat to STI/HIV prevention among FSW,12 13 with estimates of the prevalence of physical and sexual violence victimisation among FSW ranging from 40% to 75%.1 12 14–16 Whereas violence alone cannot cause STI/HIV, it is increasingly considered relevant to women's HIV risk among the general population.3 17 Experiences of violence and fear of abuse are posited to represent critical barriers to women's ability to enact HIV risk-reduction behaviours,18 and empirical data demonstrate that violence compromises condom negotiation19 and confers STI/HIV risk.20 21
To date, little is known about the impact of violence exposure on STI/HIV risk and subsequent infection among FSW, nor have HIV prevention efforts been evaluated for efforts to address violence as a potential barrier to enacting protective sex behaviours. Despite qualitative evidence that violence against FSW often leads to STI/HIV risk (eg, higher-risk sex, compromised condom negotiation),12 22–24 limited quantitative data exist to assess the relationship of violence with sexual risk behaviours and subsequent infection among representative samples of FSW. What is known indicates that violence is associated with condom failure,14 client pressure for unprotected sex10 and HIV infection16 among FSW in other regions outside of Southeast Asia. Furthermore, no studies have been conducted into the role of violence in relation to a broad range of STI/HIV risks (ie, unprotected sex, condom failure, client condom refusal and anal sex) and STI outcomes that confer HIV vulnerability for FSW.
Although extensive violence faced by FSW has been documented, and the theoretical and empirical bases suggest that violence may confer significant STI/HIV risk, quantitative data concerning these intersections among FSW, both within Southeast Asia and elsewhere, are lacking. The current study is designed to inform the state of knowledge by evaluating the prevalence of recent physical and sexual violence victimisation, and associations of violence with sexual risk and STI symptoms among a large national sample of FSW in Thailand.
The current study utilises data collected between February and March of 2007 by the Institute for Population and Social Research at Mahidol University (Bangkok, Thailand), in collaboration with local non-governmental organisations in Bangkok, Chonburi, Chiang Mai and Songkhla, Thailand, for the purpose of describing sexual and reproductive health among FSW in Thailand.25 The sample was constructed to approximate the proportional number of sex workers in each province as reported by the STI Division, Ministry of Public Health of Thailand. Local non-governmental organisations staff within each province developed maps of sex work establishments to facilitate the recruitment and understanding of the distribution of sex workers and the types (ie, massage parlours, brothels, karaoke bars, etc) of sex work venues. Venues were then stratified based on the type and size and were randomly selected within strata. Managers from selected venues were identified and asked to permit the recruitment of FSW to participate in the study from these venues. Of 301 venues selected, 39 had closed or reported that they did not employ sex workers; of the remaining 262 venues, managers of 202 agreed to participate by allowing the recruitment of FSW, resulting in a venue participation rate of 77.1%. Of the 1025 participants approached, 815 agreed to participate, for a participation rate of 79.5%. Further details concerning sampling and survey methodology are available elsewhere.25 One participant did not provide data concerning violence, thus the final sample size for the current analyses is 814.
After verbal consent was obtained, the survey was verbally administered by an interviewer trained in survey interviewing techniques and gender and sexuality sensitivity. Surveys were administered in Thai in a private location of the sex work establishment or outside within the close proximity of the establishment. All items were self-reported and pilot tested before the start of data collection. The primary exposure, physical or sexual violence victimisation, was assessed via two items; participants indicating that they had been either ‘hit’ or ‘forced to perform sex acts (they) did not want to perform’ at work in the week before the survey were classified as having experienced physical or sexual violence. Single items assessed sexual risk behaviour, including anal sex in the past month, condom failure in the past week, condom non-use for vaginal sex over the past five sex acts, and lifetime history of client condom refusal. A syndromic STI assessment was conducted as a proxy for STI. Participants endorsing any of five STI symptoms in the vaginal or anal areas, specifically ‘genital lesions’, ‘warts’, ‘itchiness’, ‘lower abdominal pain’ and ‘pain when urinating’ in the past 4 months were classified as past-4 month STI symptomatic. The syndromic STI assessment, while limited in specificity for STI diagnosis, is recommended for settings that lack diagnostic facilities.26 In recognition of the role of ulcerative STI symptoms in conferring further STI/HIV risk,27 28 a secondary STI variable was constructed to reflect participants indicating genital lesions in the 4 months before the survey. Immediately following survey data collection, all surveys were checked for completion and by field supervisors and entered into EpiData version 2.0 (EpiData Association, Denmark, 2000–9) to facilitate validity and consistency checks. Data collection was carried out in conjunction with SWING (Service Workers IN Group Foundation), which facilitated participant referrals to any needed local health and violence-related support services. All procedures pertaining to the initial data collection were approved by the Institute for Population and Social Research of Mahidol University Human Subjects Committee. The current investigation consists of secondary analysis of these anonymously collected data, and as such was deemed exempt from review by the Harvard School of Public Health Human Subjects Committee.
Prevalence estimates were calculated for physical or sexual violence among the total sample and by demographic factors. Differences in violence history based on these factors were assessed by χ2 test; significance for all analyses was set at p<0.05. Prevalence estimates were calculated for all sexual risk factors and STI symptoms for the total sample and were based on violence exposure by χ2 test;. To evaluate differences in sexual risk based on exposure to violence, log-binomial regression29 models were used to estimate adjusted risk ratios (ARR) and 95% CI; models were adjusted for demographic characteristics found to be related to violence exposure in bivariate analyses (ie, current age, type of establishment in which the respondent works and recruitment province), with respondents indicating no recent violence serving as the referent group. Current age and duration in sex work were found to be collinear, thus duration in sex work was not included in the adjusted models. Finally, both violence exposure and sexual risk factors were evaluated for their relative contributions to recent STI symptoms, that is, genital lesions or any STI symptom. Differences in STI outcomes based on these factors were assessed by χ2 test; multivariate log-binomial regression models were constructed to estimateARR and 95% CI for the associations of violence victimisation and sexual risk behaviour with each STI outcome, adjusting for current age, type of establishment and recruitment province. All statistical analyses were conducted using SAS version 9.
Violence was prevalent among the current sample of FSW in Thailand, with an estimated one in seven (14.6%) reporting being the target of physical or sexual violence in the context of sex work in the week before the survey (table 1). Younger FSW (under 18 years) appeared to be at greatest risk of such experiences, with 25.0% affected compared with 17.7% among those aged 18–25 years and 12.1% among those 26 years and over (p=0.050). Violence against FSW varied regionally, with those in Bangkok and Chonburi most likely to be affected (20.7% and 16.5%, respectively; p<0.001). The highest victimisation prevalence was identified among FSW reporting freelance work (29.7%), with lower prevalence estimates (9.8–13.0%) identified within other settings (p<0.001).
FSW exposed to recent violence were more likely to report sexual risk and STI symptoms compared with their counterparts not facing such violence (table 2). Recent anal sex was more common among those exposed to violence compared with those unexposed (6.7% vs 2.6%, p<0.05); however, these differences appeared to be explained partly by demographic characteristics (ARR 2.11, 95% CI 0.92 to 4.82). Violence was associated with recent condom failure (19.6% of those experiencing violence compared with 12.3% of those not reporting recent violence; p<0.05); this difference persisted after considering demographic factors (ARR 1.92, 95% CI 1.24 to 2.95). Client condom refusal was common overall within the sample, and was similarly more prevalent among those reporting recent physical or sexual violence in the context of sex work (85.6% vs 69.0%, respectively; ARR 1.24, 95% CI 1.14 to 1.35). Both STI symptoms overall and lesions were more common among those FSW who experienced violence compared with those not reporting this abuse (84.9% vs 62.3%, ARR 1.31, 95% CI 1.18 to 1.45 and 21.9% vs 9.6%, ARR 2.30, 95% CI 1.51 to 3.52, respectively).
Several sexual risk factors demonstrated bivariate associations with STI outcomes (table 3). Both recent lesions and recent STI symptoms were more common among FSW indicating recent condom failure (20.0% vs 10.3%, p<0.05; 79.0% vs 62.6%, p<0.05, respectively). STI symptoms were also more common among those reporting client condom refusal (68.6% vs 57.6%, p<0.05). In the final multivariate models designed to evaluate the relative contributions of sexual risk and violence to STI outcomes, recent violence victimisation remained significantly associated with both lesions (ARR 1.78, 95% CI 1.20 to 2.66) and STI symptoms (ARR 1.11, 95% CI 1.02 to 1.21). Additional factors demonstrating independent associations with lesions included recent condom non-use (ARR 2.15, 95% CI 1.06 to 4.36) and recent condom failure (ARR 1.57, 95% CI 1.03 to 2.38). Recent condom failure and client condom refusal demonstrated trends towards increased risk of STI symptoms (ARR 1.09, 95% CI 1.00 to 1.16 and ARR 1.11, 95% CI 1.00 to 1.16, respectively).
The findings demonstrate extensive violence against Thai FSW in the context of sex work, with an estimated one in seven (14.6%) exposed to physical or sexual violence in the week before the survey. Moreover, such victimisation was significantly associated with sexual risk (ie, recent condom failure and client refusal of condoms) and STI symptoms, supporting consideration of the role of violence in increasing sexual risk and as a potential mechanism for the contraction and continued transmission of HIV/STI. The findings advance earlier qualitative evidence of violence against FSW leading to STI/HIV risk,12 22–24 and build on quantitative findings from other regions indicating that violence relates to condom failure, coerced unprotected sex and HIV infection.10 14 16 Coupled with these earlier findings, current evidence strongly indicates that violence against FSW is both common and, beyond immediate injuries, represents a critical threat to the sexual health of this vulnerable population, possibly undermining efforts to reduce STI/HIV within the context of commercial sex work.
Whereas violence itself cannot cause STI, current evidence that violence is independently associated with an increased risk of STI symptoms even after adjusting for sexual risk suggests that violence may well play a mechanistic role in such infection. For example, forced sex acts, as well as unwanted and coerced sex, may cause physical trauma (ie, tearing or lacerations)30 31 and thus facilitate STI acquisition. Experiences of abuse may also constitute a marker for qualitative differences in sexual risk not fully captured within the current investigation. For example, compromised condom negotiation based on fear of abuse and other forms of coerced condom non-use not currently assessed may facilitate STI acquisition.32 Despite the relatively low levels of recent condom non-use within this sample, condom failure was common, and coerced or compromised negotiation may contribute to unsuccessful condom use. Further research should include a more comprehensive assessment of sexual risk behaviours, including those that reflect power imbalances and coercion in this context. Findings may also reflect differences in STI risk among the male clients of FSW. In the light of elevated sexual risk behaviour and STI/HIV diagnosis demonstrated among male perpetrators of violence against female partners,33–36 male clients that perpetrate violence against FSW may similarly demonstrate a greater likelihood of STI infection and subsequent transmission to FSW.
Both the high prevalence of lifetime client condom refusal (71.5%), and the elevated sexual risk observed among FSW exposed to violence, add to the growing body of research illustrating client-related constraints on the ability of FSW to ensure successful condom use as a means to protect against STI/HIV.7 11 37 This further evidence that condom use is often not within the control of FSW strongly indicates the need for continued and expanded research and programmatic efforts targeting male clients of FSW in efforts to prevent both STI/HIV as well as violence within this context.
The high prevalence of past-week violence against FSW currently observed strongly indicates the urgent need for programmatic efforts to address and prevent violence against FSW in Thailand and other settings. The HIV prevention infrastructure targeting FSW, including the 100% Condom Use Campaign, widely endorses FSW empowerment and collectivisation,3 4 with these elements generally described as also providing a mechanism to reduce and respond to violence against FSW.13 38 Given the prevalence of violence victimisation currently observed (one in seven within the past week), coupled with its strong associations with STI risk, reduction of violence against FSW must be considered a major goal within STI/HIV prevention efforts. To date, FSW empowerment and collectivisation efforts have not been evaluated to determine their capacity to reduce violence and buffer its negative impact on sexual health. Programmatic efforts designed to reduce violence against women in the general population have also not been tested among FSW.39 Further research to evaluate the ability of the existing HIV prevention and violence prevention infrastructure to respond adequately to and prevent violence against FSW must be prioritised.
Several limitations of the current study design are important to consider, most notably the inability to establish a temporal relationship between violence and sexual risk or STI. The ability to make causal inferences is hindered by both the cross-sectional nature of the investigation as well as the varied time scales used for the exposure and outcomes of interest. Whereas the past-week assessment of violence victimisation probably limits recall error, any previous experiences of violence (ie, those earlier than the past week) may similarly pose an ongoing risk of sexual abuse and STI outcomes, thus the current assessment may bias estimates towards the null, rendering them conservative. As the current investigation reflects a secondary analysis of data collected to describe sexual and reproductive health concerns, with greater detail regarding violence exposure beyond the scope of the initial study, violence indicators available for the current study were limited. Future research to clarify the present findings will benefit from greater detail concerning the context of recent violence as well as its severity and perpetrators (eg, pimps, customers, police). Such an investigation may also consider other experiences of gender-based violence across the lifespan (eg, child sexual abuse, intimate partner violence). Migrants and ethnic minorities are thought to be overrepresented among FSW in Thailand40; the relatively small proportion of ethnic minorities in the current sample suggests potential sampling bias leading to limited coverage of such groups. Given the high levels of both violence and sexual risk noted among migrant FSW, particularly those from Burma,40 41 further investigation will benefit from efforts to include this subpopulation. The syndromic assessment of STI, although recommended for settings that lack diagnostic facilities,26 is limited in specificity, rendering the interpretation of findings limited regarding actual infection. Future research should include enhanced STI/HIV assessment (eg, integrated STI/HIV testing). Finally, further qualitative and quantitative work is needed to understand the nature of the associations identified, including the extent to which violence itself may be considered a direct barrier to women's ability to negotiate safe sex and reduce STI/HIV risk, whether violence may represent a direct cause of risk and infection (eg, via genital trauma), and the extent to which identified associations may reflect other mechanistic pathways.
A high level of physical and sexual violence victimisation was identified among FSW in the current sample, with one in seven experiencing violence in the week before the survey, and increased vulnerability to sexual risk exposures and STI symptoms associated with such victimisation. Findings add to the growing body of work demonstrating violence-related threats to women's risk of STI/HIV, both generally19 21 and specific to FSW.10 12–14 16 22 24 Violence may represent a critical threat to FSW-initiated STI/HIV prevention efforts, probably undermining the ability of the existing STI/HIV prevention infrastructure to impede further transmission. Implications of the current study are particularly relevant given the role of transactional sex in facilitating the continued spread of STI/HIV both within Asia and globally.1 2 Programmatic efforts to reduce violence against FSW must be prioritised in Thailand and elsewhere, both as a means of protecting the health and wellbeing of FSW, as well as reducing the continued spread of STI/HIV.
Physical and sexual violence were common among this sample of FSW in Thailand; approximately one in seven experienced violence in the week before the survey.
Experiencing recent violence was associated with sexual risk and STI symptoms among FSW.
Efforts targeting FSW to reduce STI/HIV risk must address violence based both on the high prevalence of such experiences, as well as observed associations with reduced capacity for STI/HIV prevention.
The authors gratefully acknowledge the Institute for Population and Social Research, Mahidol University, Bangkok, Thailand.
Funding Support for analyses and manuscript development was provided to MRD via the Harvard University Center for AIDS Research (HU CFAR NIH/NIAID fund P30-AI060354). Other funders: National Institutes of Health. Collection of the original survey data was initiated and funded by the United Nations Population Fund, Thailand.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.