Abstract

Methods

Beginning in May 1996 people who had injected illicit drugs in the previous month were recruited into the Vancouver Injection Drug User Study. In brief, over 1400 subjects were recruited through self-referral and street outreach efforts. Eligible subjects were those who had injected illicit drugs at least once in the previous month, resided in the Greater Vancouver area and provided written informed consent. At baseline and semi-annually, subjects provided blood samples and completed an interviewer-administered questionnaire. The questionnaire was designed to elicit demographic data and information about drug use, HIV risk behaviours and drug treatment.

For our analyses, we included baseline data and data from the most recent follow-up questionnaire available (administered during January 1999 to October 2000). Demographic characteristics such as age, sex and ethnic background were derived from the baseline questionnaire. To evaluate current drug injection practices, information such as drug use and health-related characteristics such as HIV status were obtained from the most recent follow-up questionnaire.

We chose needle sharing as the dependent variable because this behaviour has been shown to have the highest risk of HIV transmission among IDUs. Needle sharing was defined as either the lending or borrowing of a used needle in the 6 months before the follow-up visit. We did not restrict our analyses to needle sharing with either casual or intimate partners, because we have recently shown that both risk behaviours are associated with HIV seroconversion among IDUs in Vancouver.²³ People who had not injected drugs in the 6 months before the most recent follow-up visit were excluded from our analyses, because we sought to evaluate predictors of needle sharing among people who were actively injecting drugs. We also examined, and report on briefly in this article, the proportion of IDUs who reported sharing needles as defined above in the cohort during the prior 5 follow-up visits.

Univariate and multivariate analyses were performed to determine factors associated with current needle sharing practices. Sociodemographic and behavioural characteristics considered in the analyses included age, sex, ethnic background, education level, HIV status, and self-report at baseline of ever having had a diagnosis of mental illness. We also considered information on the receipt of methadone treatment, being refused drug treatment, being refused sterile needles at pharmacies and requiring help injecting.

Characteristics of drug use considered in our analyses included difficulty getting sterile needles, frequency of cocaine and heroin injection, average use of needles (once v. more than once), nonfatal overdoses, injection in public, safety of needle disposal and frequency of injecting alone. As we have done previously,^9,10 frequent cocaine or heroin use was defined as injection of the drug once or more daily. Safe needle disposal was defined as the placement of needles in a “safe place” or a sharps container or the return of needles to the needle-exchange program. To evaluate the effect of police activities on drug use, we considered the number of subjects who reported that police activities affected their source of drugs.

Statistical analyses were used to compare IDUs who shared needles and those who did not share needles in the 6 months before the most recent follow-up visit. Categorical explanatory variables were analyzed using Pearson's χ² test, and continuous variables were analyzed using the Wilcoxon rank-sum test. All variables that were statistically significant at the 0.05 cutoff point were considered in a logistic regression analysis. All reported p values are two-sided.

Results

Since the study's inception, 124 participants have died (28 of HIV/AIDS, 41 of an overdose and 55 of other causes). A total of 962 participants responded to at least 1 questionnaire during the period January 1999 to October 2000 and were therefore eligible for our study. Compared with the 351 participants who did not come in for the most recent follow-up, those included in our study were more likely to be female (p = 0.001), Aboriginal (p = 0.001), HIV-positive at last follow-up (p = 0.001), older (p < 0.001) and have a high school education (p = 0.006). No significant differences were found between the groups with regard to difficulty acquiring clean needles (p = 0.30), frequency of cocaine injection (p = 0.72) or heroin injection (p = 0.93), or need for help injecting (p = 0.84). Of the 962 participants eligible for our study, 776 (80.7%) reported injecting drugs in the 6 months before the most recent follow-up visit and were therefore included in our analyses; 562 (72.4%) reported not sharing needles and 214 (27.6%) reported sharing needles in that 6-month period. The proportions of active IDUs who reported needle sharing at the 4 follow-up visits before the study period were 31.2%, 22.7%, 23.5% and 25.5% respectively. Of the 247 HIV-positive IDUs, 20.2% reported sharing needles in the 6 months before the most recent follow-up visit. Of the 776 IDUs in our study, 106 (13.7%) injected drugs in public, and 581 (74.9%) injected alone at least once.

The results of the univariate analysis of sociodemographic and health-related characteristics are shown in Table 1. Being refused sterile needles at pharmacies (odds ratio [OR] 2.0) and requiring help to inject drugs (OR 1.9) were positively associated with sharing needles. Although not achieving statistical significance, having a diagnosis of mental illness was associated with sharing needles (OR 1.4, p = 0.07). Alternatively, being older (OR 0.97 per year of age [95% CI 0.95–0.99]; data not shown), and being HIV positive (OR 0.6) were inversely associated with needle sharing. We found no evidence that ethnic background, education level, sex, receipt of methadone treatment or being refused drug addiction treatment were associated with needle sharing.

Table 1

Open in a separate window

The results of the univariate analysis of drug use and behavioural characteristics are shown in Table 2. Factors positively associated with needle sharing were difficulty getting sterile needles (OR 3.1), frequent cocaine injection (OR 1.6), frequent heroin injection (OR 1.8), use of needles more than once on average (OR 2.0), nonfatal overdose (OR 1.7) and reporting that police activities had affected the source of drugs (OR 1.9). We found no evidence that injection in public, unsafe needle disposal or injecting alone were associated with needle sharing.

Table 2

Open in a separate window

The variables that were found to be independently associated with needle sharing in the stepwise logistic regression analyses are listed in Table 3. Difficulty getting sterile needles (adjusted OR 2.7), requiring help to inject drugs (adjusted OR 2.0), frequent cocaine injection (adjusted OR 1.6), frequent heroin injection (adjusted OR 1.5) and use of needles more than once on average (adjusted OR 1.8) were all positively associated with needle sharing. HIV-positive status was inversely associated with needle sharing (adjusted OR 0.5).

Table 3

Open in a separate window

Interpretation

Despite targeted police efforts and a large needle-exchange program in Vancouver's Downtown Eastside, 27.6% of the IDUs included in our study reported sharing needles, and 9.7% had had a nonfatal overdose in the 6 months before the most recent follow-up visit. Having difficulty getting sterile needles, needing help injecting, reusing needles, and frequent cocaine and heroin injection were all associated with needle sharing.

Several of our findings suggest that barriers to sterile needle use persist despite the presence of a large needle-exchange program.⁹ Although expanding the program would likely help to reduce needle sharing further, several risk factors remained independently associated with needle sharing after adjustment for difficulty getting needles. Furthermore, 19.1% of the participants included in our study shared needles even though they did not report having difficulty getting sterile needles. All of these factors suggest that expansion of the needle-exchange program alone will not be sufficient to eliminate this risk.

Several conditions, such as lack of experience or physical disability, may place IDUs in need of help with injections. The high prevalence of assisted injection has been documented in other settings,²⁴ although the strong association with HIV risk behaviour has not been previously established to our knowledge. Again, adequate availability of sterile needles will probably not be sufficient to mitigate this risk behaviour.

We also found that subjects who were HIV-positive were half as likely as HIV-negative IDUs to share needles. This suggests a benefit of HIV testing and counselling in this community.²⁵ However, despite substantial outreach services offering testing and counselling to IDUs in Vancouver, only 15.0% of the participants in our study had ever had an HIV test before recruitment into the study (unpublished data).

Another major problem associated with injection drug use is death from overdose. In Canada, British Columbia has the highest number of such deaths per capita, with about 4.7 per 100 000 population annually, and in several recent years illicit drugs have been the leading cause of death among adults 30 to 49 years of age.^12,14 Although needle-exchange programs have been associated with reductions in overdoses,²⁶ overdoses continue to occur all too frequently among IDUs.

Several of the variables we examined highlight the public disorder problems associated with injection drug use. For instance, in the 6 months before the most recent follow-up visit 13.7% of the participants in our study reported injecting drugs in public and 45.6% reported that they did not always practise safe needle disposal.

We found a positive association between police activities and needle sharing. Although further study of this association is required, the potentially negative consequences of policing efforts on HIV risk behaviour, such as creating fear of possessing needles, have been reported elsewhere.^27,28

Our study has several limitations. Compared with the IDUs included in our study, those who did not come in for the most recent follow-up had a number of characteristics (e.g., younger and more likely to be HIV negative) that may make them more likely to be involved in needle sharing. Furthermore, it has been shown previously that IDUs may substantially underreport HIV risk behaviour and that HIV testing and counselling that accompany cohort studies such as ours may reduce risk behaviour over time.^29,30 Therefore, we may have underestimated the extent of needle sharing among IDUs.

In summary, our data demonstrate a continued health crisis among IDUs. Furthermore, the proportion of IDUs who reported needle sharing did not decrease over the past 5 follow-up visits, despite the presence of a large needle-exchange program and targeted law enforcement efforts.

Meanwhile, in several European cities, the risk factors we identified have proven amenable to improvement through the establishment of safer injecting rooms as part of comprehensive harm reduction strategies. Increases in HIV testing and counselling, health and social functioning, and drug addiction treatment have occurred among clients of safer injecting rooms in these cities.^15,16,31 Conversely, the incidence of HIV risk behaviour including needle sharing, hospital admissions, improper needle disposal, drug injection in public places and death from overdose have all decreased.^15,16,31,32 In fact, there has not been a fatal overdose in a safer injecting room since their establishment in the mid-1980's.¹⁷ In the present study we did not evaluate safer injecting rooms per se. We have merely identified risk factors among IDUs that have proven amenable to improvement through the availability of such facilities in other settings.^15,16,31,32 Given the high prevalence of HIV risk behaviours, overdoses and other health-related concerns that persist in Vancouver, it is crucial to evaluate whether the European experience with safer injecting rooms can be replicated in Canada.

Footnotes

References