Internet Solicitation Linked to Enhanced Occupational Health & Safety Outcomes Among Sex Workers in Metro Vancouver, Canada 2010–2019

Abstract

OBJECTIVES:

Examine the independent effect of online solicitation on sex workers’ occupational health and safety (OHS), particularly violence and work stress.

METHODS:

Data were drawn from a cohort of women sex workers (N=942, 2010 – 2019) in Vancouver, Canada. Analyses utilized descriptive statistics and bivariate and multivariable logistic and linear regression using generalized estimating equations (GEE); explanatory and confounder modelling approaches were used.

RESULTS:

33.9% (n=319) of participants solicited online and 14.1% (n=133) primarily solicited online in the last six months in at least one study visit. In multivariable GEE analysis, factors associated with primarily soliciting online included younger age (AOR 0.95 per year older, 95% CI 0.93–0.97), sexual minority status (AOR 2.57, 95% CI 1.61–4.10), gender minority status (AOR 3.09, 95% CI 1.80–5.28), higher education (AOR 2.13, 95% CI 1.34–3.40), higher sex work income (AOR 1.03 per $100 weekly, 95% CI 1.01–1.06), being an im/migrant to Canada (AOR 2.40, 95% CI 1.26–4.58), and primarily servicing in informal indoor workspaces (AOR 3.47, 95% CI 2.32–5.20). In separate GEE confounder models, primarily soliciting online significantly (1) reduced odds of physical/sexual workplace violence (AOR 0.64, 95% CI 0.39–1.06); and (2) reduced work stress (β coefficient −0.93, 95% CI −1.59 - −0.26).

DISCUSSION/CONCLUSIONS:

Younger workers, gender/sexual minorities, im/migrants, and those in informal indoor spaces had higher odds of soliciting online. Confounder models indicate access to online solicitation methods may support enhanced OHS. Decriminalization of sex work – including advertising via online platforms — remains necessary to support sex workers’ OHS.

INTRODUCTION

Sex work is an industry with numerous avenues of advertising (solicitation), including online.(1) Despite increasing internet usage, research on sex workers’ (SWs’) internet use at work is scarce. Research on SWs’ health has largely focused on street-based sex work,(2) with limited focus on occupational health and safety (OHS) of online workers. Such research is particularly needed within the context of COVID-19, as SWs report new ways of working to facilitate safety, with reported increases in online sex work (e.g. camming; online screening of clients).(3)

Numerous countries, including Canada, have implemented legislation that targets sex work clients, third parties, and online advertising platforms. In 2018 the United States (US) passed the Allow States and Victims to Fight Online Trafficking Act (FOSTA) and the Stop Enabling Sex Traffickers Act (SESTA), known as FOSTA-SESTA.(4,5) Its stated purpose is to make it illegal to knowingly assist, facilitate, or support sex trafficking, by excluding sex trafficking laws from immunity under the Section 230 of the Communications Decency Act. These steps were taken in reaction to successful use of Section 230 as a legal defense by Backpage, the largest online marketplace for buying/selling sex. Backpage’s shutdown, enacted through conspiracy and money laundering charges days before FOSTA-SESTA was signed into law, affected SWs worldwide(6). Media reports and advocacy groups suggest global impacts have been negative, including increased street-based sex work, increased risk-taking, loss of support networks, and disproportionate effects on sexual minority and gender minority SWs.(7)

In 2014 Canada introduced ‘end-demand’ legislation (the Protection of Communities and Exploited Persons Act (PCEPA)), after a legal decision overturned previous criminalized sex work laws which violated SWs’ rights.(8) While leaving sale of sex legal, the PCEPA criminalized new circumstances, including: purchasing sexual services; communication to provide sexual services in certain public areas; receiving financial or material benefits from SWs (e.g., third parties such as managers, security guards); and advertising sexual services(8). A February 2020 case pertaining to an escort agency in London, Ontario found these laws relating to purchase of sexual services, receiving financial or material benefits from SWs, and advertising are unconstitutional and violate SWs’ rights.(9) Both the PCEPA and FOSTA-SESTA limit SWs’ abilities to communicate critical OHS information (i.e., terms and conditions), creating opportunities for adverse effects on OHS (e.g., violence, conflicts).

Prior research indicates use of online services prior to meeting clients in person affords access to OHS mechanisms.(10) Benefits linked to online solicitation include: reduced female homicide rates;(11) increased feelings of control by workers;(10) flexibility with work schedules;(10,12) reduced stigma;(1,13) and increased empowerment at work, through risk management and information sharing.(10,13,14) Prior research examined population-level records (e.g. crime rates), publicly available online sex work advertising (e.g. ad content analysis), qualitative interviews, and online surveys.

Our analyses were informed by a conceptual framework of structural determinants of SWs’ OHS and intersectionality theory. Intersectionality theory, developed by Kimberlé Crenshaw,(15) posits that different elements of marginalization such as race, gender, sexual orientation, dis/ability, and socioeconomic status intersect to reflect broader macro-level axes of oppression. While SWs have been well documented to face inequities related to race, gender identity, sexual orientation, and socio-economic marginalization, limited research has used an intersectional lens.(16–18) Increasing literature demonstrates sexual minorities (e.g. lesbian, bisexual),(19) gender minorities (e.g. trans),(16) and racialized groups (e.g., im/migrants, Indigenous) are greatly overrepresented in sex work.(20,21)

Considering recent moves criminalizing advertising in Canada and takedowns and prosecution of US-based sex work solicitation websites used globally,(8) there is a scarcity of research on impacts of these policies. Given gaps in evidence regarding these impacts, the objective of this analysis was to examine the independent effect of online solicitation on OHS outcomes of violence and work stress, over a 9-year period amongst SWs in Metro Vancouver.

METHODS

Study design

Data were drawn from an ongoing open prospective cohort, An Evaluation of Sex Workers Health Access (AESHA), initiated in late January 2010. AESHA was developed based on community collaborations with sex work agencies since 2005 and is monitored and informed by a Community Advisory Board of representatives of 15+ community agencies.(22) Current eligibility includes being a self—identified woman (cisgender or transgender), 14 years and older, exchanging sex for money within the last 30 days, and providing written informed consent. Given challenges of recruiting SWs in isolated and hidden locations, time-location sampling is used to recruit SWs through day and late night outreach to outdoor/public sex work locations (i.e. streets, alleys) and indoor sex work venues (i.e. massage parlours, micro-brothels, incall locations) across Metro Vancouver.(23) Online recruitment is used to reach SWs working through online solicitation spaces. Indoor sex work venues and outdoor solicitation spaces (‘strolls’) are identified through community mapping conducted together with current/former SWs,(22) and are continually updated by the outreach team.

At enrolment and bi-annually, SWs complete a questionnaire administered by a trained interviewer (SWs and non-SWs) in English or Mandarin and HIV/STI/HCV serology testing by a project nurse. SWs visit our storefront office location in Metro Vancouver or complete the questionnaire and nursing component at their work or home location. Participants receive an honorarium of $40CAD at each visit for their time, expertise, and travel. Treatment is provided for symptomatic STI infections, and free serology and Papanicolaou testing are available for those that need them. The study holds ethical approval through Research Ethics Boards at Providence Health Care/University of British Columbia and Simon Fraser University.

Measures

Dependent variables of interest

Analyses began with explanatory models focused on two time-updated dependent variables capturing if participants did any online sex work solicitation and if they primarily solicited sex work online. Participants self-reported ways in which they solicited in the last six months at each semi-annual visit and identified their primary solicitation method. Participants may have supplemented other solicitation methods online (e.g. street-based sex work, licensed massage parlour), while those who primarily solicited sex work online may have used other solicitation avenues to supplement online methods. Subsequent analyses focused on primarily soliciting sex work online. Sensitivity analyses examined any online solicitation, as either a primary or supplementary solicitation avenue, to better understand the broader landscape of online solicitation and ensure robustness of findings.

Two time-updated OHS outcomes were examined in confounder models to investigate independent effects of online solicitation. These included experiences in the last six months of 1) physical or sexual violence by perpetrators posing as clients, or 2) work stress. Experiences of physical or sexual violence were assessed with a checklist of possible physically or sexually violent encounters from predators posing as clients, including abduction or kidnapping, rape, sexual assault, strangulation, physical assault, being locked or trapped in a car or room, being thrown out of a moving vehicle, assault with a weapon, being drugged, or ‘Other’. Work stress was assessed using an abbreviated 13 item version of Karasek’s standardized Job Content Questionnaire, used in Statistics Canada’s nationally representative Canadian Community Health Survey, adapted to sex work.(24) Responses to such measures as ‘You had the materials and equipment you needed to do your job (e.g. condoms, regular STI tests)’ were measured on a 5-point Likert scale ranging from ‘strongly agree’ to ‘strongly disagree’, with ‘not applicable’ given a score of 0. The maximum possible score for the continuous scale was 65, with higher scores indicating higher work stress. Development and use of this scale by AESHA has been detailed previously.(25)

Explanatory variables

Variables of interest at individual, workplace, and macrostructural levels were examined. Time-fixed variables included Indigenous ancestry (First Nations, Métis, or Inuit), identifying as a sexual minority (e.g. gay, lesbian, bisexual, queer, asexual, Two-Spirit), identifying as a gender minority, being an im/migrant to Canada (vs Canadian-born), and education level (high school completion vs less than high school). Remaining variables were time-updated at each semi-annual study visit to capture current measures or events in the last 6 months. These included age, residency in Vancouver’s inner city, the Downtown Eastside (DTES) (vs living in other parts of Metro Vancouver), non-injection drug use (excluding alcohol and cannabis), injection drug use, average weekly income from sex work ($CAD), and primary place of service (outdoor or public spaces such as cars; informal indoor such as worker’s or client’s home or hotel; or formal indoor such as a licensed massage parlour). Interview year was included to account for changes in internet use over time.

Statistical Analyses

Explanatory analyses of online solicitation and confounder models of physical/sexual violence were conducted among 942 participants interviewed from January 2010 — February 2019; analyses of the work stress outcome were restricted to 775 participants who were asked these questions from August 2010 – February 2019. All analyses were restricted to study visits where participants did sex work in the last six months.

Descriptive statistics at baseline were calculated for independent variables of interest, stratified by participants’ experiences with online solicitation. Differences were assessed using the Mann-Whitney test for continuous variables and Pearson’s chi-square test (or Fisher’s exact test for small cell counts) for categorical variables.

Bivariate and multivariable logistic regression using generalized estimating equations (GEE) were used to evaluate correlates of primarily soliciting sex work online; an exchangeable correlation structure was used to account for repeated measures among the same individuals. Variables hypothesized to be associated with soliciting online and those that were significant at p<0.10 in bivariate analysis were considered for inclusion in the full multivariable model. A manual backward model selection process was used to obtain the model with the best overall fit, as indicated by the lowest quasi-likelihood under the independence model criterion.(26) Sensitivity analyses examining any online solicitation for sex work used the same approach.

Finally, two separate multivariable confounder models were constructed to examine independent effects of primarily soliciting sex work online on: (1) physical or sexual violence by perpetrators posing as clients; and (2) work stress. Physical or sexual violence was analyzed with logistic regression using GEE with an exchangeable correlation structure. Work stress was analyzed using linear regression, also with GEE and an exchangeable correlation structure. The same model fitting process was employed for both models. All potential confounders identified in bivariate analyses were included in the full models. Using the process described by Maldonado and Greenland,(27) potential confounders were removed from the full models in a stepwise manner, and variables that altered the association of interest by <5% were systematically removed from the model. All statistical analyses were performed in SAS version 9.4 (SAS, Cary, NC) and all p-values are two-sided with statistical significance at <0.1.

RESULTS

Explanatory analyses of online solicitation (Tables 1–3) and our confounder model of physical/sexual violence (Table 4) are based on 942 participants (4756 observations). Median number of study visits completed by participants was 3 (interquartile range (IQR) 1–8). Analyses of the work stress outcome were restricted to a sub-sample of 775 participants who responded to the scale after its addition in August 2010 and contributed 3979 observations.

TABLE 1.

Baseline characteristics of women sex workers (N=942) in Metro Vancouver stratified by primarily soliciting sex work online in last six months (2010–2019)

Characteristic	Total (%) (n = 942)	Primarily solicited sex work online		p - value
		Yes (%) (n = 56)	No (%) (n = 886)
Individual factors
Age (years; median, IQR)	35 (28–42)	33.5 (26.5–43)	35 (28–42)	0.322
Indigenous*	358 (38.0)	25 (44.6)	333 (37.6)	0.294
Sexual minority	301 (32.0)	32 (57.1)	269 (30.4)	<0.001
Gender minority	58 (6.2)	11 (19.6)	47 (5.3)	<0.001
Im/migrant to Canada	277 (29.4)	11 (19.6)	266 (30.0)	0.097
Graduated high school	524 (55.6)	36 (64.3)	488 (55.1)	0.179
Currently live in DTES†	300 (31.9)	24 (42.9)	276 (31.2)	0.083
Non-injection drug use‡**	627 (66.6)	41 (73.2)	586 (66.1)	0.281
Injection drug use**	382 (40.6)	26 (46.4)	356 (40.2)	0.356
Average weekly income from sex work ($ CAD; median, IQR)**	500 (250–1000)	675 (200–1200)	500 (250–1000)	0.400
Structural factors
Primary place of service**
Outdoor/public spaces	360 (38.2)	10 (17.9)	350 (39.5)
Informal indoor	268 (28.5)	37 (66.1)	231 (26.1)
Formal indoor	297 (31.5)	8 (14.3)	289 (32.6)	<0.001
Occupational health and safety
Any physical/sexual violence**	166 (17.6)	7 (12.5)	159 (18.0)	0.334
Work stress (point scale; median, IQR) **•	34 (31–38)	32 (30–35)	35 (31–38)	<0.001

^*First Nations, Métis, Inuit

^†Downtown Eastside; inner city of Vancouver

^‡Excluding alcohol and cannabis

^**Instances in the last six months

^•N=775 participants responded to the work stress scale, which was added in August 2010, after the study’s initiation

All data refer to n (%) unless otherwise specified

Table 3.

Bivariate and multivariable GEE analysis of any soliciting sex work online in the last six months among women sex workers (N=942) in Metro Vancouver, 2010–2019

Characteristic	Unadjusted		Adjusted
	Odds Ratio (95% CI)	p – value	Odds Ratio (95% CI)	p - value
Age (per year older)	0.96 (0.95 – 0.98)	<0.001	0.96 (0.95 – 0.98)	<0.001
Indigenous* (yes vs. no)	0.91 (0.70 – 1.19)	0.507
Sexual minority (yes vs. no)	2.74 (2.10 – 3.57)	<0.001	1.78 (1.26 – 2.50)	0.001
Gender minority (yes vs. no)	3.83 (2.51 – 5.83)	<0.001	3.12 (1.91 – 5.11)	<0.001
Im/migrant to Canada (yes vs. no)	0.43 (0.30 – 0.61)	<0.001
Graduated high school (yes vs. no)	1.50 (1.15 – 1.95)	0.003	1.86 (1.32 – 2.62)	<0.001
Currently live in DTES† (yes vs. no)	0.84 (0.71 – 0.99)	0.038	0.74 (0.55 – 0.99)	0.044
Non-injection drug use‡** (yes vs. no)	1.44 (1.19 – 1.75)	<0.001
Injection drug use** (yes vs. no)	1.04 (0.88 – 1.23)	0.660
Average weekly income from sex work (per $100 CAD) **	1.04 (1.02 – 1.06)	<0.001	1.08 (1.05 – 1.10)	<0.001
Primary place of service **
Informal indoor (vs. outdoor/public space)	1.31 (1.13 – 1.53)	<0.001	1.85 (1.43 – 2.40)	<0.001
Formal indoor (vs. outdoor/public space)	0.48 (0.34 – 0.68)	<0.001	0.38 (0.24 – 0.60)	<0.001
Interview year (per year)	1.01 (0.98 – 1.04)	0.391

^*First Nations, Métis, Inuit

^†Downtown Eastside; inner city of Vancouver

^‡Excluding alcohol and cannabis

^**Time updated with instances in the last six months

Table 4.

Bivariate and multivariable GEE confounder analysis of the relationship between primarily soliciting sex work online and physical/sexual occupational violence among women sex workers (N=942) in Metro Vancouver (2010–2019).

Characteristic	Unadjusted		Adjusted
	Odds Ratio (95% CI)	p – value	Odds Ratio (95% CI)	p - value
Experiences of physical or sexual violence by perpetrator posing as client *
Primarily solicit online	0.63 (0.40 – 0.98)	0.041	0.64 (0.39 – 1.06)	0.082

^*Multivariable confounder model adjusted for age, sexual/gender identity, primary place of service, average weekly income from sex work, and interview year.

At baseline, 5.9% (n=56) of participants primarily solicited sex work online in the last six months and 19.1% (n=180) engaged in any online solicitation (i.e. as primary or supplementary solicitation) (Table 1). During the study, 33.9% (n=319) of participants had solicited online within the last six months of at least one study visit and 14.1% (n=133) identified primarily soliciting online in the last six months in at least one study visit.

At baseline, median age of all participants was 35 years old (IQR 28–42), 38.0% (n=358) were of Indigenous ancestry (First Nations, Métis, and Inuit), and 29.4% (n=277) were im/migrants to Canada. One-third of participants identified as a sexual minority (32.0%, n=301) and 6.2% as a gender minority (n=58). Over half of all participants were high school graduates (55.6%, n=524) and participants made an average weekly income of $500 CAD (IQR $250–1000 CAD) from sex work. Over one-third serviced primarily in outdoor/public spaces (e.g. cars) (38.2%, n=360), 28.5% (n=268) in informal indoor spaces, and 31.5% (n=297) in formal indoor spaces. At baseline, 17.6% (n=166) of participants had experienced any physical or sexual violence from perpetrators posing as clients in the last six months. Among the subsample, the median work stress score was 34 (IQR 31–38). Baseline characteristics of women SWs in Metro Vancouver stratified by any online sex work solicitation in last six months are available as Supplementary Table 1.

In multivariable GEE explanatory analysis (Table 2), factors associated with primarily soliciting sex work online included younger age (adjusted odds ratio (AOR) 0.95 per year older, 95% confidence interval (CI) 0.93–0.97), identifying as a sexual minority (AOR 2.57, 95% CI 1.61–4.10) or gender minority (AOR 3.09, 95% CI 1.80–5.28), im/migration to Canada (AOR 2.40, 95% CI 1.26–4.58), graduating high school (AOR 2.13, 95% CI 1.34–3.40), higher average weekly income from sex work (AOR 1.03 per $100 CAD, 95% CI 1.01–1.06), servicing clients in informal indoor environments (AOR 3.47, 95% CI 2.32–5.20), and interview year (AOR 1.24 per year, 95% CI 1.16–1.31). In sensitivity analysis (Table 3), the same factors were significantly associated with any online solicitation, except being an im/migrant to Canada and interview year.

Table 2.

Bivariate and multivariable GEE analysis of primarily soliciting sex work online in the last six months among women sex workers (N=942) in Metro Vancouver, 2010–2019

Characteristic	Unadjusted		Adjusted
	Odds Ratio (95% CI)	p – value	Odds Ratio (95% CI)	p - value
Age (per year older)	0.98 (0.96 – 1.00)	0.049	0.95 (0.93 – 0.97)	<0.001
Indigenous* (yes vs. no)	1.09 (0.74 – 1.62)	0.668
Sexual minority (yes vs. no)	3.61 (2.44 – 5.36)	<0.001	2.57 (1.61 – 4.10)	<0.001
Gender minority (yes vs. no)	4.63 (2.80 – 7.66)	<0.001	3.09 (1.80 – 5.28)	<0.001
Im/migrant to Canada (yes vs. no)	0.76 (0.47 – 1.23)	0.266	2.40 (1.26 – 4.58)	0.008
Graduated high school (yes vs. no)	2.00 (1.33 – 3.01)	<0.001	2.13 (1.34 – 3.40)	0.002
Currently live in DTES† (yes vs. no)	0.95 (0.74 – 1.23)	0.686
Non-injection drug use‡** (yes vs. no)	0.90 (0.68 – 1.20)	0.472
Injection drug use** (yes vs. no)	0.75 (0.57 – 0.99)	0.044
Average weekly income from sex work (per $100 CAD) **	1.02 (1.00 – 1.04)	0.015	1.03 (1.01 – 1.06)	0.022
Primary place of service **
Informal indoor (vs. outdoor/public space)	3.01 (2.21 – 4.11)	<0.001	3.47 (2.32 – 5.20)	<0.001
Formal indoor (vs. outdoor/public space)	0.79 (0.42 – 1.48)	0.460	0.57 (0.22 – 1.44)	0.232
Interview year (per year)	1.16 (1.11 – 1.21)	<0.001	1.24 (1.16 – 1.31)	<0.001

^*First Nations, Métis, Inuit

^†Downtown Eastside; inner city of Vancouver

^‡Excluding alcohol and cannabis

^**Time updated with instances in the last six months

In separate multivariable GEE confounder models, primarily soliciting sex work online was linked to a significantly reduced odds of experiencing physical or sexual violence by an aggressor posing as a client (Table 4; AOR 0.64, 95% CI 0.39–1.06), as well as reduced work stress (Table 5; β coefficient −0.93, 95% CI −1.59—−0.26) after adjustment for key confounders.

Table 5.

Bivariate and multivariable GEE confounder analysis of the relationship between primarily soliciting sex work online on work stress among women sex workers (N=775) in Metro Vancouver (2010–2019).

	Unadjusted		Adjusted
Characteristic	β coefficient (95% CI)	p – value	β coefficient (95% CI)	p – value
Work stress **
Primarily solicit online	−1.31 (−2.00 – −0.61)	<0.001	−0.93 (−1.59 – −0.26)	0.006

^**Multivariable confounder model adjusted for age, education, primary place of service, average weekly income from sex work, and interview year.

DISCUSSION

Our analysis documented a significant independent protective association of primarily soliciting sex work online on reduced odds of experiencing physical or sexual workplace violence and reduced work stress after adjustment for key confounders. Findings indicate im/migrants, younger SWs, workers with higher educational attainment and income, those in informal indoor workplaces, gender minorities, and sexual minorities had higher odds of primarily soliciting sex work clients online. Results suggest the importance of creating ‘enabling environments’ for SWs to work online, including removal of criminalization relating to online sex work.

One review suggested reductions in HIV among SWs could be achieved by addressing structural determinants, including macrostructural (e.g. legal), community organization (e.g. empowerment), work environment (e.g. venues), and individual or interpersonal characteristics (e.g. behaviour).(28) This review suggested decriminalization would have the largest impact, averting 33–46% of HIV infections in the next decade.(28) This is supported by a unique study of temporarily (2003–2009) decriminalized indoor sex work in Rhode Island. Causal effects of decriminalization of indoor sex work included 30% decrease in rape offences and 40% decrease in gonorrhea incidence, supporting evidence that structural determinants affect SWs’ OHS.(29) These studies contribute evidence that SWs’ OHS outcomes are, at least in part, shaped by structural determinants beyond SWs’ behaviour. Website takedowns have resulted in displacement online and despite increased punitive measures and censorship, volume of online sex work ads has increased.(7,30) Relevant locally, the British Columbia Centre for Disease Control included transitioning to online work as a strategy for safer sex work in the context of COVID-19.(31)

Online solicitation facilitates SWs’ autonomy, self-empowerment, and professionalism,(10,32) potentially resulting in the reduced work stress we observed. Previous research suggests online solicitation affords key risk management options for SWs. In addition to OHS-enhancing screening methods, SWs can accommodate scheduling variety, ignore client requests when uninterested, and work independently,(10,12,33) further contributing to reduced work stress. Online work is also linked to reduced stigma – a well-identified macrostructural determinant of SWs’ OHS; privacy and anonymity online may contribute to destigmatization.(1,13)

At a work environment level, SWs in informal indoor locations were found to have over three-times higher odds of primarily soliciting online. Indoor spaces are safer workplaces, facilitating OHS and other supports.(29) Working indoors provides privacy from policing, potentially contributing to reduced work stress, particularly for trans women.(34) In the Rhode Island study, mentioned above, decriminalization of indoor sex work resulted in improved OHS.(29)

Use of online services provides access to safety mechanisms afforded by communication prior to meeting clients in person.(10) An analysis on effects of Craigslist’s ‘erotic services’ (ERS) section (used almost exclusively by SWs), which expanded from 2002 to 2010 before being shut down, concluded that ERS reduced female homicide rates 10–17% by reducing street-based sex work and SWs screening out the most dangerous perpetrators posing as clients.(11) Communication begins with advertising, which provides ‘terms and conditions’: parameters of services; workers’ boundaries, restrictions, and expectations (e.g. condom use); and rates.(10,13,14,17) Once clients establish contact, SWs are able to screen for safety; online screening methods include: assessing communication style; checking online profiles; and use of webcams prior to meeting in person.(13,16,32) In addition, resources such as search engines, social media (e.g. LinkedIn), phone number verification services, and client verification services provide options for assessing clients. If a client does not pass screening, online solicitation offers distance when declining service.(13,32) In addition, some online settings (e.g. Twitter) allow SWs to interact, facilitating information exchange, destigmatization, and empowerment.(35)

Our study found several marginalized and intersecting identities, including im/migrants, younger SWs, sexual minorities, and gender minorities, had higher odds of relying on online spaces as their primary method of soliciting clients. As in previous studies,(36) older SWs had lower odds of soliciting online, likely due to normalization of internet technologies over time. Prior research has shown im/migrant workers in Vancouver and elsewhere are overrepresented in managed formal indoor sex work spaces (i.e. massage parlours),(2) where they work with third parties, whose administrative and security services provide similar protections to online screening,(37) including advertising sexual services for workers. Im/migrant SWs have been shown to experience heightened work stress relating to police encounters,(2,38) language barriers, and barriers to health care access;(2,38) access to safer options for work, including online advertising, remains particularly vital for these SWs.

Participants with higher educational attainment and income had higher odds of primarily soliciting sex work online, suggesting certain SWs have access to resources (e.g. computers) that offer protections from violence and reduce work stress. Online solicitation may reproduce class-based hierarchies, wherein sex workers with access to financial resources access safer, more lucrative sources of sex work income. Prior research that has found education is a correlate of higher sex work income.(39)

Sexual minority and gender minority sex workers face greatly increased odds of workplace, institutional, and community violence, including homophobia or transphobia, stigmatizing or non-responsive policing experiences when reporting violence, and displacement and resulting vulnerability to violence and police harassment.(34) In one study, trans women were the highest users of risk management messages when advertising online, resolving risks of systemic oppression individually.(17) Trans women have reported violence when rushed to disclose their gender or when misgendered by potential clients.(34) Thus, trans SWs without online access due to lack of finances, resources, or education may be at added risk of violence and greater work stress.

Recommendations

Structural changes addressing multiple determinants are needed to enact long-term, sustainable changes to improve SWs’ OHS and human rights. We propose:

• Decriminalization of all aspects of sex work to remove legal barriers to OHS mechanisms (e.g. advertising) and promote SWs’ health and human rights.(28) The World Health Organization, among others, has called for all countries to work towards decriminalization of sex work.(40) Towards this end, we recommend repeal of the PCEPA in Canada and repeal of FOSTA-SESTA in the US.

• Legislatively loosening restrictions regarding communications with implications for SWs’ OHS (e.g. services, restrictions, condom use, rates).

• To meet SWs’ unique needs for advertising and screening skills, we recommend services for SWs include internet safety programs.

• Seeking younger SWs, im/migrants, sexual minorities, and gender minorities, particularly trans women, when drafting OHS guidelines for online SWs. These SWs have higher odds of soliciting online, and thus higher odds of having relevant expertise.

• Partnerships between sex work advocacy organizations and non-profit electronics refurbishment and recycling organizations to facilitate SWs’ access to internet technologies, as well as trainings on their use.

Strengths and limitations

AESHA represents a cohort of SWs across diverse venues and online workers are under-represented within our sample; explanations may include limited interpersonal contact for recruitment, fear of law enforcement, and that income foregone by higher income participants to participate in research is unlikely to be covered by research honoraria. Our research is based on self-reported data, which may result in recall and social desirability biases. Interviewers include experiential staff and are trained in trauma-aware techniques and developing rapport with participants to help mitigate limitations. Our research was limited to SWs in the Metro Vancouver area and may not be generalizable to all SWs or SWs in other settings. Considering growing availability of internet access and related growth in online sex work solicitation, further research focused on understanding the experiences of online SWs is recommended.

CONCLUSIONS

Primarily soliciting sex work online was associated with reduced odds of physical or sexual violence from aggressors posing as clients and reduced work stress. These findings add to evidence suggesting SWs use screening tools available online to manage risk through advertising, screening, and scheduling. Our findings suggest primarily soliciting online is a safer solicitation avenue than alternate methods (e.g. street-based solicitation, formal licensed indoor establishment).

Access to online solicitation methods improves SWs’ OHS and is an important tool in mitigating risk and reducing stress for SWs. Decriminalization of all aspects of sex work, including advertising, is needed to support SWs’ OHS.

What is already known about this subject?

Prior research has indicated that sex workers’ occupational health outcomes are affected by structural influences including criminalization of sex work, yet we know little about the occupational health outcomes of online sex workers, including work stress and violence.

What are the new findings?

An analysis of 9 years of data from an ongoing longitudinal study suggests sex workers who solicit through online advertising experience reduced physical/sexual workplace violence (AOR 0.64, 95% CI 0.39–1.06) and reduced work stress (β coefficient −0.93, 95% CI −1.59—−0.26).

How might it impact on clinical practice in the forseeable future?

Sex workers’ occupational health outcomes can be improved through policy and practice interventions including decriminalization of all aspects of sex work (including online advertising) and improved internet access for sex workers via access to technology and computer literacy education initiatives.