Article Text
Abstract
Background This study aims to report on a newly developed Safer Indoor Work Environmental Scale that characterises the social, policy and physical features of indoor venues and social cohesion; and using this scale, longitudinally evaluate the association between these features on sex workers’ (SWs’) condom use for pregnancy prevention.
Methods Drawing on a prospective open cohort of female SWs working in indoor venues, a newly developed Safer Indoor Work Environment Scale was used to build six multivariable models with generalised estimating equations (GEE), to determine the independent effects of social, policy and physical venue-based features and social cohesion on condom use.
Results Of 588 indoor SWs, 63.6% used condoms for pregnancy prevention in the last month. In multivariable GEE analysis, the following venue-based features were significantly correlated with barrier contraceptive use for pregnancy prevention: managerial practices and venue safety policies (adjusted OR (AOR)=1.09; 95% CI 1.01 to 1.17), access to sexual and reproductive health services/supplies (AOR=1.10; 95% CI 1.00 to 1.20), access to drug harm reduction (AOR=1.13; 95% CI 1.01 to 1.28) and social cohesion among workers (AOR=1.05; 95% CI 1.03 to 1.07). Access to security features was marginally associated with condom use (AOR=1.13; 95% CI 0.99 to 1.29).
Conclusions The findings of the current study highlight how work environment and social cohesion among SWs are related to improved condom use. Given global calls for the decriminalisation of sex work, and potential legislative reforms in Canada, this study points to the critical need for new institutional arrangements (eg, legal and regulatory frameworks; labour standards) to support safer sex workplaces.
- Cohort studies
- LONGITUDINAL STUDIES
- Measurement tool Development
- POLICY
- SOCIAL FACTORS IN
Statistics from Altmetric.com
Introduction
There is a growing body of evidence globally of high levels of unmet reproductive and sexual health needs among women sex workers (SWs).1–3 Emerging research from low and middle income countries (LMIC) has demonstrated high rates of unintended pregnancies and poor access to contraceptives and other reproductive health services.4–7 The data from higher income settings are sparse; however, recent research from Canada found similarly low rates of access to sexual and reproductive health services, including among pregnant and parenting SWs, and high rates of unintended pregnancies alongside poor access to barrier and non-barrier contraceptives for pregnancy prevention.8
There is increasing global recognition of sexual and reproductive health inequities experienced by SWs.1 ,9 ,10 Barriers to safely negotiating and using condoms reflect an important feature of those inequities and are largely attributed to structural factors, particularly violence and threats of violence, and criminalisation and stigmatisation of sex work.11 Increasingly, socioecological and structural determinant frameworks are being used to map how SWs’ ability to negotiate health risks and protections are affected by structural factors, including: (A) macrostructural factors (eg, laws, policies); (B) community organisation (eg, sex work organising, community empowerment, social cohesion) and (C) work environment features (eg, physical features, venue-based policies, managerial practices).12–14 Concomitantly, there has been a shift towards complementing existing biomedical and behavioural approaches with structural and community-led interventions that account for both proximal and distal structural determinants in shaping SWs’ health outcomes.15 ,16 Though previous research on structural determinants has focused on venue and community empowerment interventions,17 ,18 this has been almost exclusively drawn from LMIC and to our knowledge, only a handful of studies have examined the multiple influences of supportive workplace models and social cohesion on barrier contraceptives usage.14 ,19 ,20
In Vancouver, Canada, despite the criminalisation of sex work, unsanctioned indoor sex work venues have long existed, as formal/‘in-call’ sex work establishments (eg, licensed massage parlours, beauty parlours, microbrothels) as well as informal or ‘out-call’ indoor venues (eg, bars, hotels, saunas). Emerging qualitative and ethnographic studies in high-income settings have demonstrated that social, policy and physical features of indoor venues and social cohesion among SWs may promote reduced violence and increase control over condom negotiation and use with clients.21–23
While a handful of studies in LMIC have examined social, policy and physical features of indoor venues, these have almost exclusively focused on HIV prevention-related features and policies (eg, workplace condom rules, manager support for condom use, price of condoms) on HIV risk.14 ,19 ,20 ,24 There remains a gap in the epidemiological literature on the influence of a broader range of workplace features, including: broader healthcare services access (eg, drug harm reduction, sexual and reproductive health services), violence prevention measures and policies,25 as well as their intersections, on SWs’ sexual and reproductive health practices. Given the absence of quantitative data on the influence of workplace models and social cohesion on SWs’ condom use for pregnancy prevention, particularly in high-income settings, this study aimed: (1) to catalogue social, policy and physical features of indoor venues to develop a safer work environment scale; and (2) to longitudinally evaluate the relationship between higher scores on the safer indoor work environment scale and social cohesion on SWs’ (male and female) condom use for pregnancy prevention in a prospective cohort of SWs in Vancouver, Canada.
Methods
Study design
Data from January 2010 to February 2013 were drawn from a longitudinal cohort known as An Evaluation of Sex Workers’ Health Access (AESHA), developed based on and monitored by a Community Advisory Board of over 15 community and sex work agencies. Eligibility includes cisgender and transgender women who exchanged sex for money within the past 30 days in street, indoor and online venues. As described previously,26 time-location sampling was used for recruitment through day and late night outreach to outdoor locations (ie, streets, alleys), indoor venues (eg, formal/‘in-call’ sex work establishments, eg, massage parlours, microbrothels; and informal or ‘out-call’ venues, eg, bars, hotels, saunas) and self-advertising spaces (eg, online, newspapers) across Metro Vancouver. Sex work venues were identified through community mapping conducted with current/former SWs and are updated regularly by the outreach team. A combination of outreach methods, and contact by mobile phone and internet were used for follow-up. Consenting SWs completed an interview-administered questionnaire by a trained interviewer at baseline and semiannual follow-up, together with an interviewer-administered nursing questionnaire on health and social supports.
The main interview questionnaire asks about individual, interpersonal and sex work patterns (eg, condom negotiation, number and characteristics of clients, fees/types of sexual services), work environment features (eg, management policies, contraceptives access, policing, intimate partner and workplace violence) and macrostructural factors (eg, changes in polices and laws, migration, stigma experiences). The nursing questionnaire asks about experiences, access and barriers to health and social supports, including sexual and reproductive health. All participants receive an honorarium of $C40 at each biannual visit for their time, expertise and travel. The study holds ethical approval from the Providence Health Care/University of British Columbia Research Ethics Board.
Outcomes: Given previous research from this cohort documenting low use of modern contraceptives (ie, birth control pills, intrauterine devices (IUDs), injectable hormones, rings and diaphragms),8 this study focused on condom use for pregnancy prevention. The primary outcome was based on a ‘yes’ response to ‘male condom’ and/or ‘female condom’ to the question “what type of contraceptives have you used for pregnancy prevention in the past month?” Participants were asked to list all contraceptives used in the last month, which also included: birth control pills, depo-provera, IUDs, vaginal rings, diaphragms, spermicides, emergency contraceptives or permanent contraceptives. This outcome was time-updated at every 6-month follow-up.
Primary exposure variable: scale development and internal validity
Early formative and qualitative research with SWs21 ,22 led to the inclusion of a diverse set of questions on the social, policy and physical features of indoor work environments (eg, in-call and out-call venues) within the AESHA questionnaire, which were then catalogued to develop the Safer Indoor Work Environment Scale (see table 2 for the list of venue features). To make use of longitudinal data, all factors were considered as time-updated variables at every 6-month interview, over the 3-year follow-up period. Based on the overwhelming distribution of responses to all questions favouring ‘always’ or ‘never’, item responses were dichotomised (ie, ‘always’, ‘usually’, ‘occasionally’ and ‘sometimes’ vs ‘never’). Consistent with a structural determinants framework, following factor analysis, the indoor work environment features were grouped as follows: (1) social and policy venue-based features, which included supportive venue-based policies and managerial practices; (2) physical venue-based features included: access to sexual and reproductive health services/supplies; access to drug harm reduction services/supplies; and access to physical security features, where relevant.
Scale items were assigned scores based on a review of the literature from LMIC, in-depth qualitative research conducted with AESHA participants,21 ,22 and consultations with the AESHA staff and community partners. Work environment features found to be ‘supportive’ of condom use were given a score of 1 vs 0. The descriptive statistics for each item, including their distributions, were examined. Exploratory factor analysis using varimax rotation was used to determine the number of factors present among the items, using a maximum likelihood method. Factors were retained that: had Eigenvalues of greater than 1; collectively accounted for 70–80% of the variance; and preceded the elbow in a Scree plot. Factor loadings were used to determine the number of items included within each factor. Additionally, Tucker and Lewis’s Reliability Coefficient yielded a score of close to 1 (0.924). Finally, Latent class analysis (specifying one class), Cronbach's α and Kuder-Richardson scores (for dichotomous variables) were used to assess internal consistency within each subscale. The final scale consisted of 19 venue-based items and is composed of five subscales (including the social cohesion scale):
Social and policy venue-based subscale: SWs were asked if their indoor workplace had in place policies and managerial practices to support SWs’ safety and control in negotiation transactions with clients, displayed in table 2.
Physical venue-based subscale: Physical venue-based subscale reflects three components: (A) access to sexual/reproductive health services and supplies; (B) access to drug harm reduction services and supplies; and (C) access to physical security features (see table 2).
Community organisation and empowerment subscale: Lippman and Kerrigan and colleagues’ Social Cohesion Scale27 was used to score the level of social cohesion among workers (eg, perceptions of peer supportiveness, trust and mutual aid), based on responses ranging from ‘strongly agree’ to ‘strongly disagree’. Lippman and Kerrigan and colleagues’ social cohesion scale is described in detail elsewhere,27 and has been previously adapted and validated with SWs in LMIC.28
Similar to work in LMIC,27 ,28 item responses to the presence of each indoor feature (1 vs 0) were summed for bivariate and multivariable longitudinal analyses. A combined score for all venue subscales was created, giving equal weight to each subscale. The primary exposure variables (The Safer Indoor Work Environment Scale and its subcomponents) were time-updated at every 6-month follow-up to account for changing work environments over time and their association with condom use for pregnancy prevention (in the last month).
Longitudinal regression analyses: Owing to low levels of missingness (<5%), a listwise deletion approach was taken for missing data. To determine if workplace features and community organisation were independently correlated with SWs’ condom use longitudinally, multivariable analyses using generalised estimating equations (GEE) were conducted. A working correlation matrix was also used to help account for repeated measures by the same respondent over 3 years of follow-up.
A series of confounding models were constructed (one for each subscale, and another for the combined scale) using an approach described by Rothman and Greenland,29 for a total of six models. Confounders were chosen based on a priori knowledge of associations with condom use for pregnancy prevention, and a statistically significant bivariate GEE correlation with our outcome (p<0.20). In addition, variables were also considered confounders if they altered the association of interest by 10%. All potential confounders were included in a full model. Backwards elimination was used to arrive at the final model. SAS statistical software package V.9.3 was used for all data analyses (SAS Institute, Cary, North Carolina, USA).
Results
Sample characteristics
Of a total of 646 SWs enrolled in AESHA, this analysis was restricted to 588 (86.1%) SWs who had worked in indoor venues over the 3-year follow-up. As the primary outcome was condom use for pregnancy prevention, SWs reporting sterilisation were excluded. The median age of the sample was 35.1 (IQR 28.0–42.0), over one-third (37.9%, n=223) of women were of Aboriginal ancestry and 23.6% (n=139) were immigrant/new migrant SWs, primarily from Asia. Overall, 66.0% were visible minorities and 34.0% were Caucasian, with approximately half (48.6%) of the participants having completed high school or some form of higher education. While 45.5% of the sample intended on becoming pregnant, over half also reported condom use for pregnancy prevention (63.6%). At baseline, among the 374 women who used condoms, all reported male condom use and 12 (2%) reported using female condoms as well (see table 1 for details). All participants had complete data on work environment features, and 1.02% (n=6) were missing among the outcome, though this did not vary between exposure and non-exposure groups.
Table 2 describes the properties of the subscales of the Safer Indoor Work Environment Scale which had satisfactory Cronbach's α and Kuder-Richardson scores. All subscales had Cronbach's α scores of >0.85, indicating a high level of internal consistency. There was one exception—the physical security features component of the physical venue-based subscale had a Cronbach's α of 0.698. The Cronbach's α for the combined scale was 0.914.
Work environment scores and condom use for pregnancy prevention
In multivariable GEE analyses, increasing scores for all but one of the work environments were significantly independently correlated with increased odds of condom use for pregnancy prevention. With every one-point increase on the combined scale, there was a 2% increase in odds of condom use (adjusted OR (AOR)=1.02; 95% CI 1.01 to 1.04). As displayed in table 3, the AORs aligned with increasing scores for each work environment subscale as follows: managerial practices and venue safety policies (AOR=1.09; 95% CI 1.01 to 1.17); access to sexual and reproductive health services and supplies (AOR=1.10; 95% CI 1.00 to 1.20); access to drug harm reduction services and supplies (AOR=1.13; 95% CI 1.01 to 1.28); as well as Lippman and Kerrigan and colleagues’ Social Cohesion Scale (AOR=1.05; 95% CI 1.03 to 1.07). Access to physical security features was marginally associated (AOR=1.13; 95% CI 0.99 to 1.29).
Discussion
This study quantifies the relationship between intersecting social, policy and physical venue-based features and social cohesion on SWs’ condom use for pregnancy prevention. The findings point to the role of venue and social cohesion in affecting SWs’ sexual and reproductive health. The study also offers a multicomponent Safer Indoor Work Environment Scale that can be used to measure the degree to which social cohesion combined with a broad range of venue features can create ‘supportive’ or enabling environments conducive to SWs’ health and safety. This Safer Indoor Work Environment Scale also has the potential to inform the development of workplace models that better support safer, more supportive sex work models. Only a handful of studies have examined the more complex and intersecting influences of structural factors on SWs’ condom use. In particular, studies from the Dominican Republic, Brazil, China and the Philippines have shown how workplace models that support HIV prevention, including venue-based sexual health policies, managerial practices (eg, managerial support for condom use) and physical layout (eg, access to condoms; health and social supports) shape negotiation of HIV risk.19 ,24 ,27 ,30 Using the Safer Indoor Work Environment Scale, our study builds on these studies and demonstrates that, alongside HIV prevention venue features and practices, a combination of social, policy and physical features that support broader sexual and reproductive health, drug harm reduction and safety, combined with higher levels of social cohesion among workers, may increase SWs’ condom use.
The implementation of venue-level managerial practices and safety policies was independently correlated with SWs’ condom use. Every one-point increase in the social and policy venue-based subscale corresponded with a 9% increased odds of condom use. Similarly, increased access to physical security features also enhanced condom use. These findings likely reflect SWs’ improved control over transactions, reduced fear of violence and enhanced ability to negotiate for condoms by clients in venues that prevent violence through safety features, policies and practices. Results from the current study also confirm findings from earlier qualitative research in this setting, in unsanctioned indoor venues within supportive women-only housing,22 as well as in licensed health enhancement and massage parlour venues.21
Physical access to sexual and reproductive health services and supplies at a venue level was also directly correlated with SWs’ condom use over a 3-year period. Other studies also have found evidence linking access to HIV prevention services and supplies, and services to consistent condom use.31–34 Our research adds to this by demonstrating that it is the synergy of social, policy and physical features at venues that promotes SWs’ condom use over time. Physical access to relevant services and supplies also may reflect important features of venue-level social relations whereby SWs, managers and peers collectively support and facilitate condom use.19 ,30 Qualitative studies on sex work policies and management in India, China and brothels in Nevada, also have reported that managers may represent important ‘nodes’ for sexually transmitted infection (STI) information, and provide protection from violence, contributing to increased condom use.23 ,35 ,36
At the community level, increasing scores for social cohesion among SWs27 were positively correlated with increased condom use. Social cohesion has been described as a ‘structural component’ of community empowerment,37 which captures trust, mutual aid and solidarity among SWs. The role of community empowerment and social cohesion on SWs’ condom use has been well documented globally, including in WHO guidelines. However, these data are almost exclusively drawn from LMICs, such as Brazil, Dominican Republic and India. These are settings where there has been significant investment in resources to support these features among SW communities.24 ,31 ,35 Such investments have not been nearly as prevalent in higher income settings (including Canadian communities) and point to an area where evidence-informed investment in novel actions and supports for SW communities could be beneficial, from a health and safety perspective. Only a handful of other studies have measured specific and intersecting facets of community empowerment, most notably data from Brazil and Dominican Republic,24 ,28 and, recently, Swaziland.38 Moreover, our results longitudinally demonstrate how a combination of social, policy and physical venue-based and community organisation characteristics (specifically, social cohesion) can promote SWs’ sexual and reproductive health.
Finally, increased access to drug harm reduction services and supplies was also associated with increased odds of condom use. Previous research has shown that among SWs who use drugs, increased control over drug preparation reduces SWs’ risk of violence (or threat of violence) at the time of drug use and increases control over negotiation of both sexual and drug risks.45 While substance use among street-based SWs and its association with inconsistent condom use have been documented,40 ,41 our findings help to address an evidence gap regarding the potentially positive effects of drug harm reduction interventions on SWs’ sexual and reproductive health. Drug harm reduction interventions and their apparent capacity to offer positive, value-added effects on the sexual and reproductive health outcomes of SWs, is a promising area of future research and practice.
Limitations
As with all research with stigmatised and criminalised populations, the clandestine nature of sex work makes identifying a sampling frame and randomly selecting a sample of representative participants challenging. To address this limitation, time-location sampling and social mapping with team members (with previous or current experience in sex work) were used that systematically sample SWs at times and venues where they work.42 Additionally, we did not measure consistency/frequency of condom use, and therefore cannot extrapolate these findings to hypothesise their effect on pregnancy outcomes or HIV transmission rates. We were unable to assess SWs’ relationships with their managers (eg, support and trust), which has been documented as a key social and venue feature in other studies.14 ,24 ,30 ,43 Additionally, we were unable to measure the influence of economic venue-based features (eg, service fees, payment structure, income) on condom use. Together, these features represent areas for research in future iterations of the scale.
As our analysis was focused on condom use for pregnancy prevention, our outcome may not capture SWs who consider condoms solely as a HIV prevention practice rather than pregnancy prevention method. Contraceptives and pregnancy can be a sensitive topic for already stigmatised populations and thus, as with all self-reported data, responses may be subject to social desirability bias. There is a possibility of over-reporting of condom use among SWs in venues that encourage condom use with clients (and conversely under-reporting in environments unsupportive of condom use), which may bias our results. However, given the study's protocols to ensure patient confidentiality and interviewers’ good rapport with participants, this bias is expected to be minimal. The interviewers also have strong community linkages, with some having sex work experience themselves. As well, our study findings are specific to SWs in a higher income setting, but they under-represent higher income SWs in our setting (eg, escort and independent workers). Therefore, our study findings should be interpreted cautiously across sex work contexts.
Conclusions
This study is among the first to examine the longitudinal relationship of safer indoor work environments and social cohesion on SWs’ condom use in a high-income setting. The results provide key implications for future sexual and reproductive health policy and programming, including HIV/STI prevention. Specifically, this research highlights the role of venue-based policies and managerial practices, as well as access to health and social supports, on SWs’ condom use for pregnancy prevention. Substantial evidence exists in Canada and globally that criminalisation and unsafe work environments reduce SWs’ ability to safely negotiate condom use due to fear of violence and arrest. Safer indoor venues with supportive policies and practices, together with strong policy support for SWs to work collectively, represent structural avenues through which to promote and protect SWs’ safety and health. Furthermore, these results suggest that even within contexts that feature substantial barriers for SWs to access healthcare,44 supportive sex workplaces may also help by providing such services themselves or by facilitating better access to appropriate supplies, services and programmes, including condoms.
The WHO/UN is calling for decriminalisation of sex work globally.45 In 2013, the Supreme Court of Canada struck down sex work laws that prohibited (among other things) working in indoor settings, or the hiring or accessing of managers and other third parties, such as security. This study indicates that safer workplace models that include supportive venue and management practices (eg, security and access to other health resources and services) are key to SWs’ health and safety. The evidence documented in the current study provides crucial insights to inform a new legal/policy framework. Furthermore, the findings shed light on how safer workplace models and the ability of SWs to work together remains a critical conduit to promoting SWs’ condom use.
What is already known on this subject?
Structural interventions (eg, supportive work environments and community empowerment) that extend beyond solely biomedical and/or behavioural interventions have been identified as a cornerstone for HIV prevention, including contraceptive use, among sex workers.
Qualitative research among sex workers has highlighted the important role that venue-level social and structural interventions play (eg, supportive policies, management and services) on barrier contraception negotiation and use.
Quantitative evidence from high-income settings that disentangle the independent effects of various venue-level environments and sex worker cohesion on broader reproductive health of sex workers, including condom use for pregnancy prevention is lacking.
What this study adds?
This study offers a novel ‘Safer Indoor Work Environment Scale’ that catalogues a range of venue-level work environment factors and examines their relationship with sex workers’ condom use.
Study findings highlight the importance of structural venue-based (eg, supportive management and policies, onsite access to sexual and reproductive health services) and social cohesion among workers in promoting condom use among sex workers.
This study supports global calls for decriminalisation to support the ability of sex workers to more formally collectivise and ensure access to safer work environment models that are essential in promoting condom use.
Acknowledgments
The authors thank all those who contributed their time and expertise to this project, including participants, partner agencies and the AESHA Community Advisory Board. They wish to acknowledge Peter Vann, Calvin Lai, Eric Fu, Ofer Amram, Jill Chettiar, Alex Scot and Kathleen Deering for their research and administrative support. We thank Dr Melanie Rusch and Adriana Nóhpal de la Rosa for their consultations regarding scale development. We would also like to thank Dr. Stefan Baral, Dr. Ruth Elwood-Martin and Dr. Sam Sheps for their review and feedback on earlier versions of this manuscript.
References
Footnotes
Contributors KS had access to the data and takes full responsibility for the integrity of the data. PD and KS developed the analyses plan with advice from SD and JC. PD conducted the statistical analyses, wrote the first draft of the manuscript and integrated suggestions from all authors. All authors made significant contributions to the interpretation of the data, drafting of the manuscript, and all authors approved the final manuscript.
Funding US Department of Health and Human Services-National Institutes of Health (R01DA028648). Government of Canada-Canadian Institutes of Health Research (HHP-98835). PD was supported by PHIRNET (Population Health Interventions Research), an initiative of CIHR, and the Liu Institute for Global Issues at the University of British Columbia. KS is supported by the Michael Smith Foundation for Health Research and the Canadian Institutes of Health Research.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.