UBC Faculty Research and Publications

Physical violence among a prospective cohort of injection drug users : a gender-focused approach. Marshall, Brandon D.; Fairbairn, Nadia; Li, Kathy; Wood, Evan; Kerr, Thomas Oct 1, 2008

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-Marshall_B_et_al_Physical_violence.pdf [ 303.19kB ]
JSON: 52383-1.0314911.json
JSON-LD: 52383-1.0314911-ld.json
RDF/XML (Pretty): 52383-1.0314911-rdf.xml
RDF/JSON: 52383-1.0314911-rdf.json
Turtle: 52383-1.0314911-turtle.txt
N-Triples: 52383-1.0314911-rdf-ntriples.txt
Original Record: 52383-1.0314911-source.json
Full Text

Full Text

Physical Violence Among a Prospective Cohort of Injection DrugUsers: A Gender-Focused ApproachBrandon D.L. Marshalla,b, Nadia Fairbairna, Kathy Lia, Evan Wooda,c, and Thomas Kerra,c,*a British Columbia Centre for Excellence in HIV/AIDS, St. Paul's Hospital, 608-1081 Burrard Street,Vancouver, BC, Canada, V6Z 1Y6b Department of Health Care and Epidemiology, University of British Columbia, 5804 Fairview Avenue,Vancouver, BC, Canada, V6T 1Z3c Department of Medicine, University of British Columbia, Vancouver Hospital & Health Sciences Centre,Room 3300-950 West 10th Avenue Vancouver, BC, Canada, V5Z 4E3AbstractAlthough dramatically heightened rates of violence have been observed among injection drug users(IDU), little is known about the gender differences associated with violence among this population.Employing a risk environment framework, we performed an analysis of the factors associated withexperiencing violence among participants enrolled in a prospective cohort study of IDU during theyears 1996-2005 using generalized estimating equations (GEE). Among 1114 individuals, 291 (66%)of females and 470 (70%) of males reported experiencing violence during the study period. Inmultivariate analyses, mental illness, frequent alcohol use, frequent crack use, homelessness,Downtown Eastside residency, and requiring help injecting were positively associated withexperiencing violence for both sexes (all p < 0.05). For females, binge drug use (AOR = 1.30) anddrug dealing (AOR = 1.42) were positively associated with violence, while younger age (AOR =1.02), frequent heroin injection (AOR = 1.24), and incarceration (AOR = 1.50) were significant formales. Women were more likely to be attacked by acquaintances, partners, and sex trade clients,while men were more likely to experience violence from strangers and the police. These findingsindicate that susceptibility to violence among IDU is structured by environmental factors such ashomelessness and drug-related factors such as frequent alcohol use and involvement in drugeconomies. Furthermore, important gender differences with respect to the predictors andcharacteristics of violent attacks do exist. These findings indicate an urgent need for the developmentof comprehensive programs and structural interventions that take a gender-focused approach toviolence among IDU.KeywordsViolence; Gender; Injection Drug Use; HIV; Canada* Corresponding author: Thomas Kerr, PhD, Research Scientist, B.C. Centre for Excellence in HIV/AIDS, University of British Columbia,St. Paul's Hospital, 608-1081 Burrard Street, Vancouver, B.C. V6Z 1Y6, CANADA, Tel: (604) 806-9116, Fax: (604) 806-9044, Email:urhi@cfenet.ubc.ca.Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customerswe are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resultingproof before it is published in its final citable form. Please note that during the production process errors may be discovered which couldaffect the content, and all legal disclaimers that apply to the journal pertain.NIH Public AccessAuthor ManuscriptDrug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.Published in final edited form as:Drug Alcohol Depend. 2008 October 1; 97(3): 237–246. doi:10.1016/j.drugalcdep.2008.03.028.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author Manuscript1. IntroductionViolence is a major cause of morbidity and mortality among injection drug users (IDU) thatoften manifests at the everyday, structural and symbolic levels (Bourgois et al., 2004; Kohli etal., 2005). Many studies have documented drastically elevated rates of physical violence amongIDU recruited from both drug treatment programs and street-based settings (Chermack andBlow, 2002; Farris and Fenaughty, 2002; Vlahov et al., 1998). For example, a study of partnerviolence among women in methadone treatment found that over three-fourths reported everexperiencing violence, while over one quarter reported physical violence in the past year (El-Bassel et al., 2000). Beyond the direct physical injury that results from violence is a range ofother health-related harms. Impairments in mental and emotional health resulting from violentencounters include depression, anxiety, suicidal ideation, posttraumatic stress disorder, moodand eating disorders, and substance dependence (Farley and Barkan, 1998; Fischbach andHerbert, 1997; Taylor and Jason, 2002).The experience of violence among IDU has also been associated with an array of HIV-relatedrisk behaviours (Braitstein et al., 2003; El-Bassel et al., 2000; Gilbert et al., 1997; Vlahov etal., 1998). The majority of these studies have focused on the past traumatic experiences andcurrent social- and individual-level factors that result in a concomitant relationship betweenviolence and HIV risk among marginalized, injection drug using women. For example, onestudy examining violence and HIV risk behaviours among female partners of male IDUobserved that almost half had been physically assaulted by their sex partners; furthermore,those who had experienced physical violence were more likely to engage in unprotected analintercourse (He et al., 1998). A study of IDU living in Vancouver found historical sexual abuseto be associated with sex trade work, sharing syringes with HIV-positive people, and otherhealth-related harms such as accidental overdose (Braitstein et al., 2003). Intimate partnerviolence among women in methadone treatment has been associated with visiting shootinggalleries, living with someone with drug or alcohol abuse problems, and exchanging sex formoney or drugs (El-Bassel et al., 2000). Other risk factors and correlates of physical violencethat have been documented among populations of drug-using women include: younger age(Vlahov et al., 1998); being separated or divorced (Gruskin et al., 2002); inconsistent condomuse (El-Bassel et al., 2005); having multiple sex partners (Wenzel et al., 2004a); unstablehousing (Wenzel et al., 2004b); marijuana use (Burke et al., 2005); and frequent alcohol use(Chermack and Blow, 2002). Childhood sexual and physical abuse is also commonly found tobe a strong and independent predictor of adult physical violence among women currently usingillicit drugs and among those enrolled in drug treatment programs (Gilbert et al., 1997).Although few studies have examined the risk factors for received violence among malesubstance users, the prevalence of physical violence among this population is also strikinglyhigh (Finlinson et al., 2003). One study examining received violence in a sample of individualsparticipating in substance abuse treatment found no gender difference in the proportion ofparticipants reporting violence from partners (61% versus 65% among men and women,respectively), while males reported higher rates of violence from non-partners (75% versus45% among men and women, respectively) (Chermack et al., 2001). Although theexpression of violence has traditionally been associated with masculine gender roles, which inturn are at least partially responsible for the linking of partner violence perpetuation and sexualrisk behaviours (Santana et al., 2006), received violence among male IDU is also common andlikely underreported. As noted by Chermack et al. (2001), studies that do not include both maleand female participants and do not examine violence outside of the partner or marital contextresult in a limited understanding of the gender-related issues that influence physical violence.Furthermore, studies that examine primarily individual-level factors (e.g., age, ethnicity,education) cannot account for the gendered structures and social relations within the drugMarshall et al. Page 2Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author Manuscripteconomy and street environment that perpetuate everyday violence experienced by IDU (Epele,2002).Over the past several decades, numerous theoretical models have been developed to guide ourunderstanding of the complex relationship between drugs, alcohol and violence (Parker andAuerhahn, 1998). One of the most commonly used frameworks is that first proposed byGoldstein, in which three categories are used to describe the factors that link drugs and violence(Fagan, 1993; Goldstein, 1985; Goldstein et al., 1989): (1) pharmacological (i.e.,pharmacological effect of substances enhance violent behaviour and individuals under theinfluence of drugs are less aware of surroundings making them vulnerable to violence); (2)economical compulsive (i.e., engagement in economically oriented violent crime to sustaindrug use behaviours); and (3) systemic (i.e., the endemicity of violent interactions within thesystem of drug distribution and use). Although this framework is often used as a theoreticalexplanation for the observed link between violence, crime and substance use, several authorshave noted that only limited empirical evidence exists to support all three components of theconceptual model (Martin and Bryant, 2001; Parker and Auerhahn, 1998). For example, in areview of illicit drug use and violence, Martin and Bryant (2001) found only limited evidenceof a pharmacological basis for an association between illicit drug use and violence, and go onto suggest that most researchers emphasize a social rather than psychopharmacological basisfor the link between cocaine use and violence. We hypothesize that “risk environment” theorymay provide a more appropriate theoretical approach to understanding the multilevel factorswhich influence exposure to physical violence among IDU. Risk environment theory positsthat the physical, social, economic, and policy environments in which drug use takes placestructures and shapes the production and re-production of HIV risk (Rhodes, 2002). Given thatthis theory has been used effectively to conceptualize the multifactorial influences that produceHIV risk among IDU (Rhodes et al., 2005), a similar approach may be useful for describingthe “risk environment” factors that structure susceptibility to physical violence among thispopulation.Although it has been shown in various settings that epidemics of HIV and violence are closelylinked (Maman et al., 2000; Quinn and Overbaugh, 2005), little is known about the experiencesof physical violence among IDU in the Downtown Eastside area of Vancouver, the city's drugscene epicenter. Given that an HIV epidemic among IDU in this area has persisted since themid 1990s, with HIV prevalence rates estimated to be greater than 30% (Kuyper et al., 2004),characterizing physical violence and its potential association with HIV risk production may beuseful for informing programs, policies and interventions targeting this population.Furthermore, there is a lack of information regarding the gender differences associated withexperiencing physical violence among individuals who inject drugs. In reponse to theseconcerns, the present study was conducted to examine physical violence, both partner violenceand other types of assault, among a prospective community-recruited cohort of male and femaleIDU. Thus, the objectives of this study were to: (1) determine the prevalence of experiencingphysical violence among this cohort of IDU; (2) examine the gender differences associatedwith risk factors for physical violence; and (3) compare the characteristics of experiencingviolence, including the perpetrator and type of attack (beating, attack with a weapon, etc.)experienced by men and women in this setting.2. MethodsThe Vancouver Injection Drug Users Study (VIDUS) is an ongoing prospective cohort studyof injection drug using individuals recruited through self-referral and street outreach fromVancouver's Downtown Eastside since May 1996. The study has been described in detailpreviously (Tyndall et al., 2003; Wood et al., 2001). Briefly, persons were eligible to participatein VIDUS if they had injected illicit drugs at least once in the previous six months, resided inMarshall et al. Page 3Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author Manuscriptthe Greater Vancouver region at time of enrolment, and provided written informed consent.At baseline and semi-annually, subjects provide blood samples and complete an interviewer-administered questionnaire. The questionnaire elicits demographic data as well as informationregarding drug use, HIV risk behaviours, sexual activity, and drug or alcohol treatment.Participants receive $20 (CDN) for each study visit. The study has been approved by theUniversity of British Columbia's Research Ethics Board.Reports of experiencing violence among men and women were identified by examiningresponses to the question, “Have you been attacked, assaulted, or suffered any kind of violencein the past six months?” Since this question has been asked consistently for over 5 years offollow-up, we were able to analyse these data longitudinally. In order to obtain more detailedinformation regarding the perpetrator and type of attacks, responses to the following questionswere also examined for those who reported received violence during the last six months: 1)“Who has attacked you?” and 2) “What type of attack was it?” These questions were used todifferentiate between partner and non-partner violence, as well as the many different forms ofviolent attack (i.e., beating, attack involving a weapon or gun, strangling, robbery, or beingphysically threatened). Participants were able to provide more than one answer to theseadditional questions. We restricted our analysis to reports of physical violence since a studyexamining the characteristics and predictors of sexual violence among this cohort of IDU hasbeen reported previously (Braitstein et al., 2003).Risk environment theory was used as the theoretical basis for the selection of primary predictorsof interest in this study. Factors such as homelessness, Downtown Eastside (DTES) residency(i.e., Vancouver's epicenter of injection drug use and HIV and HCV outbreaks), currentenrolment in any drug or alcohol treatment program, recent incarceration, requiring helpinjecting, sex trade involvement, and drug dealing were included to assess the potential impactof the social, environmental and economic conditions in which injection drug use and exposureto violence are situated. Other drug-related variables of interest included: years injecting,frequent cocaine injection, frequent heroin injection, frequent crack use, alcohol use of greaterthan four drinks per day, and binge drug use. Finally, factors that were included as potentialconfounders due to their known or a priori hypothesized relationship with both violence andone or more independent variables listed above included a range of socio-demographic andbehavioural variables: age, ethnicity (Aboriginal versus other), marital status (married versusother), mental illness, and sexual abuse. To be consistent with our previous work (Craib et al.,2003; Wood et al., 2005; Wood et al., 2001), we defined: “mental illness” as self-reporteddepression, anxiety and/or other mental health illness in the past 6 months; “frequent use” tobe daily injection of cocaine or heroin or daily smoking of crack; “binge drug use” to be self-reported periods when drugs were injected more frequently than usual; “sexual abuse” to beever having been forced to have sex against one's will; and “incarceration” as being in detention,jail or prison overnight or longer. All drug use variables along with homelessness, DTESresidency, recent incarceration, and sex trade involvement were treated as time-updatedcovariates that refer to activities or situations occurring during the past six months.Initially, we examined bivariate associations between the risk environment, drug-related, andpotentially confounding variables and reported violence using generalized estimatingequations (GEE). We used GEE for binary outcomes with logit link for the analysis ofcorrelated data since the factors potentially associated with violence during follow-up wereserial (time-dependent) measures. GEE models account for the correlation between repeatedmeasures for each subject, and as such data from every participant follow-up visit wasconsidered in the analysis. Therefore, this technique permitted the determination of whichfactors were associated with received violence prior to the six-month follow-up interviewthroughout the 60-month observation period. The GEE method is commonly used for studiesin which a repeated measure binary dependent variable is analysed longitudinally, and has beenMarshall et al. Page 4Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author Manuscriptdescribed in our work previously (Kerr et al., 2005a). In order to construct gender-specificexplanatory models for experiencing violence and to adjust for potential confounding, we alsofit multivariate logistic GEE models adjusting for all variables that were found to besignificantly associated (p < 0.05) with violence in bivariate analyses. To compare thecharacteristics of physical violence by sex, Chi-square tests were conducted. Fisher's exact testwas also used to compute p-values when one or more of the observations was less than or equalto five. The Wilcoxon rank sum test was used to compare between men and women the mediannumber of follow-up visits and the median number of violent experiences reported over thestudy period. All statistical analyses were performed using SAS software version 8.0 (SAS,Cary, NC), and all reported p-values are two-sided.3. ResultsA total of 1114 participants completed at least one follow-up interview during the period fromDecember 2000 to December 2005 and were eligible for this analysis. Out of 10 interviewperiods, the median number of follow-up visits was 8 (interquartile range [IQR]: 4 – 10), withno significant difference in follow-up between men and women (p = 0.99). The median agewas 36.8 (IQR: 28.3 – 44.2), 444 (40%) were women, and 346 (31%) were of Aboriginalancestry. At baseline, we observed several gender differences with respect to the primaryvariables of interest. Men were more likely to report recent incarceration (22% vs. 14%, p =0.003), while women were more likely to report recent sex trade involvement (31% vs. 4%,p < 0.001), drug dealing (34% vs. 21%, p < 0.001), enrolment in drug or alcohol treatment(56% vs. 49%, p = 0.047), and requiring help injecting (23% vs. 13%, p < 0.001). At baseline,98 (22%) females and 138 (21%) males reported having suffered physical violence in the lastsix months (p = 0.56). Over the follow-up period, a further 193 (43%) females and 332 (50%)males reported experiencing physical violence. Therefore, over the entire study period, a totalof 291 (66%) females and 470 (70%) males reported experiencing violence at least once (p =0.11). Of those who reported at least once incident of violence, the median number of reportsof violence over the study period was 2 (IQR: 1 – 3) for both men and women.The bivariate GEE analyses of associations between variables of interest and self-reportedviolence are shown in Tables 1 and 2 for females and males, respectively. For females, thefollowing variables were positively associated with experiencing violence: younger age (oddsratio [OR] = 1.03, 95% confidence interval [CI]: 1.01 – 1.05); mental illness (OR = 1.64, 95%CI: 1.34 – 2.00); sexual abuse (OR = 1.36, 95% CI: 1.10 – 1.69); alcohol use (OR = 1.51, 95%CI: 1.25 – 1.83); frequent heroin injection (OR = 1.64, 95% CI: 1.29 – 2.10); frequent cocaineinjection (OR = 1.28, 95% CI: 1.02 – 1.62); frequent crack smoking (OR = 1.85, 95% CI: 1.48– 2.33); binge drug use (OR = 1.60, 95% CI: 1.28 – 2.02); homelessness (OR = 1.93, 95% CI:1.43 – 2.62); Downtown Eastside (DTES) residency (OR = 1.96, 95% CI: 1.56 – 2.46);incarceration (OR = 1.68, 95%CI: 1.28 – 2.21); sex trade involvement (OR = 1.59, 95% CI:1.24 – 2.04); drug dealing (OR = 2.04, 95%CI: 1.62 – 2.58); and requiring help injecting (OR= 1.91, 95% CI: 1.47 – 2.49). Being married or common-law (OR = 0.76, 95% CI: 0.61 – 0.94)and current enrolment in a drug or alcohol treatment program (OR = 0.78, 95% CI: 0.63 – 0.99)were negatively associated with experiencing violence. Among women in the study, Aboriginalethnicity and years injecting were not statistically associated with experiencing violence inbivariate analyses.For males, the following variables were positively associated with experiencing violence:younger age (OR = 1.03, 95% CI: 1.02 – 1.04); mental illness (OR = 1.56, 95% CI: 1.33 –1.83); alcohol use (OR = 1.39, 95% CI: 1.17 – 1.64); frequent heroin injection (OR = 1.57,95% CI: 1.30 – 1.88); frequent cocaine injection (OR = 1.25, 95% CI: 1.04 – 1.51); frequentcrack smoking (OR = 1.65, 95% CI: 1.39 – 1.96); homelessness (OR = 1.86, 95% CI: 1.48 –2.35); DTES residency (OR = 1.59, 95% CI: 1.33 – 1.91); incarceration (OR = 1.77, 95% CI:Marshall et al. Page 5Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author Manuscript1.45 – 2.17); drug dealing (OR = 1.57, 95% CI: 1.29 – 1.91); and requiring help injecting (OR= 1.67, 95% CI: 1.36 – 2.06). Being married or common-law (OR = 0.79, 95% CI: 0.64 – 0.97)and years injecting (OR = 0.99, 95%CI: 0.97 – 1.00) were negatively associated with reportedviolence. Among men in the study, Aboriginal ethnicity, sex trade work, sexual abuse, bingedrug use, and enrolment in any drug or alcohol treatment program were not statisticallyassociated with experiencing violence in bivariate analyses.The results of the multivariate GEE analyses are also shown in Tables 1 and 2 for females andmales, respectively. For both sexes, mental illness (adjusted odds ratio for females [AORf] =1.69, 95%CIf: 1.35 – 2.10; adjusted odds ratio for males [AORm] = 1.54, 95%CIm: 1.29 – 1.82),alcohol use (AORf = 1.49, 95% CIf: 1.22 – 1.82; AORm = 1.42, 95% CIm: 1.20 – 1.69), frequentcrack smoking (AORf = 1.34, 95% CIf: 1.04 – 1.73; AORm = 1.38, 95% CIm: 1.15 – 1.65),homelessness (AORf = 1.43, 95% CIf: 1.03 – 1.98; AORm = 1.53, 95% CIm: 1.20 –1.94), DTESresidency (AORf = 1.48, 95% CIf: 1.18 – 1.87; AORm = 1.38, 95% CIm: 1.15 – 1.66), andrequiring help injecting (AORf = 1.51, 95%CIf: 1.15 – 1.99; AORm = 1.51, 95% CIm: 1.22 –1.87) were positively and independently associated with received violence. Predictors that wereunique to females included binge drug use (AORf = 1.30, 95% CIf: 1.03 – 1.65) and drugdealing (AORf = 1.42, 95% CIf: 1.11 – 1.81). Predictors that were unique to males included:younger age (AORm = 1.02, 95%CIm: 1.00 – 1.03); frequent heroin injection (AORm = 1.24,95% CIm: 1.02 – 1.51); and recent incarceration (AORm = 1.50, 95% CIm: 1.22 – 1.84).Acquaintances and strangers were the most commonly reported perpetrator of physicalviolence. As shown in Table 3, women were more likely than men to report being attacked byacquaintances (43.5% versus 37.0% of reports, p = 0.017), partners (5.2% vs. 2.6%, p = 0.010),and sex trade clients and workers (4.6% vs. 0.9%, p < 0.001), while men were more likely toreport being attacked by strangers (40.7% vs. 31.7%, p < 0.001) and by the police (8.7% vs.3.8%, p < 0.001). The most commonly reported type of attack for both genders was a beating(65.1% of reports from women and 61.1% of reports from men). Women were more likely toreport being strangled (2.0% vs. 0.5%, p = 0.014) and physically threatened (2.4% vs. 0.4%,p = 0.002), while men were significantly more likely to report being attacked with weapons(22.6% versus 10.8%, p < 0.001).4. DiscussionIn the present study, we observed a very high prevalence of received physical violence amongboth male and female IDU. Similar proportions of men and women reported experiencing arecent violent attack at baseline (21% vs. 22%) and over the follow-up period (50% vs. 43%).These results are consistent with previous studies that have observed comparable overall ratesof received violence among drug-using men and women (Chermack et al., 2001; Finlinson etal., 2003). In longitudinal multivariate analyses examining the risk factors associated withreceived violence, many similarities between sexes were observed. Mental illness, alcohol use,frequent crack smoking, homelessness, DTES residency, and requiring help injecting werepositively and independently associated with experiencing violence for both sexes. Althoughmany of the risk factors examined did not vary considerably by sex, several significantdifferences were observed. Factors that were positively and independently associated withviolence for females included binge drug use and drug dealing, while risk factors unique tomales included frequent heroin injection and recent incarceration. Several important genderdifferences were also observed when the perpetrator and nature of the violent experience wereexamined. Men were more likely to report being attacked by strangers and the police, while agreater proportion of women reported violence from acquaintances, partners, and individualsinvolved in the sex trade. Furthermore, men were significantly more likely to report beingattacked with weapons, while women were more likely to report being strangled or physicallythreatened.Marshall et al. Page 6Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author ManuscriptConsistent with other studies examining illicit drug use and violence (Martin and Bryant,2001; Parker and Auerhahn, 1998), the frequent consumption of cocaine and heroin (i.e., thedirect pharmacology of these substances) were poor predictors of experiencing violence amongIDU. However, other drug-related factors more closely linked to the social and environmentalcharacteristics of drug economies and substance use (e.g., drug dealing, requiring helpinjecting) were independently associated with experiencing violence and thus may play a largerrole in the perpetuation of violence among IDU. An exception is the frequent smoking of crack,which was positively and independently associated with violence among both males andfemales. The association between daily women's crack use and violence has been noted inprevious studies (Brewer et al., 2005; Vlahov et al., 1998; Wechsberg et al., 2003). Our resultsalso suggest that frequent crack use is an equally strong predictor of experiencing violenceamong male IDU. Frequent alcohol use was also one of the strongest predictors of experiencingphysical violence, with an adjusted odds ratio higher than that for frequent heroin, cocaine, orcrack consumption. This finding is consistent with a previous study of HIV positive drug-usingindividuals that observed an association between recent received violence and alcohol useseverity (Liebschutz et al., 2005).Although the Goldstein tripartite model has been used extensively to understand the interactionof drugs and the expression of violence, we propose that the above results are better explainedusing a “risk environment” approach that accounts for the social, environmental and structuralinfluences on violence and illicit drug use. Many factors that were associated with violence inour study appear to be closely related to the characteristics of the neighbourhood, physicalenvironment, and social and economic context in which IDU are situated. For example,Downtown Eastside (DTES) residency was strongly and independently associated withreceived violence for both sexes. The DTES in Vancouver is one of the most economicallydisadvantaged neighbourhoods in Canada, and as such has the lowest per capita income of anyurban region in the country (Statistics Canada, 1996). Homelessness, a common feature of thisand many other impoverished neighbourhoods, has been recognized as a major determinant ofpoor health among IDU, including an elevated risk of HIV infection (Fisher et al., 1995; Rhodeset al., 2005; Zolopa et al., 1994). The observed association between homelessness and violencemay be attributed to a variety of factors. First, individuals who are homeless are moresusceptible to violence due to a lack of protective shelter. Second, homeless individuals mayhave to engage in income generation activities that carry increased risk for violence (e.g., pettycrime, sex work) (Erickson, 2001; Johnson et al., 1985). Finally, homeless individuals are morelikely to participate in the street-based drug economy (Fisher et al., 1995; Wechsberg et al.,2003), a culture which has been associated with the normalization of extreme levels of violence(Bourgois et al., 2004). Further research must be conducted to elucidate the multi-level factorsthat perpetuate both homelessness and violence within disadvantaged and drug usingpopulations.Risk environment theory also accounts for violence attributed to gender constructs whichoperate within unregulated drug-based cultures and economies (Rhodes et al., 2005). Inelucidating the complex intersections of gender, HIV risk, and violence in the lives of injectiondrug using women, Epele (2002) has argued that gender inequality promotes HIV vulnerabilityas a consequence of multidimensional violence. The author argues that gender inequalityresults in both subordinated positions of women within the street-involved drug economy andprecarious subsistence strategies such as sex trade work (Epele, 2002). In our study, weobserved that drug dealing and binge drug use were positively associated with violence forwomen but not for men. These findings can be explained in terms of the social, economic, andstructural gender inequalities described by Epele. For example, women are known to besystematically excluded from higher-level roles in the hierarchy of drug dealing due to a male-centered street ideology that enforces the perception that women should be denied power overresources derived from the illicit drug trade (Maher and Daly, 1996). Recent evidence hasMarshall et al. Page 7Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author Manuscriptsuggested that this relegation of women to lower-level roles puts them at increased risk ofviolence and even death (Miller et al., 2007). We also postulate that the physical harmsassociated with sex trade work and its association with binge drug use may explain theconnection between bingeing and exposure to violence among women in the cohort. Miller etal. (2006) have suggested that women on a binge may be more likely to encounter violencewith “bad dates” due to lower selectively of clients in an attempt to generate income quickly.Further research should be conducted to examine the multiple gender inequalities thatperpetuate violence and HIV risk among injection drug using women who trade sex or dealdrugs.The finding that IDU who experience violence are more likely to require assistance injectingis of particular concern because violence may serve to compound the many other well-knownharms associated with assisted injection, including risk for HIV infection (O'Connell et al.,2005; Wood et al., 2003). Recently, requiring assistance injecting has also been associated withinjection in public (McKnight et al., 2007). Individuals injecting in public may be moredistressed and agitated due to increased pressure to inject quickly in order to avoid police andother threats (Darke et al., 2001; Wood et al., 2004). Since IDU who inject in public aresusceptible to and feel threatened by street violence from police and other street predators(Small et al., 2007), our results provide further evidence that heavy enforcement in areas wherepublic and assisted injection is known to occur may only serve to increase the likelihood ofviolence and assault against IDU. Additionally, requiring assistance injecting often involvescompensation to the individual who assists with the injection (i.e., a “hit doctor”), and disputesover what constitutes appropriate compensation may explain why assisted injection isassociated with experiencing violence (Fairbairn et al., 2006).The finding that almost 4% of violent attacks reported by women and 9% of attacks reportedby men were attributed to interactions with police indicates that police presence is acontributing factor to violent experiences among IDU in this setting. This finding is consistentwith previous reports of police-related violence in the DTES (Human Rights Watch, 2003;Pivot Legal Society, 2004). Beyond the direct result of violent interactions with police, thepresence of heavy enforcement within unregulated drug market environments may indirectlyimpact the incidence of violence through a number of mechanisms (Erickson, 2001; Kerr etal., 2005c). For example, although increased policing (“crackdowns”) has been a commontactic aimed at upsetting illegal drug markets and restoring public order, the success of suchinterventions are often time-limited and can be completely offset by displacement of drugmarkets into non-public locations or neighbouring areas (Kerr et al., 2005c). As a result of thedisplacement of established relationships within the drug economy, drug market enforcementhas also been associated with increased violence and volatility among drug users and dealers(Brownstein et al., 2000; Goldstein, 1989; Maher and Dixon, 1999; Small et al., 2006). Theseresults provide further support to the notion that heavy drug market enforcement interacts withand transforms various practices and social dynamics within the broader risk environment ofIDU (Rhodes et al., 2005), and thereby constitutes a potential source of violence that adverselyaffects the health of this population.These findings have significant implications for education and prevention initiatives, harmreduction programs, and the development of policies and laws. Principally, our results suggestthat structural and other non-individual level interventions may be effective in reducing therisk of violence among both male and female IDU. For example, an important way to diminishthe risk of violence associated with the consumption of illicit drugs within dangerousenvironments such as the street is the provision of monitored enclosed spaces where individualsmay use pre-obtained illicit drugs under the supervision of health care providers, (e.g.,supervised injection facilities). Within such facilities, the risk of physical violence isdiminished because injections take place in a supervised environment free of threats posed byMarshall et al. Page 8Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author Manuscriptpolice and street predators. Since our results also indicate that frequent crack use is associatedwith experiencing violence, the implementation of safer smoking facilities, where individualsmay smoke crack while supervised by health care providers, is an additional micro-environmental intervention that may reduce the risk of experiencing violence (Shannon et al.,2006). Furthermore, since requiring assistance injecting was associated with experiencingviolence for both sexes, permitting assisted injection at these facilities may also provide anopportunity for IDU to learn to self-inject, or to receive injections in an environment wherethe harms associated with this dangerous practice are reduced (Kerr et al., 2005b; Wood et al.,2003). In light of the observed association between violence and homelessness, increasedprogramming to provide supportive housing and emergency shelters are urgently needed todiminish the risk of violence among IDU who lack affordable housing. Since severity(frequency and duration) of homelessness has in itself been associated with experiencingviolence among homeless women (Wenzel et al., 2001), access to safe and stable shelter islikely a necessary component of any intervention seeking to address this issue. Furthermore,since involvement in the drug trade and street economy appears to carry greater risks forwomen, gender-focused programming that addresses the economic and educationaldisadvantage of injection drug using women should be a priority. Given that improvedeconomic status and decreased financial dependence on male partners have been associatedwith reduced HIV-related risks among women elsewhere (Blankenship et al., 2006), we suggestthat income support and vocational programming would carry other benefits in terms ofreducing women's exposure to violence. And finally, given the evidence indicating thatescalated prohibition and police enforcement increases the level of violence between policeand drug users and between individuals within the street-based drug market (Erickson, 2001;Kerr et al., 2005c), continued reliance on such measures to minimize drug-related violence isnot recommended. Health-focused policies, interventions, and evidence-based harm reductionprogramming that addresses the social, structural, and environmental inequalities so persuasivein the lives of injection drug users may ultimately be more successful in reducing the highlevels of physical violence observed in this setting.There are several limitations to the present study. First, we have restricted our analysis toexperiencing violence among IDU and have therefore not considered violence perpetuated bywomen or men. Second, as with many other prospective cohort studies of IDU, VIDUS is nota random sample, and as such these findings may not generalize to other IDU populations.Third, this study relied on self-reported information and is hence susceptible to sociallydesirable reporting. In the present study, this may have led to an underestimation of theprevalence of violence or other stigmatized activities among either male or female IDU.Previous studies have indicated that in certain settings violence against women and/or violenceagainst men can be underreported (Koss, 1992; Watkins and Bentovim, 1992; Watts andZimmerman, 2002).In summary, the results of our study suggest that violence is a common experience among maleand female IDU in Vancouver. Susceptibility to violence among this population is moderatedby gender, and is driven by a combination of factors. Although mental illness, frequent alcoholuse and crack use were associated with violence, structural and environmental factors such ashomelessness, DTES residency and recent incarceration were also strong predictors ofexperiencing physical violence in this population. We have argued that such results supportthe adoption of a risk environment approach, which posits that the interaction of environmentalfactors with individual-level determinants structures susceptibility to physical violence amongIDU. Furthermore, we recommend that frameworks which seek to describe violence amongIDU should be updated to reflect the gendered influences on the social and structural productionof violence. These results also indicate that IDU who experience physical assault are amongthe most marginalized, and as such violence should be a primary area of focus for preventionand support services. Such interventions include safe and stable housing options, educationalMarshall et al. Page 9Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author Manuscriptand economic programming, reforms to drug enforcement strategies, and the expansion ofharm reduction programs such as supervised injecting facilities.ReferencesBlankenship KM, Friedman SR, Dworkin S, Mantell JE. Structural interventions: concepts, challengesand opportunities for research. J Urban Health 2006;83:59–72. [PubMed: 16736355]Bourgois P, Prince B, Moss A. The everyday violence of hepatitis C among young women who injectdrugs in San Francisco. Hum Org 2004;63:253–264.Braitstein P, Li K, Tyndall M, Spittal P, O'Shaughnessy MV, Schilder A, Johnston C, Hogg RS, SchechterMT. Sexual violence among a cohort of injection drug users. Soc Sci Med 2003;57:561–569. [PubMed:12791497]Brewer DD, Fleming CB, Haggerty KP. Past and future research on the crack cocaine-domestic violencelink. Am J Public Health 2005;95:1493–1494. [PubMed: 16051926]Brownstein HH, Crimmins SM, Spunt BJ. A conceptual framework for operationalizing the relationshipbetween violence and drug market stability. Contemp Drug Problems 2000;27:867–890.Burke JG, Thieman LK, Gielen AC, O'Campo P, McDonnell KA. Intimate partner violence, substanceuse, and HIV among low-income women: Taking a closer look. Violence Against Wom 2005;11:1140–1161.Chermack ST, Blow FC. Violence among individuals in substance abuse treatment: The role of alcoholand cocaine consumption. Drug Alcohol Depend 2002;66:29–37. [PubMed: 11850133]Chermack ST, Walton MA, Fuller BE, Blow FC. Correlates of expressed and received violence acrossrelationship types among men and women substance abusers. Psychol Addictive Behav 2001;15:140–151.Craib KJ, Spittal PM, Wood E, Laliberte N, Hogg RS, Li K, Heath K, Tyndall MW, O'Shaughnessy MV,Schechter MT. Risk factors for elevated HIV incidence among Aboriginal injection drug users inVancouver. CMAJ 2003;168:19–24. [PubMed: 12515780]Darke S, Kaye S, Ross J. Geographical injecting locations among injecting drug users in Sydney,Australia. Addiction 2001;96:241–246. [PubMed: 11182868]El-Bassel N, Gilbert L, Schilling R, Wada T. Drug Abuse and Partner Violence among women inmethadone treatment. J Family Viol 2000;15:209–228.El-Bassel N, Gilbert L, Wu E, Go H, Hill J. HIV and intimate partner violence among methadone-maintained women in New York City. Soc Sci Med 2005;61:171–183. [PubMed: 15847970]Epele ME. Gender, violence and HIV: Women's survival in the streets. Cult Med Psychiatry 2002;26:33–54. [PubMed: 12088097]Erickson, PG. Drugs, violence, and public health: What does the harm reduction approach have to offer?.The Fraser Institute; Vancouver: 2001.Fagan J. Interactions among drugs, alcohol, and violence. Health Aff 1993;12:65–79.Farley M, Barkan H. Prostitution, violence, and posttraumatic stress disorder. Women Health1998;27:37–49. [PubMed: 9698636]Farris CA, Fenaughty AM. Social isolation and domestic violence among female drug users. Am J DrugAlcohol Abuse 2002;28:339–351. [PubMed: 12014819]Finlinson HA, Oliver-Vélez D, Deren S, Cant JGH, Colón HM, Robles RR, Andía JF. Puerto Rican drugusers' experiences of physical and sexual abuse: Comparisons based on sexual identities. ViolenceAgainst Wom 2003;9:839–858.Fischbach RL, Herbert B. Domestic violence and mental health: Correlates and conundrums within andacross cultures. Soc Sci Med 1997;45:1161–1176. [PubMed: 9381230]Fisher B, Hovell M, Hofstetter CR, Hough R. Risks associated with long-term homelessness amongwomen: Battery, rape, and HIV infection. Int J Health Serv 1995;25:351–369. [PubMed: 7622323]Fairbairn N, Wood E, Small W, Stoltz JA, Li K, Kerr T. Risk profile of individuals who provide assistancewith illicit drug injections. Drug Alcohol Depend 2006;82:41–46. [PubMed: 16150556]Marshall et al. Page 10Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author ManuscriptGilbert L, El-Bassel N, Schilling RF, Friedman E. Childhood abuse as a risk for partner abuse amongwomen in methadone maintenance. Am J Drug Alcohol Abuse 1997;23:581–595. [PubMed:9366975]Goldstein P. The drugs/violence nexus: A tripartite conceptual framework. J Drug Issues 1985;39:143–174.Goldstein, P. Drugs and violent crime. In: Weiner, NA.; Wolfgana, ME., editors. Pathways to CriminalViolence. Sage Publications; Beverly Hills: 1989. p. 16-48.Goldstein P, Brownstein H, Ryan P, Bellucci P. Crack and homicide in New York City 1988: Aconceptually based event analysis. Contemp Drug Probl 1989;16:651–687.Gruskin L, Gange SJ, Celentano DD, Schuman P, Moore JS, Zierler S, Valhov D. Incidence of violenceagainst HIV-infected and uninfected women: Findings from the HIV Epidemiology Research (HER)study. J Urban Health 2002;79:512–524. [PubMed: 12468671]He H, McCoy HV, Stevens SJ, Stark MJ. Violence and HIV sexual risk behaviors among female sexpartners of male drug users. Women Health 1998;27:161–175. [PubMed: 9640640]Human Rights Watch. Abusing the user: Police misconduct, harm reduction and HIV/AIDS inVancouver. 15. Human Rights Watch; New York: 2003.Johnson, B.; Goldstein, P.; Preble, E.; Schneidler, J.; Lipton, D.; Spunt, B.; Miller, T. Taking Care ofBusiness: The Economics of Crime by Heroin Abusers. Lexington Books; Lexington, Kentucky:1985.Kerr T, Marsh D, Li K, Montaner J, Wood E. Factors associated with methadone maintenance therapyuse among a cohort of polysubstance using injection drug users in Vancouver. Drug Acohol Depend2005a;80:329–335.Kerr T, Oleson M, Tyndall MW, Montaner J, Wood E. A description of a peer-run supervised injectionsite for injection drug users. J Urban Health 2005b;82:267–275. [PubMed: 15872193]Kerr T, Small W, Wood E. The public health and social impacts of drug market enforcement: A reviewof the evidence. Int J Drug Policy 2005c;16:210–220.Kohli R, Lo Y, Howard AA, Buono D, Floris-Moore M, Klein RS, Schoenbaum EE. Mortality in anurban cohort of HIV-infected and at-risk drug users in the era of highly active antiretroviral therapy.Clin Infect Dis 2005;41:864–872. [PubMed: 16107987]Koss MP. The under-detection of rape: Methodological choices influence incidence estimates. J SocialIssues 1992;48:61–75.Kuyper LM, Hogg RS, Montaner JSG, Schechter MT, Wood E. The cost of inaction on HIV transmissionamong injection drug users and the potential for effective interventions. J Urban Health 2004;81:655–660. [PubMed: 15466846]Liebschutz JM, Geier JL, Horton NJ, Chuang CH, Samet JH. Physical and sexual violence and healthcare utilization in HIV-infected persons with alcohol problems. AIDS Care 2005;17:566–578.[PubMed: 16036243]Maher L, Daly K. Women in the street-level drug economy: Continuity or change? Criminol1996;34:465–491.Maher L, Dixon D. Policing and public health: Law enforcement and harm minimization in a street-leveldrug market. Brit J Criminol 1999;39:488–512.Martin SE, Bryant K. Gender differences in the association of alcohol intoxication and illicit drug abuseamong persons arrested for violent and property offenses. J Substance Abuse 2001;13:563–581.Maman S, Campbell J, Sweat MD, Gielen AC. The intersections of HIV and violence: Directions forfuture research and interventions. Soc Sci Med 2000;50:459–478. [PubMed: 10641800]McKnight I, Maas B, Wood E, Tyndall MW, Small W, Lai C, Montaner JSG, Kerr T. Factors associatedwith public injecting among users of Vancouver's supervised injection facility. Am J Drug AlcoholAbuse 2007;33:319–325. [PubMed: 17497555]Miller CL, Kerr T, Frankish JC, Spittal PM, Li K, Schechter MT, Wood E. Binge drug use independentlypredicts HIV seroconversion among injection drug users: Implications for public health strategies.Subst Use Misuse 2006;41:199–210. [PubMed: 16393742]Miller CL, Kerr T, Strathdee SA, Li K, Wood E. Factors associated with premature mortality amongyoung injection drug users in Vancouver. Harm Reduct J 2007;4:1. [PubMed: 17201933]Marshall et al. Page 11Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author ManuscriptO'Connell JM, Kerr T, Li K, Tyndall MW, Hogg RS, Montaner JSG, Wood E. Requiring help injectingindependently predicts incident HIV infection among injection drug users. J Acquir Immune DeficSynd 2005;40:83–88.Parker RN, Auerhahn K. Alcohol, drugs, and violence. Annu Rev Sociol 1998;24:291–311.Pivot Legal Society. Voices for dignity: A call to end the harms caused by Canada's sex trade laws. PivotLegal Society; Vancouver: 2004.Quinn TC, Overbaugh J. HIV/AIDS in women: An expanding epidemic. Science 2005;308:1582–1583.[PubMed: 15947174]Rhodes T. The ‘risk environment’: A framework for understanding and reducing drug-related harm. IntJ Drug Policy 2002;13:85–94.Rhodes T, Singer M, Bourgois P, Friedman SR, Strathdee SA. The social structural production of HIVrisk among injecting drug users. Soc Sci Med 2005;61:1026–1044. [PubMed: 15955404]Santana MC, Raj A, Decker MR, La Marche A, Silverman JG. Masculine gender roles associated withincreased sexual risk and intimate partner violence perpetration among young adult men. J UrbanHealth 2006;83:575–585. [PubMed: 16845496]Shannon K, Ishida T, Morgan R, Bear A, Oleson M, Kerr T, Tyndall MW. Potential community andpublic health impacts of medically supervised safer smoking facilities for crack cocaine users. HarmReduct J 2006;3:1. [PubMed: 16403229]Small W, Kerr T, Charette J, Schechter MT, Spittal P. Impacts of intensified police activity on injectiondrug users: Evidence from an ethnographic investigation. Int J Drug Policy 2006;17:85–95.Small W, Rhodes T, Wood E, Kerr T. Public injection settings in Vancouver: Physical environment,social context and risk. Int J Drug Policy 2007;18:27–36. [PubMed: 17689341]Statistics Canada. Population Census of Canada. Government of Canada; Ottawa: 1996.Taylor RR, Jason LA. Chronic fatigue, abuse-related traumatization, and psychiatric disorders in acommunity-based sample. Soc Sci Med 2002;55:247–256. [PubMed: 12144139]Tyndall MW, Currie S, Spittal P, Li K, Wood E, O'Shaughnessy MV, Schechter MT. Intensive injectioncocaine use as the primary risk factor in the Vancouver HIV-1 epidemic. AIDS 2003;17:887–893.[PubMed: 12660536]Vlahov D, Wientge D, Moore J, Flynn C, Schuman P, Schoenbaum EE, Zierler S. Violence among womenwith or at risk for HIV infection. AIDS Behav 1998;2:53–60.Watkins B, Bentovim A. The sexual abuse of male children and adolescents: A review of current research.J Child Psychol Psychiatry 1992;33:197–248. [PubMed: 1737828]Watts C, Zimmerman C. Violence against women: Global scope and magnitude. Lancet 2002;359:1232–1237. [PubMed: 11955557]Wechsberg WM, Lam WK, Zule W, Hall G, Middlesteadt R, Edwards J. Violence, homelessness, andHIV risk among crack-using African-American women. Subst Use Misuse 2003;38:669–700.[PubMed: 12747401]Wenzel SL, Leake BD, Gelberg L. Risk factors for major violence among homeless women. JInterpersonal Viol 2001;16:739–752.Wenzel SL, Tucker JS, Elliott MN, Hambarsoomians K, Perlman J, Becker K, Kollross C, Golinelli D.Prevalence and co-occurrence of violence, substance use and disorder, and HIV risk behavior: Acomparison of sheltered and low-income housed women in Los Angeles County. Prev Med 2004b;39:617–624. [PubMed: 15313103]Wenzel SL, Tucker JS, Elliott MN, Marshall GN, Williamson SL. Physical violence against impoverishedwomen: A longitudinal analysis of risk and protective factors. Women Health Iss 2004a;14:144–154.Wood E, Li K, Small W, Montaner JSG, Schechter MT, Kerr T. Recent incarceration independentlyassociated with syringe sharing by injection drug users. Pub Health Rep 2005;120:150–156.[PubMed: 15842116]Wood E, Spittal PM, Kerr T, Small W, Tyndall MW, O'Shaughnessy MV, Schechter MT. Requiring helpinjecting as a risk factor for HIV infection in the Vancouver epidemic: Implications for HIVprevention. Can J Pub Health 2003;94:355–359. [PubMed: 14577743]Marshall et al. Page 12Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author ManuscriptWood E, Spittal PM, Small W, Kerr T, Li K, Hogg RS, Tyndall MW, Montaner JSG, Schechter MT.Displacement of Canada's largest public illicit drug market in response to a police crackdown. CMAJ2004;170:1551–1556. [PubMed: 15136548]Wood E, Tyndall MW, Spittal PM, Li K, Kerr T, Hogg RS, Montaner JSG, O'Shaughnessy MV, SchechterMT. Unsafe injection practices in a cohort of injection drug users in Vancouver: Could safer injectingrooms help? CMAJ 2001;165:405–410. [PubMed: 11531048]Zolopa AR, Hahn JA, Gorter R, Miranda J, Wlodarczyk D, Peterson J, Pilote L, Moss AR. HIV andtuberculosis infection in San Francisco's homeless adults. Prevalence and risk factors in arepresentative sample. JAMA 1994;272:455–461. [PubMed: 8040981]Marshall et al. Page 13Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author ManuscriptMarshall et al. Page 14Table 1Bivariate and multivariate GEEa analysis of factors associated with received violence for females during follow-up (n = 444)Characteristic Unadjusted Odds Ratio(95% CIb) p - valueAdjusted Odds Ratio(95% CIb) p - valueAge (per year younger) 1.03 (1.01 – 1.05) 0.001 1.02 (1.00 – 1.03) 0.084Aboriginal ethnicity (yes vs. no) 0.75 (0.56 – 1.01) 0.059Married (yes vs. no) 0.76 (0.61 – 0.94) 0.012 0.82 (0.66 – 1.03) 0.083Mental illness† (yes vs. no) 1.64 (1.34 – 2.00) < 0.001 1.69 (1.35 – 2.10) < 0.001Sexual abuse‡ (yes vs. no) 1.36 (1.10 – 1.69) 0.006 1.25 (0.99 – 1.58) 0.057Years injecting (per year older) 0.99 (0.97 – 1.01) 0.271Alcohol use† (yes vs. no) 1.51 (1.25 – 1.83) < 0.001 1.49 (1.22 – 1.82) < 0.001Frequent heroin injection† (> daily vs. < daily) 1.64 (1.29 – 2.10) < 0.001 1.03 (0.78 – 1.36) 0.842Frequent cocaine injection† (> daily vs. < daily) 1.28 (1.02 – 1.62) 0.037 0.91 (0.71 – 1.18) 0.483Frequent crack smoking† (> daily vs. < daily) 1.85 (1.48 – 2.33) < 0.001 1.34 (1.04 – 1.73) 0.023Binge drug use† (yes vs. no) 1.60 (1.28 – 2.02) < 0.001 1.30 (1.03 – 1.65) 0.027Homelessness†† (yes vs. no) 1.93 (1.43 – 2.62) < 0.001 1.43 (1.03 – 1.98) 0.034DTESc residency†† (yes vs. no) 1.96 (1.56 – 2.46) < 0.001 1.48 (1.18 – 1.87) 0.001Incarceration† (yes vs. no) 1.68 (1.28 – 2.21) < 0.001 1.28 (0.96 – 1.72) 0.097Sex trade involvement† (yes vs. no) 1.59 (1.24 – 2.04) < 0.001 1.02 (0.78 – 1.33) 0.880Drug dealing† (yes vs. no) 2.04 (1.62 – 2.58) < 0.001 1.42 (1.11 – 1.81) 0.005Any treatment†† (yes vs. no) 0.78 (0.63 – 0.99) 0.037 0.82 (0.65 – 1.03) 0.089Require help injecting† (yes vs. no) 1.91 (1.47 – 2.49) < 0.001 1.51 (1.15 – 1.99) 0.003aGEE = Generalized Estimating Equation;bCI = Confidence Interval;cDTES = Downtown Eastside;†Denotes activities in the previous 6 months;††Denotes current activities;‡Denotes lifetime history; all variables significant at p<0.05 in bivariate GEE analyses were included in the multivariate GEE analysis.Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author ManuscriptMarshall et al. Page 15Table 2Bivariate and multivariate GEEa analysis of factors associated with received violence for males during follow-up (n = 670)Characteristic Unadjusted Odds Ratio(95% CIb) p - valueAdjusted Odds Ratio(95% CIb) p - valueAge (per year younger) 1.03 (1.02 – 1.04) < 0.001 1.02 (1.00 – 1.03) 0.015Aboriginal ethnicity (yes vs. no) 0.79 (0.60 – 1.04) 0.090Married (yes vs. no) 0.79 (0.64 – 0.97) 0.025 0.84 (0.68 – 1.03) 0.098Mental illness† (yes vs. no) 1.56 (1.33 – 1.83) < 0.001 1.53 (1.29 – 1.82) < 0.001Sexual abuse‡ (yes vs. no) 1.23 (0.95 – 1.60) 0.109Years injecting (per year older) 0.99 (0.97 – 1.00) 0.011 1.00 (0.99 – 1.02) 0.960Alcohol use† (yes vs. no) 1.39 (1.17 – 1.64) < 0.001 1.42 (1.20 – 1.69) < 0.001Frequent heroin injection† (> daily vs. < daily) 1.57 (1.30 – 1.88) < 0.001 1.24 (1.02 – 1.51) 0.029Frequent cocaine injection† (> daily vs. < daily) 1.25 (1.04 – 1.51) 0.017 0.92 (0.75 – 1.13) 0.424Frequent crack smoking† (> daily vs. < daily) 1.65 (1.39 – 1.96) < 0.001 1.38 (1.15 – 1.65) 0.001Binge drug use† (yes vs. no) 1.15 (0.94 – 1.42) 0.171Homelessness†† (yes vs. no) 1.86 (1.48 – 2.35) < 0.001 1.50 (1.20 – 1.94) < 0.001DTESc residency†† (yes vs. no) 1.59 (1.33 – 1.91) < 0.001 1.38 (1.15 – 1.66) < 0.001Incarceration† (yes vs. no) 1.77 (1.45 – 2.17) < 0.001 1.50 (1.22 – 1.84) < 0.001Sex trade involvement (yes vs. no) 1.08 (0.71 – 1.63) <0.728Drug dealing† (yes vs. no) 1.57 (1.29 – 1.91) < 0.001 1.20 (0.97 – 1.48) 0.099Any treatment†† (yes vs. no) 1.01 (0.85 – 1.20) 0.909Require help injecting† (yes vs. no) 1.67 (1.36 – 2.06) < 0.001 1.51 (1.22 – 1.87) < 0.001aGEE = Generalized Estimating Equation;bCI = Confidence Interval;cDTES = Downtown Eastside;†Denotes activities in the previous 6 months;††Denotes current activities;‡Denotes lifetime history; all variables significant at p<0.05 in bivariate GEE analyses were included in the multivariate GEE analysis.Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author ManuscriptMarshall et al. Page 16Table 3Physical abuse characteristics by genderCharacteristic Female(n = 501) †Male(n = 929) † p - valuePerpetrator of attack Stranger 159 (31.7%) 378 (40.7%) < 0.001 Acquaintance 218 (43.5%) 344 (37.0%) 0.017 Partner 26 (5.2%) 24 (2.6%) 0.010 Friends/Family 22 (4.4%) 33 (3.6%) 0.431 Drug Dealer 18 (3.6%) 43 (4.6%) 0.355 Police 19 (3.8%) 81 (8.7%) < 0.001 Sex trade client/worker 23 (4.6%) 8 (0.9%) < 0.001 Other/Refused 16 (3.2%) 18 (1.9%) 0.137Type of attack Beating 326 (65.1%) 568 (61.1%) 0.143 Attacked with weapons 54 (10.8%) 210 (22.6%) < 0.001 Strangled 10 (2.0%) 5 (0.5%) 0.014 Attacked with a gun 4 (0.8%) 6 (0.6%) 0.747 Robbery 62 (12.4%) 110 (11.8%) 0.767 Physically threatened 12 (2.4%) 4 (0.4%) 0.002 Other/Refused 33 (6.6%) 26 (2.8%) 0.001†Total number of reports of received violence over follow-up period. Fisher's exact test was used to compare proportions if one or more counts was lessthan or equal to five.Drug Alcohol Depend. Author manuscript; available in PMC 2009 October 1.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items