High HIV testing and low HIV prevalence among injecting drug users attending the Sydney Medically Supervised Injecting Centre
Correspondence to: Allison M. Salmon, National Centre HIV Epidemiology and Clinical Research, University of New South Wales, Level 2, 376 Victoria Street, Darlinghurst, New South Wales, 2010. Fax: (02) 9385 0920; e-mail: firstname.lastname@example.org
Objective: Measure the self-reported prevalence of HIV, history of HIV testing and associated risk factors among injecting drug users (IDUs) attending the Sydney Medically Supervised Injecting Centre (MSIC).
Methods: Cross-sectional survey of IDUs attending the Sydney MSIC (n=9,778).
Results: The majority of IDUs had been tested for HIV (94%), most within the preceding 12 months. Self-reported prevalence of HIV was only 2% (n=162) and homosexuality (AOR 20.68), bisexuality (AOR 5.30), male gender (AOR 3.33), mainly injecting psychostimulants (AOR 2.02), use of local health service (AOR 1.56) and increasing age (AOR 1.62) were independently associated. Among the 195 homosexual male sample 23% were self-reported being HIV positive. HIV positive homosexual males were more likely to report mainly psychostimulant injecting than other drugs, a finding not replicated among the heterosexual males.
Conclusions: The associations in this sample are consistent with other data indicating Australia has successfully averted an epidemic of HIV among heterosexual IDUs. The absence of any significant associations between HIV positive sero-status and the injecting-related behaviours that increase vulnerability to BBV transmission suggests that HIV infection in this group may be related to sexual behaviours. In particular, the strong associations between homosexual males and psychostimulant injectors with HIV positive sero-status suggests that patterns of infection within this group reflect the epidemiology of HIV in Australia more generally, where men who have sex with men remain most vulnerable to infection.
Human immunodeficiency virus (HIV) antibody prevalence has remained low among injecting drug users in Australia over the past decade or more, ranging from 0.8% to 1.5% among those attending Needle and Syringe Programs (NSPs) in the years 1995 to 2007.1 In contrast to Australian levels, a recent global overview found that while HIV prevalence among injecting drug user/s (IDUs) was less than 5% in 43 of the countries/territories reviewed, it was 20% or more in 25 of the 78 countries countries/territories which included a number of developed countries such as the US and Canada.2
The low prevalence of HIV among IDUs in Australia has been attributed to the various harm reduction initiatives implemented in the 1980s and 1990s, and specifically the early and wide spread availability of Needle and Syringe Programs.3–5 This public health measure is one example of Australia's commitment to harm reduction which includes the trial supervised injecting facility, which opened in Sydney in May 2001.
This supervised injecting facility, the first of its kind established outside of Europe, is situated in Kings Cross, which is the ‘red light’ district of Sydney and an area where there had been a visible increase in public drug use and associated public health and public order problems during the 1990s. One of the key political arguments for the establishment of the Sydney Medically Supervised Injecting Centre (MSIC) was the potential to reduce the spread of HIV and the hepatitis C virus (HCV).6 The facility was designed to accommodate a high throughput of IDUs for injecting episodes and recovery, while health and social welfare needs are assessed and where appropriate, referred to other services.7 Data from the first six years of operation provide evidence that the facility has been successful in reaching a marginalised population of IDUs, that is, long-term injectors, homeless injectors, those who inject frequently and in public places, who are not accessing health care services, injectors with a history of unemployment and imprisonment, those with low education levels and those engaged in sex work.8 The objective of this study was to use the MSIC setting to explore of HIV testing histories and HIV prevalence among a large sample of IDUs and an assessment of Australia's HIV testing and prevention policies/strategies among this marginalised group of IDUs.
In our study, we present the prevalence of, and factors associated, with self-reported lifetime HIV testing and self-reported HIV positive sero-status in IDUs attending the Sydney MSIC. HIV positive sero-prevalence by sexual orientation and psychostimulant use was further examined.
Between May 2001 and April 2007 data were collected from IDUs registering for the first time to inject drugs at the Sydney MSIC as part of the routine assessment of their eligibility to use the service.7 The eligibility criteria include being 18 years of age or over; having injected drugs previously; not being known to be or obviously pregnant; not being accompanied by a child/ren and not being intoxicated. Clients are asked to consent to use of their data for service evaluation.
Self-reported HIV testing history and HIV positive sero-status at registration were collected. We were interested in potential associated factors with both outcome factors, including: gender, age, indigenous status, injecting frequency, history of public injecting, main drug injected, sharing of needles and syringes and injecting equipment in the past month, accommodation status, education status, overdose history, use of local IDU health services, imprisonment in the past 12 months, main source of income, sexual orientation, recent sex work and years of injecting drug use. Client characteristics were examined at univariate level using odds ratios (ORs) and 95% confidence intervals. Multiple logistic regression analysis, using forward stepwise methods determined characteristics independently associated with HIV testing and HIV positive sero-status. A p-values of <0.05 were considered statistically significant and p-trends are reported for ordinal variables and show whether the observed trend is statistically significant. The prevalence of self-reported HIV positive sero-status by sexual orientation and psychostimulant use was further examined. The study was conducted as part of the Phase 2 Evaluation of the Sydney MSIC and approved by the University of NSW Human Research Ethics Committee. All analyses were conducted using STATA 8.2 (STATA Corporation, College Station, TX).
In the six years from May 2001 to April 2007, 9,778 IDUs registered to inject drugs at the Sydney MSIC. This sample of clients were mostly male (74%), were aged 33 years on average (range 18-70 years), had been injecting drugs for 14 years on average (range <1-51 years); the majority had not completed high school (71%) and were unemployed (61%). Approximately a quarter of the sample had unstable accommodation (24%) and had been imprisoned in the 12 months prior to registering (24%). More than half (51%) of the sample reported heroin as the main drug they injected in the past month and 20% were mainly psychostimulant users (cocaine and/or methamphetamines). Sixty per cent had received some form of drug treatment and 35% reported having experienced one or more drug overdoses; 4% identified as homosexual or lesbian and 8% as bisexual.
History of HIV testing
A subset (n=8,889) of the sample answered “Have you ever been tested for HIV?”. Of this subset 8,344 or 94% had been tested for HIV prior to registering with the Sydney MSIC. More specifically, 7,091 reported being tested for HIV in the preceding year while 1,253 reported being tested more than a year ago and 545 IDUs reported never having been tested.
After adjustment for those factors significantly associated at the univariate level, it was found that IDUs tested, either in the past twelve months or prior, were significantly more likely to be homosexual than heterosexual (AOR 3.24; CI 1.54-6.70; p-value=0.001) and have a history of drug treatment (AOR 2.33; CI 1.85-2.94; p-value<0.001). Other characteristics significantly associated with HIV testing were imprisonment in the previous twelve months (AOR 1.87; CI 1.41-2.49; p-value<0.001); being female (AOR 1.83; CI 1.39-2.40; p-value<0.001); being aged 25 years or over (AOR 1.25; CI 1.12-1.39; p-trend<0.001); bisexual compared to heterosexual (AOR 1.66; CI 1.03-2.66; p-value<0.001); having a history of one or more drug overdoses (AOR 1.32; CI 1.14-1.52; p-value<0.001); and unemployed (AOR 1.23; CI 1.00-1.51; p-value=0.05).
Among those who had been tested, a total of 7,624 reported knowing the results of their test. Of this sub-group, 2% (n=192) reported HIV positive sero-status, and of these only 6% (n=10) reported ever having received antiretroviral treatment. At the multivariate level, IDUs who were HIV positive were more likely to report being homosexual (AOR 20.43; CI 13.21-31.59; p-value<0.001) or bisexual (AOR 5.30; CI 3.15-8.93; p-value<0.001) than heterosexual and were more likely to be male than female (AOR 3.33; CI 1.96-5.56; p-value<0.001). HIV positive IDUs were more likely to report being users of the local IDU health care services (AOR 1.56; CI 1.07-2.27; p-value=0.025) and be aged 30 years or more (AOR 1.62; CI 1.35-1.96; p-trend=<0.001). They were also more likely to report the drug most injected was a psychostimulant (AOR 2.02; CI 1.38-2.96; p-value<0.001) as compared to heroin (Table 1).
Table 1. Characteristics of IDUs associated with self- reported negative and positive HIV test results: May 2001 to end April 2007
|Male||136 (84)||5454 (73)||3.33 (1.96-5.56)||<0.001|
|Female||24 (15)||1973 (26)||1.00||-|
|Transgender||2 (1)||34 (1)||-||-|
|Missing||0 (0)||0 (0)|| || |
|Age in years|
|<25||13 (8)||1344 (18)||1.00||-|
|25 to 29||18 (11)||1710 (23)||0.72 (0.32-1.62)||-|
|30 to 34||44 (27)||1563 (21)||2.38 (1.21-4.67)||-|
|35 or more||87 (54)||2845 (38)||3.12 (1.67-5.82)||<0.001*|
|Main drug injected, last month|
|Heroin||55 (34)||4053 (54)||1.00||-|
|Other opiates||4 (2)||239 (3)||1.34 (0.47-3.84)||0.582|
|Psychostimulants||76 (47)||2231 (30)||2.02 (1.38-2.96)||<0.001|
|Other||1 (1)||76 (1)||-||-|
|Missing||17 (10)||847 (11)||-||-|
|User of local IDU services|
|No||92 (57)||5193 (70)||1.00||-|
|Yes||70 (43)||2269 (30)||1.56 (1.07-2.27)||0.025|
|Imprisoned, last 12 months|
|No||133 (82)||5636 (76)||1.00|| |
|Yes||25 (15)||1781 (24)||0.76 (0.47-1.24)||0.275|
|Missing||4 (2)||45 (1)||-||-|
|Heterosexual||75 (1)||6104 (99)||1.00||-|
|Gay/lesbian**||52 (16)||269 (84)||20.43 (13.21-31.59)||<0.001|
|Bisexual||25 (4)||606 (96)||5.30 (3.15-8.93)||<0.001|
|Unspecified||10 (2)||483 (98)||-||-|
|Sex work, last month|
|No||142 (88)||6784 (91)||1.00||-|
|Yes||20 (12)||578 (8)||1.77 (0.94-3.32)||0.078|
|Missing||0 (0)||100 (1)||-||-|
|Years of injecting drug use|
|>12 years||93 (57)||3599 (48)||1.00||-|
|7 to 12 years||44 (27)||2088 (28)||1.27 (0.78-2.07)||-|
|1 to 6 years||25 (15)||1715 (23)||0.87 (0.48-1.58)||0.616*|
|<1 years||0 (0)||60 (1)||-||-|
Sexual orientation and psychostimulant use
A total of 195 homosexual males were in the sample, of whom 23% (n=45) self-reported being HIV positive. This group of HIV positive homosexual males were significantly more likely to report mainly injecting psychostimulants versus other drugs (OR 3.50; CI 1.63-7.54; p-value=0.001) than HIV negative homosexual males. Among the 4,759 heterosexual males, only 1% self-reported as HIV positive and among this group, there was no significant difference in those who reported mainly injected psychostimulants versus other drugs (OR 1.46; CI 0.83-2.58; p-value = 0.194).
This study has one of the largest samples of IDUs surveyed internationally. We found that the large majority (94%) had been tested for HIV, most of whom had been tested in the 12-months prior to first use of the Sydney MSIC. There is limited international evidence to contextualise these results; however, one cross-sectional survey of people at risk of HIV in the US in the mid 1990s showed that 82% IDUs recruited through street outreach had been tested for HIV.9 More recently, 83% of IDUs in contact with urban drug services self-reported ever having had an HIV test and women were more likely to have ever been tested men.10 Australian NSP Survey's data for 2003 to 2007 also shows high testing levels with 88-89% of IDUs having ever been tested for HIV. As revealed in our multivariate analysis, recent imprisonment is associated with higher testing levels and 23% of MSIC clients had been imprisoned in the past 12 months, compared to 13-18% in the NSP sample.1 In line with clinical protocols for drug treatment, a history of drug treatment was found to be significantly associated with HIV testing. Our sample had somewhat higher levels of reported HIV testing than IDUs in other Australian samples, suggesting uptake of HIV screening is reaching even this marginalised population. This high level of testing points to an effective testing policy in Australia and also to high numbers of IDUs having received the pre- and post test counselling for HIV and the associated risk awareness messages.
We found 2% of IDUs self-reported as HIV positive, consistent with surveillance results (i.e. 0.9% to 1.5% from 2003-2007)1 and the strongest predictor was being homosexual, while factors with moderate predictive value were bisexuality, male gender and the main use of psychostimulants versus heroin. These findings confirm what has been known to increase vulnerability to HIV transmission in Australia and developed countries.11,12 The prevalence among those surveyed supports the assumption that Australia has successfully averted an epidemic among heterosexual IDUs.
While the prevalence of HIV among the whole sample are low, the prevalence in the homosexual male group of IDUs attending the service is concerning at 23%, which is a similar percentage to that found in the general injecting drug using population globally.2 We found that this group of men was more likely to report mainly psychostimulants injecting than other drug injecting. Also concerning was the very low levels of antiretroviral treatment (6%) reported by those in the sample who reported HIV positive sero-status, a finding which warrants future investigation as low treatment uptake among this high-risk group has broad public health and individual health implications.
One of the strengths of our study is the size of the data set, while the limitations of the study include the limitation of self-report data. The opportunity to validate self-report data in most research contexts is limited and difficult, and it should be noted that IDU self-report data were collected by Sydney MSIC staff and as such may be subject to social desirability biases, along with biases associated with the recall of events over time which may have affected the results of the regression analysis presented. The validation of self-report data provided by IDUs has been attempted with the general, although not universal, consensus that in the reporting of risk behaviours IDUs are reliable.13
The trial of a supervised injecting facility followed Australia's early and effective response to HIV via needle and syringe programs and the high level of HIV testing reported here points to an effective testing policy. The associations between socio-demographic characteristics and self-reported HIV status in this large, cross-sectional sample are consistent with other data indicating that Australia has successfully averted an epidemic of HIV among heterosexual IDUs. It is important, and an ongoing public health challenge, to maintain the historical levels of protection from HIV among this highly vulnerable group. The absence of any significant associations between self-reported HIV positive sero-status and a priori injecting risk behaviours that increase vulnerability to BBV transmission suggests that HIV infection in this group is likely to be related to sexual behaviours, while acknowledging the complex nexus between sexuality, drug use and HIV status. Also, the strong associations between homosexual male psychostimulant injectors and HIV positive sero-status suggests that patterns of HIV infection within this group reflect the epidemiology of HIV in Australia more generally, where men who have sex with men remain most vulnerable to infection.
The authors would like to acknowledge the staff and clients of the Sydney MSIC and thank Dr Libby Topp and NCHECR PhD Group for their input and advice. NCHECR is funded by the Australian Government Department of Health & Ageing and is affiliated with the Faculty of Medicine, The University of New South Wales.