The relationship between violence and engagement in drug dealing and sex work among street-involved youth.

Hayashi, Kanna ; Daly-Grafstein, Ben ; Dong, Huiru 等

Typically unstably housed, street-involved youth are vulnerable to an array of health-related harms, including violence and the associated adverse health consequences. (1) The number of street-involved youth in Canada has increased dramatically since the 1980s. (1,2) Recent national estimates indicated that approximately 1 in 230 Canadians used an emergency shelter per year, and of these, youth aged 16-24 accounted for 20%. (2,3) In addition, the size of the "hidden homeless" population (i.e., non-shelter users) is likely much bigger, with a conservative estimate suggesting that it is three times that of shelter users. (3)

In general, youth are almost 15 times more likely to be victims of violence than the general adult population in Canada. (4) The rates of experiencing violence are further elevated among street-involved youth; previous studies reported that 82% of homeless youth in Toronto had experienced violence during the past year, and 48% of street-involved youth in Vancouver had experienced violence during the previous six months. (5,6) While immediate adverse health consequences of victimization include physical injuries or death, experiencing violence has also been associated with an array of psychiatric and psychological disorders, suicidal behaviour, substance use, and high-risk sexual behaviour, all of which may have a lasting impact on mortality and morbidity throughout adulthood. (7) These negative health sequelae of violence have prompted the World Health Organization to call for greater public health efforts to prevent violence and associated harms. (7)

Previous research suggests that for street-involved youth, not just being homeless but a variety of other factors as well are associated with their experiencing violence. For instance, street-involved youth are known to have high rates of substance use, and frequent alcohol use has been associated with falling victim to violence. (6) Further, it has been shown that youth who have experienced childhood abuse or mental health disorders are more likely to be victims of violence. (8) Another key social factor that has been consistently identified as a significant correlate of experiencing violence is involvement in informal income-generation activities, such as drug dealing and sex work. (5,6,8) Previous studies have suggested that limited employment opportunities condition street-involved youth to resort to informal income-generation activities for subsistence, which in turn serve to constrain their access to safety measures and increase their vulnerability to violence. (5,9)

While many factors have been indicated as risk factors for experiencing violence and might be interrelated, few studies have estimated the independent impact of informal income-generation activities on the likelihood of being a victim of violence, adjusting for the effect of other potential individual, social, structural and environmental confounders. As well, the relative contributions of different income-generating activities remain understudied. Such a study would help inform the development and refinement of interventions to reduce violence among vulnerable street-involved youth. As such, we sought to examine the relationship between engagement in drug dealing and sex work and being a victim of violence among street-involved youth who use illicit drugs in Vancouver, British Columbia.

METHODS

Data for this study were derived from the At Risk Youth Study (ARYS). ARYS is an open prospective cohort study of street-involved youth who use illicit drugs in Vancouver. The study methods have been described in detail previously. (10) In brief, study recruitment involves extensive street-based outreach and snowball sampling. To be eligible, participants must: provide written informed consent, have used illicit drugs other than marijuana in the previous month, be between the ages of 14 and 26, and be street-involved. Street involvement is defined as being absolutely or temporarily without stable housing, or having accessed street-based youth services in the previous six months. At baseline and subsequent semi-annual follow-up interviews, participants complete an interviewer-administered questionnaire, which includes items on socio-demographic information, drug use patterns, sexual and drug-related risk behaviours, and engagement with the criminal justice system and health and social services. After each visit, participants are provided with compensation in the amount of $30 CAD. The ARYS cohort has been approved by the University of British Columbia/Providence Healthcare Research Ethics Board.

For the current longitudinal data analysis, the sample was restricted to ARYS participants who had completed a baseline interview between September 1, 2005 and May 31, 2014. We used data collected at both baseline and any subsequent follow-ups in the analysis. The outcome of interest was having experienced violence in the previous six months (yes vs. no), ascertained through the question "Have you been attacked or assaulted (including sexual assault), or suffered any kind of violence in the past six months?" The variable definition was consistent with a previous study. (6) For those who experienced violence, we also asked them what was (or were) the type(s) of violence and who was (or were) the perpetrator(s).

The variable selection for this study was informed by the "Risk Environment" framework, which suggests that a range of individual, social, structural and environmental factors interact with each other to shape the production of drug-related harm. (11) The primary explanatory variable of interest was involvement in drug dealing and/or sex work in the past six months. This variable had four categories: involvement in both drug dealing and sex work; involvement in drug dealing but not sex work; involvement in sex work but not drug dealing; and as the reference category, involvement in neither of the two. As in a previous study, sex work was defined as having exchanged sex for gifts, food, clothing, shelter, money or drugs. (12)

We also selected a range of secondary explanatory variables that we hypothesized might confound the relationship between involvement in drug dealing and/or sex work, and being a victim of violence. These included such demographic characteristics as year of age, ancestry (Caucasian vs. other), and being in a stable relationship at the time of the interview (i.e., being married or having a regular partner) (yes vs. no), as well as a self-reported history of diagnosed mental illness (any vs. none). As previous studies have suggested that high-intensity and/or risky substance use patterns might increase vulnerability to violence as well as prolong involvement in informal income-generation activities, (6,13,14) we considered substance-using behaviours as potential confounders. These included injection or non-injection use of heroin, cocaine, crack cocaine and methamphetamine respectively (all: daily vs. <daily); injection drug use (yes vs. no); binge drug use, defined as having an episode of using drugs via injection or non-injection more than usual during the previous six months (yes vs. no); binge alcohol use, defined as having an episode of consuming alcohol more than usual during the previous six months (yes vs. no); and receiving assistance with injecting drugs (yes vs. no). Other social, structural and environmental factors included homelessness (yes vs. no), residency in Vancouver's Downtown Eastside (i.e.,the city's open drug scene epicentre) (yes vs. no), incarceration (yes vs. no) and childhood emotional abuse assessed at baseline with the Childhood Trauma Questionnaire (CTQ) subscale (moderate to extreme [a CTQ subscale score of [greater than or equal to] 13] vs. none to moderate [a CTQ subscale score of <13]). A subscale for the childhood emotional abuse was chosen based on a previous study showing an independent association with being a victim of violence. (15) All behavioural variables referred to the previous six months unless otherwise stated, and were treated as time-varying variables.

As a first step, we examined sex-based differences in baseline characteristics of our sample, using the Pearson's [chi square] test (for categorical variables) and the Wilcoxon rank sum test (for continuous variables). Since there were significant differences in the primary explanatory variable and the outcome by sex, all subsequent univariable and multivariable analyses were stratified by sex. Next, we used the generalized estimating equation (GEE) for the binary outcome with logit link to examine the univariable and multivariable associations with the aforementioned primary and secondary explanatory variables and being a victim of violence. This method of regression modeling allowed us to account for correlation between covariates from the same individual over time and examine the population-averaged effects of each variable on experiencing violence independently. To deal with uneven follow-up, we used the exchangeable working correlation structure, which assumes that the correlation between any pair of measurements on the same individual is the same. GEE models also assumed that missing assessments were missing completely at random. To account for possible confounding and calculate the best effect estimate, we fit a multivariable model using an a priori-defined modeling strategy proposed by Greenland et al. (16) We used a conservative p-value of 0.10 in the univariable analyses to determine whether a secondary explanatory variable was considered as a potential confounder in the relationship between the primary explanatory variable and the outcome and for inclusion in a full multivariate model. Then, reduced models were constructed in a manual stepwise approach, removing a single secondary explanatory variable each time. Specifically, the value of the coefficient for the primary explanatory variable in the full model and each reduced model was compared, and the secondary explanatory variable corresponding to the smallest relative change was removed. The process was repeated until the smallest relative change in the coefficient for any category of the primary explanatory variable from the full model was greater than 5%. In a sub-analysis, we replaced the primary explanatory variable with an interaction term for sex work and drug dealing involvement (and dummy variables for these two activities) in the final multivariable models to examine whether the effect of drug dealing (or sex work) on violence might differ depending on the involvement in sex work (or drug dealing). All p-values were two-sided. All statistical analyses were performed using SAS software version 9.3 (SAS, Cary, NC).

RESULTS

A total of 1,152 participants were included in the study. Of these, the median age at baseline was 21.8 years (interquartile range [IQR]: 19.8-23.6), and 364 (31.6%) were female. In this sample, 368 (31.9%) had a baseline assessment only, and the remaining 784 (68.1%) were followed for a median of 24.6 (IQR: 13.4-53.1) months. At baseline, 531 (46.1%) individuals reported having experienced violence in the past six months, and 740 (64.2%) reported having experienced violence at some point during the study period. The most commonly reported types of violence experienced included beating (71.7%), attacked with weapons (club, knife, belt, etc.) (19.5%), robbery (rolling for drugs or money) (7.4%), strangled (5.5%), and sexual assault including rape (3.9%). The most commonly reported perpetrators of violence consisted of strangers (45.8%), police officers (18.8%), acquaintances (16.3%), friends (15.3%), and intimate partners/ex-partners, including husband/wife, boyfriend/girlfriend and regular sex partner (13.5%). Table 1 provides further information on the baseline characteristics of the sample stratified by sex. As shown, at baseline, males were more likely than females to experience violence and engage in drug dealing but not sex work in the previous six months, while females were more likely to engage in both drug dealing and sex work in the previous six months (all p < 0.05).

The results of the univariable and multivariable GEE analyses of factors associated with experiencing violence are shown in Table 2. As shown, among females, compared to the reference category (i.e., involvement in neither of the two), involvement in drug dealing only was the only category that remained independently associated with experiencing violence in the final multivariable GEE model (adjusted odds ratio [AOR]: 1.43; 95% confidence interval [CI]: 1.08-1.90). Among males, involvement in drug dealing only (AOR: 1.50; 95% CI: 1.25-1.80), and involvement in both drug dealing and sex work (AOR: 1.74; 95% CI: 1.00-3.03) remained independently associated with experiencing violence in the final multivariable GEE model. In sub-analysis, the interaction between drug dealing and sex work involvement was not statistically significant among females (p = 0.505) or males (p = 0.618).

DISCUSSION

We found a high prevalence of violence among our sample of drug-using, street-involved youth in Vancouver, BC, with two thirds having been a victim of violence during an average of two years follow-up. In multivariable analysis, after adjusting for a range of individual, social, structural and environmental factors, involvement in drug dealing but not sex work was independently associated with violence among both males and females. Among males, involvement in both drug dealing and sex work was also independently associated with being a victim of violence; this was not the case among females. In sub-analysis, involvement in sex work did not show statistically significant modified effect on the relationship between drug dealing and violence among both genders, suggesting that engagement in drug dealing is likely driving the observed association with violence.

Our prevalence of past-6-month experiences of violence (64%) was slightly lower than that of past-year experiences of violence (82%) identified among homeless youth in Toronto. (5) However, it was much higher than the lifetime prevalence of experiencing violence (42%) reported among inner-city high school students in Baltimore, Maryland (USA), who were referred for mental health care, (17) and that (46%) reported among 32-year-old African Americans living in socially disadvantaged neighbourhoods in Chicago, Illinois (USA). (18) As well, it was higher than the prevalence (48%) reported in our previous cross-sectional study from the same cohort, (6) indicating that experiencing violence is common among this population. Taken together, our findings reinforce the previous call for attention to high rates of violence among street-involved youth (1,5) by providing more recent, longitudinal data.

Consistent with previous research, (5) we found that street-involved youth who were engaged in drug dealing were more likely to experience violence. Our findings extend previous research by demonstrating that this particular socio-economic factor increases the likelihood of experiencing violence, regardless of a range of other individual, social, structural and environmental markers of risk of violence. While many people who use drugs are known to engage in drug dealing to sustain their drug use, (19,20) for some individuals, drug market involvement preceded the initiation of illicit drug use, which may be the case for some street-involved youth. (12,21) Among street-involved youth engaging in drug dealing in this setting, the majority (79%) have been shown to assume a role of a direct drug seller, followed by a "middler" (i.e., coordinating a deal between a dealer and client) (28%) and a "holder" (i.e., carrying drugs during a drug trade) (6%). (19) All these roles operate at the lowest end of the drug market hierarchy and thereby increase vulnerability to violence, including encounters with violent customers or police officers, and the risk of being punished when unable to fulfill assigned quotas. (20,22,23)

The findings that involvement in sex work on its own was not independently associated with experiencing violence and that sex work involvement did not modify the effect of drug dealing on violence were unexpected given a large body of research indicating high rates of sexual and physical violence among sex workers around the world. (24) This may be due to the relatively small number of participants who reported being engaged in sex work in our study. It is noteworthy that in our study, the association between engagement in sex work and violence appeared stronger among males than among females. Literature on male sex workers is scarce and explanations for this difference are not apparent from our data. (24) Future research should seek to conduct a more in-depth investigation of experiences and gendered dynamics of violence among street-involved youth who engage in sex work in this setting.

Our findings indicate a need for novel interventions to prevent violence among youth who are involved in the drug trade. It is possible that those who engage in drug dealing may be prone to violence via gang affiliations and therefore interventions focusing on preventing gang involvement may be effective. (25) However, prior investigation found that while male street-involved youth were significantly more likely than their female counterparts to have a history of gang involvement, males with gang ties were not more likely to experience physical violence, indicating that gang involvement likely does not account for the higher risk of violence among males involved in drug dealing in this setting. (26) Continued investment in established evidence-based interventions such as addiction treatment that help youth exit street life is a promising approach to prevent drug dealing and subsequent increased risk of violence; (14,26) however, the level of violence observed among youth in our study indicates that current approaches are not effective enough in preventing violence among vulnerable youth, and novel interventions for those who actively use drugs and engage in drug dealing are needed.

Although previous studies report that approximately half of street-involved youth who both used and dealt drugs in this setting showed willingness to cease dealing drugs, those who were deeply entrenched in drug addiction were less likely to cease drug dealing. (14,19) As well, addiction treatment has not shown consistent impacts on employment outcomes, (27) suggesting a need for novel approaches to transition youth away from drug dealing towards less risky forms of income generation. Low-threshold employment opportunities that are readily accessible for those who have ongoing drug use is one promising approach. (28) The potential for low-threshold opportunities to attract youth away from drug dealing, thereby preventing related harms, including violence, warrants further investigation.

Further, as a growing body of research suggests that the criminalization of drug use makes people who use drugs vulnerable to drug dealing and violence, (29) decriminalization of personal drug use may serve to promote transitions from informal income-generation activities to formal employment among street-involved youth. Future research should investigate whether such transitions result in reductions in violent victimization among this population.

Limitations

There are some limitations with this study that are of importance. First, the ARYS cohort is not a randomly generated sample. Therefore, the sample may not be representative of the street-involved youth in Vancouver or elsewhere. Furthermore, our data relied on self-reported responses, allowing for the possibility of biases in the participants' answers. This may be due to socially desirable reporting, and conceivably lead to an underestimation of reports of violence or informal income-generation activities. Nonetheless, self-reported data have been commonly utilized in observational studies involving both adult and youth drug-using populations and found to be valid. (30)

CONCLUSION

We found that experiences of violence were common among our sample of street-involved youth, and independently associated with involvement in drug dealing. These findings indicate that in addition to conventional interventions, such as addiction treatment, novel approaches are needed to reduce the risk of violence for drug-using youth who are actively engaged in drug dealing.

REFERENCES

(1.) Boivin J-F, Roy E, Haley N, Galbaud du Fort G. The health of street youth: A Canadian perspective. Can J Public Health 2005; 96(6):432-37. PMID: 16350867.

(2.) Gaetz S. Coming of Age: Reimagining the Response to Youth Homelessness in Canada. Toronto, ON: Canadian Homelessness Research Network Press, 2014.

(3.) Gaetz S, Donaldson J, Richter T, Gulliver T. The State of Homelessness in Canada 2013. Toronto, ON: Canadian Homelessness Research Network Press, 2013.

(4.) Perreault S, Brennan S. Criminal victimization in Canada, 2009. Juristat 2010; 30(2). Available at: http://www.statcan.gc.ca/pub/85-002-x/85-002-x2010002-eng.htm (Accessed October 22, 2015).

(5.) Gaetz S. Safe streets for whom? Homeless youth, social exclusion, and criminal victimization. Can J Criminol Crim 2004; 46(4):423-56. doi: 10.3138/ qccj.46.4.423.

(6.) Martin I, Palepu A, Wood E, Li K, Montaner J, Kerr T. Violence among street-involved youth: The role of methamphetamine. Eur Addict Res 2009; 15(1):32-38. PMID: 19052461. doi: 10.1159/000173007.

(7.) Krug EG, Mercy JA, Dahlberg LL, Zwi AB. The world report on violence and health. Lancet 2002; 360(9339):1083-88. PMID: 12384003. doi: 10.1016/ S0140-6736(02)11133-0.

(8.) Chen X, Thrane L, Whitbeck LB, Johnson KD, Hoyt DR. Onset of conduct disorder, use of delinquent subsistence strategies, and street victimization among homeless and runaway adolescents in the Midwest. J Interpers Violence 2007; 22(9):1156-83. PMID: 17704461. doi: 10.1177/0886260507303731.

(9.) Ferguson KM, Bender K, Thompson S, Xie B, Pollio D. Correlates of street-survival behaviors in homeless young adults in four U.S. cities. Am J Orthopsychiatry 2011; 81(3):401-9. PMID: 21729020. doi: 10.1111/j.1939-0025.2011.01108.x.

(10.) Wood E, Stoltz JA, Montaner JS, Kerr T. Evaluating methamphetamine use and risks of injection initiation among street youth: The ARYS study. Harm Reduct J 2006; 3:18. PMID: 16723029. doi: 10.1186/1477-7517-3-18.

(11.) Rhodes T. The "risk environment": A framework for understanding and reducing drug-related harm. Int J Drug Policy 2002; 13(2):85-94. doi: 10.1016/ S0955-3959(02)00007-5.

(12.) Uhlmann S, DeBeck K, Simo A, Kerr T, Montaner JSG, Wood E. Crystal methamphetamine initiation among street-involved youth. Am J Drug Alcohol Abuse 2014; 40(1):31-36. PMID: 24191637. doi: 10.3109/00952990. 2013.836531.

(13.) Marshall BDL, Fairbairn N, Li K, Wood E, Kerr T. Physical violence among a prospective cohort of injection drug users: A gender-focused approach. Drug Alcohol Depend 2008; 97(3):237-46. PMID: 18487025. doi: 10.1016/j. drugalcdep.2008.03.028.

(14.) Werb D, Bouchard M, Kerr T, Shoveller J, Qi J, Montaner J, Wood E. Drug dealing cessation among a cohort of drug users in Vancouver, Canada. Drug Alcohol Depend 2011; 118(2-3):459-63. PMID: 21664770. doi: 10.1016/j. drugalcdep.2011.05.008.

(15.) Lake S, Wood E, Dong H, Dobrer S, Montaner J, Kerr T. The impact of childhood emotional abuse on violence among people who inject drugs. Drug Alcohol Rev 2015; 34(1):4-9. PMID: 24635836. doi: 10.1111/dar.12133.

(16.) Maldonado G, Greenland S. Simulation study of confounder-selection strategies. Am J Epidemiol 1993; 138(11):923-36. PMID: 8256780.

(17.) Weist MD, Acosta OM, Youngstrom EA. Predictors of violence exposure among inner-city youth. J Clin Child Psychol 2001; 30(2):187-98. PMID: 11393919. doi: 10.1207/S15374424JCCP3002_6.

(18.) Doherty EE, Robertson JA, Green KM, Fothergill KE, Ensminger ME. A longitudinal study of substance use and violent victimization in adulthood among a cohort of urban African Americans. Addiction 2012; 107(2):339-48. PMID: 21939463. doi: 10.1111/j.1360-0443.2011.03665.x.

(19.) Werb D, Kerr T, Li K, Montaner J, Wood E. Risks surrounding drug trade involvement among street-involved youth. Am J Drug Alcohol Abuse 2008; 34(6):810-20. PMID: 19016187. doi: 10.1080/00952990802491589.

(20.) Kerr T, Small W, Johnston C, Li K, Montaner JS, Wood E. Characteristics of injection drug users who participate in drug dealing: Implications for drug policy. J Psychoactive Drugs 2008; 40(2):147-52. PMID: 18720663. doi: 10. 1080/02791072.2008.10400624.

(21.) Fast D, Small W, Krusi A, Wood E, Kerr T. "I guess my own fancy screwed me over": Transitions in drug use and the context of choice among young people entrenched in an open drug scene. BMC Public Health 2010; 10(1):126. PMID: 20222984. doi: 10.1186/1471-2458-10-126.

(22.) Johnson BD, Golub A. The potential for accurately measuring behavioral and economic dimensions of consumption, prices, and markets for illegal drugs. Drug Alcohol Depend 2007; 90(Suppl 1):S16-26. PMID: 16978801. doi: 10. 1016/j.drugalcdep.2006.08.005.

(23.) Small W, Maher L, Lawlor J, Wood E, Shannon K, Kerr T. Injection drug users' involvement in drug dealing in the downtown eastside of Vancouver: Social organization and systemic violence. Int J Drug Policy 2013; 24(5):479-87. PMID: 23664788. doi: 10.1016/j.drugpo.2013.03.006.

(24.) Decker MR, Crago A-L, Chu SKH, Sherman SG, Seshu MS, Buthelezi K, et al. Human rights violations against sex workers: Burden and effect on HIV. Lancet 2015; 385(9963):186-99. PMID: 25059943. doi: 10.1016/S0140-6736 (14)60800-X.

(25.) O'Brien K, Daffern M, Chu CM, Thomas SDM. Youth gang affiliation, violence, and criminal activities: A review of motivational, risk, and protective factors. Aggress Violent Behav 2013; 18(4):417-25. doi: 10.1016/j.avb.2013.05. 001.

(26.) Marshall BDL, DeBeck K, Simo A, Kerr T, Wood E. Gang involvement among street-involved youth in a Canadian setting: A gender-based analysis. Public Health 2015; 129(l):74-77. PMID: 25542743. doi: 10.1016/].puhe.2014. 10.017.

(27.) Richardson L, Wood E, Montaner J, Kerr T. Addiction treatment-related employment barriers: The impact of methadone maintenance. J Subst Abuse Treat 2012; 43(3):276-84. PMID: 22301085. doi: 10.1016/j.jsat.2011.12.008.

(28.) Richardson L, Sherman SG, Kerr T. Employment amongst people who use drugs: A new arena for research and intervention? Int J Drug Policy 2012; 23(1):3-5. PMID: 21996164. doi: 10.1016/j.drugpo.2011.08.004.

(29.) Werb D, Rowell G, Guyatt G, Kerr T, Montaner J, Wood E. Effect of drug law enforcement on drug market violence: A systematic review. Int J Drug Policy 2011; 22(2):87-94. PMID: 21392957. doi: 10.1016/j.drugpo.2011.02.002.

(30.) Darke S. Self-report among injecting drug users: A review. Drug Alcohol Depend 1998; 51(3):253-63; discussion 267-68. PMID: 9787998. doi: 10.1016/ S0376-8716(98)00028-3.

Received: July 13, 2015

Accepted: November 28, 2015

Kanna Hayashi, PhD, [1,2] Ben Daly-Grafstein, [1,3] Huiru Dong, MSc, [1] Evan Wood, PhD, MD, ABIM, FRCPC, [1,2] Thomas Kerr, PhD, [1,2] Kora DeBeck, PhD [1,4]

Author Affiliations

[1.] British Columbia Centre for Excellence in HIV/AIDS, St. Paul's Hospital, Vancouver, BC

[2.] Department of Medicine, University of British Columbia, St. Paul's Hospital, Vancouver, BC

[3.] Human Biology Program, University of Toronto, Toronto, ON

[4.] School of Public Policy, Simon Fraser University, Vancouver, BC

Correspondence: Kanna Hayashi, PhD, Research Scientist, BC Centre for Excellence in HIV/AIDS, St. Paul's Hospital, 608-1081 Burrard Street, Vancouver, BC V6Z 1Y6, Tel: 604-558-6680, E-mail: khayashi@cfenet.ubc.ca

Acknowledgements: The authors thank the study participants for their contribution to the research, as well as current and past researchers and staff. The study was supported by the US National Institutes of Health (U01DA038886) and the Canadian Institutes of Health Research (CIHR) (MOP-102742). This research was undertaken, in part, thanks to funding from the Canada Research Chairs program through a Tier 1 Canada Research Chair in Inner City Medicine which supports EW. KH is supported by the CIHR New Investigator Award (MSH-141971). KD is supported by a MSFHR/St. Paul's Hospital-Providence Health Care Career Scholar Award and a CIHR New Investigator Award.

Conflict of Interest: None to declare. Table 1. Baseline sample characteristics, stratified by sex (n =1152) Characteristic Total n (%) Females n (%) 1152 (100) 364 (31.6) Involvement in drug dealing and sex work * Neither 510 (44.3) 174 (47.8) Sex work but not drug dealing 42 (3.6) 20 (5.5) Drug dealing but not sex work 522 (45.3) 125 (34.3) Both 78 (6.8) 45 (12.4) Median age (IQR) 22 (20-24) 21 (19-23) Caucasian 780 (67.7) 236 (64.8) Currently in a stable relationship 313 (27.2) 146 (40.1) Homelessness * 848 (73.6) 244 (67.0) DTES residency * 325 (28.2) 105 (28.8) Daily heroin use *, ([dagger]) 117 (10.2) 47 (12.9) Daily cocaine use *, ([dagger]) 36 (3.1) 9 (2.5) Daily crack use *, ([dagger]) 180 (15.6) 54 (14.8) Daily methamphetamine use *, ([dagger]) 142 (12.3) 54 (14.8) Injection drug use * 366 (31.8) 132 (36.3) Binge drug use *, ([dagger]) 530 (46.0) 169 (46.4) Binge alcohol use * 122 (10.6) 35 (9.6) Require help injecting * 128 (11.1) 60 (16.5) Mental illness ever 766 (66.5) 269 (73.9) Incarceration * 204 (17.7) 46 (12.6) Childhood emotional abuse ([double dagger]) 549 (47.7) 216 (59.3) Experienced violence * 531 (46.1) 151 (41.5) Characteristic Males n (%) 788 (68.4) Involvement in drug dealing and sex work * Neither 336 (48.2) Sex work but not drug dealing 22 (2.8) Drug dealing but not sex work 397 (50.4) Both 33 (4.2) Median age (IQR) 22 (20-24) Caucasian 544 (69.0) Currently in a stable relationship 167 (21.2) Homelessness * 604 (76.6) DTES residency * 220 (27.9) Daily heroin use *, ([dagger]) 70 (8.9) Daily cocaine use *, ([dagger]) 27 (3.4) Daily crack use *, ([dagger]) 126 (16.0) Daily methamphetamine use *, ([dagger]) 88 (11.2) Injection drug use * 234 (29.7) Binge drug use *, ([dagger]) 361 (45.8) Binge alcohol use * 87 (11.0) Require help injecting * 68 (8.6) Mental illness ever 497 (63.1) Incarceration * 158 (20.1) Childhood emotional abuse ([double dagger]) 333 (42.3) Experienced violence * 380 (48.2) Characteristic Odds ratio p-value (95% CI) Involvement in drug dealing and sex work * Neither 1.00 Sex work but not drug dealing 1.76 (0.93-3.31) 0.081 Drug dealing but not sex work 0.61 (0.46-0.80) <0.001 Both 2.63 (1.62-4.28) <0.001 Median age (IQR) <0.001 Caucasian 0.83 (0.64-1.08) 0.157 Currently in a stable relationship 2.48 (1.89-3.25) <0.001 Homelessness * 0.61 (0.47-0.81) <0.001 DTES residency * 1.05 (0.80-1.38) 0.745 Daily heroin use *, ([dagger]) 1.52 (1.02-2.25) 0.037 Daily cocaine use *, ([dagger]) 0.72 (0.34-1.55) 0.398 Daily crack use *, ([dagger]) 0.91 (0.65-1.29) 0.606 Daily methamphetamine use *, ([dagger]) 1.39 (0.97-2.01) 0.073 Injection drug use * 1.34 (1.03-1.75) 0.027 Binge drug use *, ([dagger]) 1.02 (0.80-1.31) 0.872 Binge alcohol use * 0.86 (0.57-1.30) 0.470 Require help injecting * 2.08 (1.43-3.02) <0.001 Mental illness ever 1.66 (1.26-2.18) <0.001 Incarceration * 0.57 (0.40-0.82) 0.002 Childhood emotional abuse ([double dagger]) 2.05 (1.58-2.65) <0.001 Experienced violence * 0.76 (0.59-0.98) 0.033 Note: CI = confidence interval; IQR = interquartile range; DTES = Downtown Eastside. * Denotes activities in the previous six months. ([dagger]) Refers to any route of consumption (i.e., sniffing, snorting, smoking or injecting). ([double dagger]) Moderate to extreme childhood emotional abuse as assessed by the Childhood Trauma Questionnaire. Table 2. Univariable and multivariable GEE analyses of factors associated with experiencing violence in the past six months among street-involved youth in Vancouver, Canada, stratified by sex (n =1152) Characteristic Females (n = 364) Unadjusted OR (95% CI) Involvement in drug dealing and sex work * Involved in both vs. neither 2.15 (1.40-3.31) Involved in drug dealing but not sex work 1.84 (1.41-2.40) vs. neither Involved in sex work but not drug dealing 1.41 (0.92-2.15) vs. neither Age (per year older) 0.91 (0.87-0.94) Ethnicity (Caucasian vs. other) 1.25 (0.92-1.70) Currently in a stable relationship (yes vs. no) 0.65 (0.51-0.82) Homelessness * (yes vs. no) 2.60 (2.04-3.32) DTES residency * (yes vs. no) 1.18 (0.92-1.52) Daily heroin use *, ([dagger]) (yes vs. no) 1.34 (0.95-1.88) Daily cocaine use *, ([dagger]) (yes vs. no) 2.62 (1.14-6.02) Daily crack use *, ([dagger]) (yes vs. no) 1.10 (0.77-1.59) Daily methamphetamine use *, ([dagger]) 1.30 (0.91-1.87) (yes vs. no) Injection drug use * (yes vs. no) 1.06 (0.81-1.39) Binge drug use *, ([dagger]) (yes vs. no) 1.85 (1.50-2.29) Binge alcohol use * (yes vs. no) 1.54 (1.03-2.30) Require help injecting * (yes vs. no) 1.10 (0.80-1.51) Mental illness ever (yes vs. no) 1.24 (0.88-1.75) Incarceration* (yes vs. no) 2.33 (1.57-3.46) Childhood emotional abuse (CTQ subscales) at 1.81 (1.32-2.48) baseline (moderate to extreme [[greater than or equal to] 13] vs. none to moderate [<13]) Characteristic Females (n = 364) Adjusted OR (95% CI) Involvement in drug dealing and sex work * Involved in both vs. neither 1.29 (0.77-2.18) Involved in drug dealing but not sex work 1.43 (1.08-1.90) vs. neither Involved in sex work but not drug dealing 1.15 (0.76-1.74) vs. neither Age (per year older) 0.95 (0.91 -0.99) Ethnicity (Caucasian vs. other) Currently in a stable relationship (yes vs. no) 0.76 (0.58-1.00) Homelessness * (yes vs. no) 2.06 (1.58-2.69) DTES residency * (yes vs. no) Daily heroin use *, ([dagger]) (yes vs. no) Daily cocaine use *, ([dagger]) (yes vs. no) Daily crack use *, ([dagger]) (yes vs. no) Daily methamphetamine use *, ([dagger]) (yes vs. no) Injection drug use * (yes vs. no) Binge drug use *, ([dagger]) (yes vs. no) 1.45 (1.14-1.85) Binge alcohol use * (yes vs. no) Require help injecting * (yes vs. no) Mental illness ever (yes vs. no) Incarceration* (yes vs. no) 1.70 (1.12-2.57) Childhood emotional abuse (CTQ subscales) at 1.57 (1.13-2.20) baseline (moderate to extreme [[greater than or equal to] 13] vs. none to moderate [<13]) Characteristic Males (n = 788) Unadjusted OR (95% CI) Involvement in drug dealing and sex work * Involved in both vs. neither 2.46 (1.47-4.11) Involved in drug dealing but not sex work 1.85 (1.57-2.18) vs. neither Involved in sex work but not drug dealing 1.73 (1.03-2.91) vs. neither Age (per year older) 0.89 (0.86-0.92) Ethnicity (Caucasian vs. other) 0.92 (0.74-1.14) Currently in a stable relationship (yes vs. no) 0.83 (0.69-1.00) Homelessness * (yes vs. no) 1.74 (1.49-2.02) DTES residency * (yes vs. no) 1.13 (0.95-1.34) Daily heroin use *, ([dagger]) (yes vs. no) 0.91 (0.69-1.19) Daily cocaine use *, ([dagger]) (yes vs. no) 1.32 (0.84-2.06) Daily crack use *, ([dagger]) (yes vs. no) 0.99 (0.78-1.25) Daily methamphetamine use *, ([dagger]) 1.40 (1.12-1.75) (yes vs. no) Injection drug use * (yes vs. no) 1.05 (0.87-1.27) Binge drug use *, ([dagger]) (yes vs. no) 1.55 (1.33-1.81) Binge alcohol use * (yes vs. no) 1.59 (1.24-2.05) Require help injecting * (yes vs. no) 1.47 (1.12-1.92) Mental illness ever (yes vs. no) 1.10 (0.88-1.37) Incarceration* (yes vs. no) 1.80 (1.49-2.16) Childhood emotional abuse (CTQ subscales) at 1.51 (1.23-1.86) baseline (moderate to extreme [[greater than or equal to] 13] vs. none to moderate [<13]) Characteristic Males (n = 788) Adjusted OR (95% CI) Involvement in drug dealing and sex work * Involved in both vs. neither 1.74 (1.00-3.03) Involved in drug dealing but not sex work 1.50 (1.25-1.80) vs. neither Involved in sex work but not drug dealing 1.42 (0.81-2.48) vs. neither Age (per year older) 0.90 (0.87-0.93) Ethnicity (Caucasian vs. other) Currently in a stable relationship (yes vs. no) Homelessness * (yes vs. no) 1.37 (1.15-1.62) DTES residency * (yes vs. no) Daily heroin use *, ([dagger]) (yes vs. no) Daily cocaine use *, ([dagger]) (yes vs. no) Daily crack use *, ([dagger]) (yes vs. no) Daily methamphetamine use *, ([dagger]) 1.36 (1.07-1.74) (yes vs. no) Injection drug use * (yes vs. no) Binge drug use *, ([dagger]) (yes vs. no) 1.28 (1.08-1.51) Binge alcohol use * (yes vs. no) 1.57 (1.20-2.07) Require help injecting * (yes vs. no) Mental illness ever (yes vs. no) Incarceration* (yes vs. no) 1.70 (1.39-2.08) Childhood emotional abuse (CTQ subscales) at 1.58 (1.28-1.95) baseline (moderate to extreme [[greater than or equal to] 13] vs. none to moderate [<13]) Note: GEE = generalized estimating equations; CI = confidence interval; DTES = Downtown Eastside; CTQ = Childhood Trauma Questionnaire. * Denotes activities in the previous six months. ([dagger]) Refers to any route of consumption (i.e., sniffing, snorting, smoking or injecting).