Use of North America’s first medically supervised safer injecting facility among HIV-positive injection drug users

Abstract

The objective of this study was to examine supervised injecting facility (SIF) use among a cohort of 395 HIV-positive injection drug users (IDUs) in Vancouver, Canada. The correlates of SIF use were identified using generalized estimating equation analyses. In multivariate analyses, frequent SIF use was associated with homelessness (adjusted odds ratio [AOR] = 1.90), daily heroin injection (AOR = 1.56), and daily cocaine injection (AOR = 1.59). The reasons given for not using the SIF included a preference for injecting at home and already having a safe place to inject. The SIF services most commonly used were needle exchange and nursing services. The SIF appears to have attracted a high-risk subpopulation of HIV-positive IDUs; this coverage perhaps could be extended with the addition of HIV- specific services such as disease monitoring and the provision of antiretroviral therapy.

INTRODUCTION

Many cities throughout the world are contending with an array of health consequences associated with the injection of illicit substances (Mathers et al., 2008). Prominent health problems resulting from injection drug use include HIV/AIDS, hepatitis C, abscesses, osteomyelitis, endocarditis, and high rates of overdose mortality (Milloy, Kerr, Mathias et al., 2008; Patrick et al., 2001). Nearly one third of all new HIV infections outside of subSaharan Africa are attributed to injection drug use (UNAIDS, 2006). The provision of health care for injection drug users (IDUs) is often severely compromised by limited uptake of primary care services and over dependence on emergency departments (Kerr et al., 2004; O’Connor, Selwyn, & Schottenfeld, 1994; Palepu et al., 2001).

In response to the ongoing public health issues arising from injection drug use, supervised injecting facilities (SIFs) have been implemented in various international settings (Dolan et al., 2000; Hedrich, 2004). SIFs permit the injection of preobtained drugs under the supervision of health care staff (Kimber, Dolan, van Beek, Hedrich, & Zurhold, 2003). Although these facilities have existed in European countries since the 1980s, the first North American SIF opened in 2003 in the Downtown Eastside of Vancouver (Stoltz et al., 2007). This facility, known as Insite, has been functioning at full capacity (700 visits per day) since its inception (Wood et al., 2006). The specific operating procedures vary across SIFs, but these sites typically provide the supplies necessary to inject drugs in a sterile manner, including sterile needles, alcohol swabs, sterile water, tourniquets, and cookers, if necessary (Kimber, Dolan, & Wodak, 2005). Like most other SIFs, Insite offers overdose response, basic primary care, safe injecting education, and referrals to treatment services (Kerr, Kimber, DeBeck, & Wood, 2007).

The vast majority of prevention interventions that target IDUs focus on preventing the acquisition of HIV infection among HIV-negative IDUs, although some interventions targeting HIV-positive IDUs have also been developed, evaluated, and found to be effective (Collins et al., 2010; Lyles et al., 2007). Educational interventions, addiction treatment, and needle exchange programs have successfully reduced the rate of new HIV infections (Des Jarlais et al., 1998; Mehta et al., 2006), but unacceptably high HIV incidence levels persist in many settings with established IDU-driven HIV epidemics (Choopanya et al., 2002; Hamers & Downs, 2003; Wood et al., 2008). This is because of low coverage of these programs (Mathers et al., 2010), as well as overreliance on individually focused prevention strategies (Rhodes, 2002). Reviews of the available evidence suggest that HIV prevention interventions for IDUs that focus solely on individual behavior change are likely to result in only a partial reduction of HIV transmission risk, on the order of 25% to 40% (Coyle, Needle, & Normand, 1998; Heimer, Bray, Burris, Khoshnood, & Blankenship, 2002). This has led to the call for social and structural interventions to reduce drug-related harms among HIV-negative and HIV-positive IDUs (Rhodes, Singer, Bourgois, Friedman, & Strathdee, 2005).

In response to the limitations of conventional programs for IDUs, a novel area of investigation has focused on the broader “risk environment” of IDUs and the role that social, environmental, and structural factors play in mediating risk-taking and health outcomes (Rhodes, 2002; Rhodes et al., 2005). Rhodes’s Risk Environment Framework reflects recent theorizing in the field of health and social behavior that explores how the broader environment links to risk behavior and health inequality (Bourgois, 1998; Burris, Blankenship, & Donoghoe, 2004; Des Jarlais, 2000; Galea, Ahern, & Vlahov, 2003; Singer et al., 2000). The framework provides a heuristic model for delineating the multiple social, structural, and environmental factors that produce and reduce both health-related risks and health service utilization (Rhodes, 2002). By focusing on the risk environment of IDUs, it is possible to examine the interactions of physical, social, and structural factors with individual factors at micro and macro levels and to study the effects of these interactions on risk reduction practices and health service access. The Risk Environment Framework also highlights the need for microenvironmental interventions, such as SIFs, which serve to reduce unsafe injecting (Rhodes et al., 2006). Given these issues, and the lack of research pertaining to SIF use among HIV-positive IDUs, we sought to investigate the prevalence and correlates of SIF use among HIV-positive IDUs in Vancouver, Canada.

METHODS

The AIDS Care Cohort to evaluate Exposure to Survival Services (ACCESS) is an open prospective cohort of HIV-positive IDUs in Vancouver, Canada. The cohort was populated through snowball sampling and extensive street outreach methods in the city’s Downtown Eastside (the drug use and HIV epicenter). To be enrolled into the ACCESS study, participants had to be 18 years of age or older, HIV-positive, to have used injection drugs, and to have provided written informed consent. Participants answered a standardized interviewer-administered questionnaire and provided blood samples for serologic analysis at baseline and at every 6-month follow-up visit. The study protocol has received ethical approval by the Providence Health Care—University of British Columbia Research Ethics Board.

The local setting is unusual in that antiretroviral therapy and laboratory monitoring are available free of cost to HIV-infected British Columbia residents. As well, there is a provincewide centralized antiretroviral dispensation program and HIV/AIDS laboratory, which enables a complete prospective profile of all patient CD4 cell count determinations and plasma HIV-1 RNA levels (Amplicor Monitor assay, Roche Molecular Systems, Mississauga, Canada), as well as a complete prospective profile of antiretroviral therapy use among cohort participants. This consists of the specific antiretroviral agents prescribed, as well as a previously validated measure of patient adherence derived from prescription refill compliance (Wood, 2008). The universal health care system and centralized nature of the HIV/AIDS treatment program make possible the examination of HIV-related outcomes in a setting where financial barriers to health care and HIV treatment are largely eliminated.

In the present study, we included all participants who were recruited and who completed at least one follow-up interview between December 2005 and May 2008. The primary outcome of interest was reporting frequent SIF use, which was operationally defined as using the SIF for more than 25% of injections. This definition is consistent with previous studies (Stoltz et al., 2007). Consistent with the Risk Environment Framework (Rhodes, 2002), we selected a range of individual and contextual variables that were deemed to be potentially associated with the outcome. These variables were age, gender, Aboriginal ancestry (yes vs. no), residence in the Downtown Eastside (yes vs. no), homelessness (yes vs. no), frequent crack use (yes vs. no), daily heroin injection (yes vs. no), daily cocaine injection (yes vs. no), daily methamphetamine use (yes vs. no), participation in the sex trade (yes vs. no), public injecting (yes vs. no), borrowing needles/syringes (yes vs. no), lending needles/syringes (yes vs. no), nonfatal overdose (yes vs. no), and methadone maintenance therapy (yes vs. no). All behavioral variables referred to the 6-month period prior to the interview; for consistency with previous literature, we defined public injecting as any public injections in the last 6 months (DeBeck et al., 2009). From the confidential linkage to the centralized antiretroviral dispensary, we added these variables: current ARV use, HIV-1 RNA level (≥ 500 copies/mL vs. < 500 copies/mL), and CD4 cell count (≥ 200 cells/μL vs. > 200 cells/μL). If more than one measure of viral load or CD4 cell count was available in any 6-month period, we used the mean of all available measurements. If no measures of either viral load or cell count were available, we used the most recent measure available prior to the interview.

To identify factors associated with SIF use at baseline, a univariate analysis was conducted whereby Pearson’s chi-square text was used to examine associations between SIF use and the independent variables of interest. Next, we examined univariate associations between all potential explanatory variables and SIF use, including baseline and all available follow-up data. Because serial measures for each variable were available for each subject, we used generalized estimating equation (GEE) logistic regression for the analysis of correlated data. This method allowed for the identification of factors associated with the outcome over the entire study period and calculated standard errors using an exchangeable correlation structure, adjusted by multiple observations for each individual. As each individual could report SIF use or not during different study visits, GEE analyses identified factors correlated with periods of SIF use both within and between individuals. Because GEE models are a form of marginal longitudinal analysis, they have been used to analyze datasets containing repeated measures, such as longitudinal cohorts (Wood et al., 2004). Following examination of the univariate results, we fitted a multivariate GEE model using a model-building protocol defined a priori to include all explanatory variables significant (p < .05) in unadjusted analyses. All statistical procedures were performed using SAS software (SAS, Cary, NC). All p values are two-sided.

We then conducted subanalyses to examine barriers to use and services used within the facility. Specifically, individuals who reported never using the SIF were asked to indicate why they did not use the facility. Among individuals who reported use of the SIF, we also examined which health-related services they used during their visits to the SIF. For these questions, individual responses were checked on a list of possible responses, and responses that were not already on the list were captured in a space for “other” responses. Descriptive data obtained via these subanalyses are presented as proportions of total observations recorded over the study period.

RESULTS

In total, 395 participants were enrolled during the study period and were eligible for this analysis, including 142 (35.9%) women and 156 (39.5%) individuals of Aboriginal ancestry. The median age of the participants was 41.9 (interquartile range 36.3–47.7). Among these individuals, 104 (26.4%) reported using the SIF for more than 25% of their injections.

Table 1 shows the factors associated with frequent SIF use stratified by sociodemographic and behavioral variables at baseline. As shown here, factors positively associated with frequent SIF use at baseline included gender, homelessness, daily heroin injection, daily cocaine injection and non-fatal overdose, whereas factors negatively associated with SIF use included age and current receipt of antiretroviral therapy.

Table 1.

Factors associated with frequent SIF use among HIV-positive injection drug users at baseline (n = 395)

Characteristic	≥ 25% of Injections n (%) n = 104	< 25% of Injections n (%) n = 291	Odds ratio (95% CI)	p Value
Age
≥ 41.9 years of age	33 (31.7)	160 (55.2)	0.38 (0.24–0.61)	<.001
< 41.9 years of age	71 (68.3)	130 (44.8)
Gender
Female	47 (45.2)	95 (32.6)	1.70 (1.08–2.69)	.022
Male	57 (54.8)	196 (67.4)
Aboriginal ancestry
Yes	41 (39.4)	115 (39.5)	1.00 (0.63–1.58)	.986
No	63 (60.6)	176 (60.5)
Residence in DTES*
Yes	84 (80.8)	207 (71.1)	1.70 (0.98–2.95)	.056
No	20 (19.2)	84 (28.9)
Homelessness
Yes	38 (36.5)	50 (17.2)	2.78 (1.68–4.59)	<.001
No	66 (63.5)	241 (82.8)
Daily crack use
Yes	54 (51.9)	134 (46.0)	1.27 (0.81–1.98)	.303
No	50 (48.1)	157 (54.0)
Daily heroin injection
Yes	48 (46.2)	55 (18.9)	3.68 (2.27–5.97)	<.001
No	56 (53.8)	236 (81.1)
Daily cocaine injection
Yes	24 (23.1)	27 (9.3)	2.93 (1.60–5.36)	<.001
No	80 (76.9)	264 (90.7)
Daily methamphetamine injection
Yes	6 (5.8)	14 (4.8)	1.21 (0.45–3.24)	.702
No	98 (94.2)	277 (95.2)
Public injecting
Yes	87 (83.7)	192 (66.0)	2.64 (1.49–4.68)	.001
No	17 (16.3)	99 (34.0)
Syringe borrowing
Yes	10 (9.6)	33 (11.3)	0.83 (0.40–1.75)	.628
No	94 (90.4)	258 (88.7)
Syringe lending
Yes	2 (1.9)	15 (5.2)	0.36 (0.08–1.61)	.163
No	102 (98.1)	276 (94.8)
Sex trade involvement
Yes	19 (18.3)	47 (16.2)	1.16 (0.65–2.09)	.619
No	85 (81.7)	244 (83.8)
Non-fatal overdose
Yes	16 (15.4)	19 (6.5)	2.60 (1.28–5.28)	.006
No	88 (84.6)	272 (93.5)
Methadone use
Yes	36 (34.6)	131 (45.0)	0.65 (0.41–1.03)	.065
No	68 (65.4)	160 (55.0)
CD4 cell count
≥ 200 cells/μL	72 (69.2)	212 (72.9)	1.19 (0.73–1.95)	.481
> 200 cells/μL	32 (30.8)	79 (27.1)
Viral load
≥ 500 copies/mL	29 (27.9)	73 (25.1)	0.65 (0.35–1.18)	.150
< 500 copies/mL	75 (72.1)	218 (74.9)
Antiretroviral use
Yes	38 (40.9)	141 (55.7)	0.55 (0.34–0.89)	.014
No	55 (59.1)	112 (44.3)

^*DTES — the local epicenter of injection drug use and drug-related infections.

Table 2 shows the bivariate GEE analyses of factors associated with SIF use. As indicated here, these analyses yielded significant correlations between frequent SIF use and: age (odds ratio [OR] = 0.63; 95% confidence interval [CI]: 0.01–0.06); gender (OR = 1.47; 95% CI: 1.01–2.15); homelessness (OR = 2.08; 95% CI: 0.34–0.68); daily heroin injection (OR = 2.27; 95% CI: 1.61–3.20); daily cocaine injection (OR = 1.93; 95% CI: 1.32–2.84); and antiretroviral use (OR = 0.70; 95% CI: 0.51–0.96).

Table 2.

Bivariate and multivariate GEE analysis of factors associated with frequent supervised injection facility use among HIV-positive injection drug users (n = 395)

	Unadjusted		Adjusted
Characteristic	Odds Ratio (95% CI)	p Value	Odds Ratio (95% CI)	p Value
Age
≥ median (41.9) vs. < median	0.63 (0.01–0.06)	.003	0.78 (0.54–1.11)	.166
Gender
(female vs. male)	1.47 (1.01–2.15)	.046	1.19 (0.79–1.81)	.410
Aboriginal ancestry
(yes vs. no)	1.04 (0.71–0.52)	.854
Reside in DTES*
(yes vs. no)	1.36 (0.91–2.04)	.133
Homelessness
(yes vs. no)	2.08 (0.34–0.68)	<.001	1.90 (1.30–2.77)	.001
Daily crack use
(yes vs. no)	1.01 (0.78–1.30)	.939
Daily heroin injection
(yes vs. no)	2.27 (1.61–3.20)	<.001	1.56 (1.08–2.25)	.017
Daily cocaine injection
(yes vs. no)	1.93 (1.32–2.84)	<.001	1.59 (1.05–2.42)	.030
Daily methamphetamine injection
(yes vs. no)	0.93 (0.45–1.93)	.848
Public injecting
(yes vs. no)	2.10 (1.53–2.89)	<.001	1.26 (0.90–1.75)	.181
Sex trade involvement
(yes vs. no)	1.37 (0.88–2.11)	.160
Non-fatal overdose
(yes vs. no)	1.27 (0.83–1.97)	.277
Methadone maintenance therapy enrollment
(yes vs. no)	0.85 (0.61–1.18)	.328
CD4 cell count
(≥ vs. < 200 cells)	1.00 (0.71–1.42)	.983	0.97 (0.66–1.43)	.890
Viral load
(≥ vs. < 500 copies/mL)	1.00 (0.91–1.34)	.549
Antiretroviral use
(yes vs. no)	0.70 (0.51–0.96)	.026	0.86 (0.62–1.21)	.393

^*DTES — Downtown Eastside.

The multivariate GEE analysis is also shown in Table 2 and reveals that frequent SIF use was independently associated with: homelessness (adjusted odds ratio [AOR] = 1.90; 95% CI: 1.30–2.77); daily heroin injection (AOR = 1.56; 95% CI: 1.08–2.25); and daily cocaine injection (AOR = 1.59; 95% CI: 1.05–2.42).

The subanalysis of reasons for not using Insite yielded 341 observations between baseline and the most recent follow-up. Of these 341 observations, 106 (31%) included a stated preference for injecting at home, 73 (21%) included a report of having another safe place to inject, 34 (10%) included a stated preference for keeping drug use private, 27 (8%) included a report of living too far from Insite, 21 (6%) involved some dislike of the design or operations of the facility (e.g., “entrance is too public,” “don’t want to register”), 20 (6%) involved a report of requiring help injecting (which is prohibited at Insite), 14 (4%) involved a report of wanting to avoid the Downtown Eastside, and the remaining 46 (13%) reported various other reasons. From baseline to the most recent follow-up, those who used Insite reported which services they used there. Of the 1,051 observations, 612 (58%) reports included an account of using services other than the injecting room. In terms of health-related service use, there were 513 (49%) reports of needle exchange use, 174 (17%) reports of receiving care from a nurse, 79 (8%) reports of alcohol and drug counseling use, and 24 (2%) reports of receiving a health care referral.

DISCUSSION

Our results demonstrate that 26% of HIV-positive IDUs frequently attended a local supervised injection facility. Consistent with the Risk Environment Framework (Rhodes, 2002), in multivariate analysis, frequent SIF use was independently associated with both individual and contextual factors, including homelessness, daily heroin injection, and daily cocaine injection. The primary reasons given for not accessing the SIF included a preference for injecting at home and already having a safe place to inject. The health care services most commonly used at the SIF, aside from the supervised injecting room, included the needle exchange and nursing care.

Our findings indicate that the SIF has attracted a population of HIV-positive IDUs who are at heightened risk for adverse health outcomes, which is consistent with previous studies involving other IDU populations (Wood et al., 2006). To our knowledge, though, this is the first study to specifically examine SIF use among HIV-positive IDUs. As in previous studies, IDUs in this study who used the SIF frequently were more likely to inject heroin and cocaine at least daily (Tyndall et al., 2006). High-intensity cocaine use is explained by the brevity of cocaine’s half-life, which ranges between 40 and 60 minutes (McCoy, Lai, Metsch, Messiah, & Zhao, 2004). As a result, cocaine users often inject 20–30 times per day and are at high risk to acquire or transmit infectious diseases, including HIV (Tyndall et al., 2003). Therefore, it is encouraging that the SIF is attracting HIV-positive IDUs who frequently inject cocaine. The association between SIF use and daily heroin injection has also been reported in previous studies (Wood et al., 2006). Given that heroin-related overdose accounts for a significant amount of preventable morbidity and mortality among IDUs (Darke & Hall, 2003), it is encouraging that the SIF is attracting this subgroup and therefore likely reducing the risk of harm associated with opiate-related overdose (Kerr, Small, Moore, & Wood, 2007; Milloy, Kerr, Tyndall, Montaner, & Wood, 2008).

Our finding that homelessness was associated with more frequent SIF use has important implications. Homeless individuals are at heightened risk for various negative health outcomes (Culhane, Gollub, Kuhn, & Shpaner, 2001; Galea & Vlahov, 2002). As suggested by the Risk Environment Framework (Rhodes et al., 2006), social conditions such as homelessness shape behavioral practices such as risky injecting. Homelessness has been strongly associated with public injecting, and HIV-positive IDUs who inject in public are vulnerable to bacterial infections arising from unsterile injecting (Young et al., 2004). Indeed, because of compromised immunity, HIV-positive IDUs have a heightened susceptibility to bacterial infections that commonly result from frequent nonsterile injections (Lloyd-Smith et al., 2005). These injection-related bacterial infections are associated with considerable morbidity and account for the majority of emergency room visits and hospitalizations among IDUs (Ebright & Pieper, 2002; Palepu et al., 2001). These issues and the present findings should be interpreted in the context of an earlier qualitative study that revealed that the Vancouver SIF has been effective in increasing access to assessment, care, and treatment of injection-related infections (Small, Wood, Lloyd-Smith, Tyndall, & Kerr, 2008).

Although frequent SIF use was not independently associated with use of antiretroviral therapy (ARV), the bivariate association between SIF use and a reduced likelihood of being on ARV raises questions about the opportunities to provide enhanced HIV treatment via the SIF, especially in settings such as British Columbia, Canada, where ARV is provided free under the universal health care system. Previous studies have indicated that HIV-positive IDUs experience comparatively low rates of access and adherence to HIV treatment, and consequently this population has not benefited from modern HIV treatments to the same extent that other HIV-affected populations have (Kresina, Bruce, & McCance-Katz, 2009; Wood et al., 2003). Given that the SIF in Vancouver is attracting high-risk IDUs with a lower likelihood of receiving HIV treatment, the SIF could provide a venue for enhanced HIV care, including viral load and CD4 count testing, vaccinations and other preventive measures, and distribution of antiretroviral therapies. It has been suggested previously that the integration of existing harm reduction services, including SIFs, into treatment efforts that target IDUs could enhance HIV disease monitoring and treatment provision (Kerr, Kimber et al., 2007), and it is noteworthy that another local SIF (the Dr. Peter Centre) limits its clientele to HIV-positive IDUs and provides many HIV-specific services (Krüsi, Small, Wood, & Kerr, 2009).

The subanalyses of SIF use revealed several reasons why some HIV-positive IDUs refrain from using Insite. Already injecting in a safe place, preferring to inject at home, the distance from the site, preferring to inject in private, and requiring help injecting were frequently reported barriers to using the SIF. Consistent with the Risk Environment Framework (Rhodes et al., 2006), efforts to remove structural barriers to prevention and treatment services are critical to ensuring the optimal impact of such services. Efforts aimed at increasing coverage of SIFs locally should therefore focus on both increasing the number of SIFs (and their geographic coverage) and modifying rules that prevent assisted injecting. Future research should seek to evaluate the impact of these types of structural changes to SIF program delivery.

We also found that HIV-positive participants in this study used various services at the SIF aside from the supervised injecting service. Given the high intensity of injecting among SIF users, it is encouraging that many IDUs were using the syringe exchange and drug and alcohol counseling services. As well, in light of the fact that many HIV-positive IDUs were using the SIF’s nursing services, the SIF could potentially expand the provision of HIV treatment and care by providing a range of HIV-specific services (e.g., disease monitoring, referral to HIV specialist physicians, and provision of daily dispensed antiretroviral therapy). In this particular setting, where health care is available to all free of charge, the barriers to treatment and care faced by many HIV-positive IDUs are other than financial ones. The SIF is therefore an ideal environment in which to offer HIV-specific services to IDUs, as it makes treatment and care more readily accessible to this population.

This study has several limitations. First, like most other cohort studies involving high-risk IDUs, ACCESS is not a random sample. Therefore, our study findings may not generalize well to the larger population of HIV-positive IDUs in Vancouver. Second, given the variations across settings, including differences in high-risk behaviors in other urban environments, our findings may not generalize well to HIV-positive IDUs in other locations. Third, we relied on a dichotomous outcome, and several of our independent variables were also dichotomized. Although the use of a continuous outcome may have allowed for a more nuanced analysis and less restricted variance across measures, we note that we were able to detect several significant associations between the outcome and independent variables that were considered. Last, we relied on self-reported measures, which may have introduced response biases into our results, such as socially desirable responding. Thus, we may have underestimated the sensitive behaviors and experiences, such as injection drug use and sex work involvement, among the participants.

CONCLUSION

In summary, approximately one-quarter of the HIV-positive IDUs participating in our study used the local SIF on a frequent basis. These individuals were more likely than less frequent SIF users to be high-intensity injectors and homeless. A significant number of IDUs reported not using the SIF because they preferred to inject at home or had another place to safely inject. HIV-positive IDUs using the SIF made use of a range of health-related services at the facility, including needle exchange, nursing care, and drug and alcohol counseling. To our knowledge, this is the first study to examine SIF use, as well as barriers to SIF use, among HIV-positive IDUs. Although the SIF appears to be an effective micro-environmental intervention that is attracting high-risk HIV-positive IDUs, its full potential has yet to be realized, as several barriers remain that discourage certain IDUs from using the facility. Modifications to and expansion of the SIF program may result in higher coverage of SIF services among HIV-positive IDUs. Further, the services offered at the SIF could be expanded to include HIV-specific services such as disease monitoring and the provision of antiretroviral therapy.