Proceedings--Workshop on Needle Exchange and Bleach Distribution Programs (1994)

ANNEKE VAN DEN HOEK and ROEL COUTINHO Municipal Health Service, Department of Public Health and Environment, The Netherlands

Through June 1993 a cumulative total of 2678 cases of AIDS have been reported in the Netherlands (circa 15 million inhabitants). Homosexual men are the most important risk group (78%), followed by injecting drug users (9%); 93% of the cumulative AIDS cases are men. In 1992 481 new cases were diagnosed and in 1991 437. Most of the AIDS cases in the Netherlands were reported from Amsterdam (700,000 inhabitants).

The total number of HIV infected persons in the Netherlands is estimated at 6,000-10,000. In Amsterdam the total number of homosexual men between 18 and 55 is estimated at 20,000 of whom 2,000-4,000 are infected with HIV. The number of drug users in the city is estimated at 7,000 of whom approximately 800 are HIV infected.

The estimate of the number of hard drug users in Amsterdam is based on a capture-recapture method and is a year prevalence. The estimated number of drug users staying on a regular day in Amsterdam is lower, approximately 5,500. This smaller number is due to the large number of foreign drug users who only stay briefly in Amsterdam.

Based on data of participants of the low threshold methadone programs, it is estimated that about 40% of the drug users in Amsterdam inject their drugs. The prevalence of the current injection of drugs among drug users differs according to country of origin: circa 40% of the Dutch drug users inject their drugs, compared to circa 70% of drug users of foreign origin (mainly German and South-European) and circa 5% of the ethnic drug users (from Surinam, the Netherlands Antilles, Morocco, and Turkey).

The assistance system for drug users in Amsterdam can be described in three phases: getting in contact, harm reduction and treatment.

Contact with drug users is made by 1) street corner workers, 2) physicians visiting drug users arrested in police-cells and 3) social nurses visiting all hospitalized drugs patients.

Through regular contact appropriate medical and social care can be given, which is considered "to be beneficial for drug users themselves and the society at large". This policy is called the harm reduction approach.

The main instrument for harm reduction (as long as the drug user is not able or willing to stop his/her drug use) is the large methadone program with a low level of threshold.

Another activity of the harm reduction approach is the needle and syringe exchange program, aimed at the reduction of the harm by injecting. This program was initially started in 1984-through an initiative of the drug users organization, the "Junkiebond"-to prevent hepatitis B, but was soon overshadowed by the more important goal of AIDS prevention. In 1985, 100,000 needles and syringes were handed out and this number has gradually risen to circa 700,000 in 1988 and to approximately one million in 1991 and 1992. In 1992, 92% of the distributed needles/syringes had been exchanged for a used needle/syringe. Presently Amsterdam has 14 needle exchange locations. It is possible to exchange needles and syringes from 10 a.m. till 4 a.m. the next day. During the night, two slot machines are in operation for purchasing syringes. Participation in the exchange program does not require identification or registration. For this reason, no information is available on the number of participants or on their demographic characteristics.

As the needle/syringe exchange program is a low threshold project, there is no registration or monitoring of clients. Evaluation of the impact of the exchange program on injecting behavior and the spread of HIV, has therefore mainly taken place in our cohort study on HIV infection and AIDS.

The open cohort study started at the end of 1985. At that time only one drug user with AIDS had been reported in the Netherlands.

The aims of the study are

1. to study the prevalence and incidence of HIV infection and AIDS in relation to (changes in) drug use and sexual behavior;

2. to evaluate the impact of various HIV-prevention programs for drug users;

3. to study determinants of risky injecting and sexual behavior; and

4. to study the natural history of HIV infection.

Participants are recruited at methadone outposts, the special STD clinic for drug using prostitutes and by word of mouth. Eligible for the study are men and women who use or have used drugs, either by injection or otherwise. Blood samples for serology, virology and immunology are taken and participants are interviewed using a standard questionnaire which includes questions concerning clinical symptoms, medical history, lifestyle, use of oral and intravenous drugs (methadone included), and

prostitution. Participants are asked to return for a follow-up visit every four months. Twenty five Dutch guilders are paid per follow-up visit to encourage continued participation.

Through July 1993 a total of 1,012 drug users had entered the study, 258 HIV positives and 754 HIV negatives. The HIV prevalence among drug users with a history of injecting drug use was approximately 30% (¹) and remained more or less stable among new intakes in this group in following years (²). The annual HIV incidence per 100 person-years was 9.2 in 1986, varied between 2 and 5% in the years 1987-1991 (³) and was 2.5 in 1992. To date a total number of 52 seroconversions have occurred.

The first study on risk reduction among the participants (December 1985-April 1988) showed that during follow-up, a strong reduction in borrowing and lending occurred, and that this behavioral was not dependent on being informed of HIV serostatus (⁴). Over time, the use of the needle and syringe exchange program increased. However, reduction in needle sharing was not seen among new entrants to the study. Therefore, we concluded that the risk reduction observed during follow-up was mainly an effect of the study (with counselling), with the exchange program only having a limited effect.

The next study (⁵) looked into factors related to regular participation in the exchange program and the borrowing of syringes in 131 HIV seronegative current injecting drug users (1989-1990). A total of 29% of the users reported borrowing syringes in the past 4-6 months. Users at increased risk of borrowing are previous borrowers, long term moderate-to-heavy alcohol users, current cocaine injectors, and drug users without permanent housing. Regular clients of the syringe exchange, when compared with other injecting drug users, were found more often to be frequent, long term injectors. They borrowed slightly less often than other users, but this was not statistically significant, even after controlling for frequency of injecting or other potential confounders. These results suggest that 5 years after the start of the exchange program, drug use characteristics govern an individual injecting drug user's choice of exchanging or not exchanging. We concluded that it seems more important to direct additional preventive measures at injecting drug users with an increased risk of borrowing rather than at users who do not participate in the syringe exchange or who do so irregularly.

Another study (³) assessed risk factors for seroconversion to HIV, between December 1985 and November 1991. The behaviors of 31 seroconverters were compared with those of 202 seronegative injecting drug users (controls). Three

independent risk factors for seroconversion were found in logistic regression: 1) living > 10 years in Amsterdam (OR=2.45, 95%CI 1.09-5.53); 2) first injection < 2 years ago (OR=3.43, 95%CI 1.20-9.81); and 3) injecting mainly at home (OR=0.39, 95%CI 0.18-0.88). No evidence was found that obtaining new needles/syringes via the exchange program was protective. However, the data suggest that exchanging needles/syringes may have been protective at the start of this program. In the discussion of this finding we mentioned that this may be explained by an overall increased availability of needles/syringes, which enabled non-exchangers to more easily obtain new needles/syringes. Another explanation we mentioned was that, at the beginning of the program, a desire for risk reduction was the motive for exchanging, while later on exchanging became just a way to obtain injection equipment.

The methodological problems encountered in evaluating prevention programs are many. In general, little is known about the representativeness of the study sample of drug users. Furthermore, participants are self selected, and self-selection occurs again with respect to participation in the follow-up study. Self-reports on injecting and sexual behavior may be unreliable and are difficult to validate.

To evaluate the impact of prevention-programs, random allocation of drug users to the various programs would be the best study design. However, this allocation would be in conflict with the harm reduction policy which includes large accessibility of the programs for all drug users. Another problem in evaluating the impact of the programs on risk reduction is that drug users may attend programs for other reasons than risk reduction and the longer low threshold programs exist, the more this may be the case. On the other hand, health education messages have also reached drug users who do not want to use the needle and syringe exchange program to obtain clean needles and syringes and prefer to buy their needles and syringes at pharmacies and certain shops. These considerations may imply that the impact of a prevention-program cannot be assessed by studying differences in risk behavior between attenders and non-attenders.

Indeed, a last study (⁶) that studied serial, cross-sectional trends in injecting behavior from 1986 to 1992 showed that the proportion of drug users who reported borrowing and lending used injection equipment and re-using needles/syringes (in the 6 months preceding intake) continuously declined from 51% to 20%, from 46% to 10% and from 63 to 39%, respectively and that non-attenders of exchange programs reduced their risk behavior to the same extent as attenders. This finding explains why, in comparing attenders with non-attenders we were not able to demonstrate any impact of the exchange (and other prevention) programs on risk reduction*.

We conclude, therefore, that the evaluation of specific measures is difficult. Although we have not been able to demonstrate any impact of specific prevention measures, we think that all prevention activities taken together in Amsterdam (exchange programs, over-the-counter sales of needles/syringes by pharmacies, low threshold methadone programs, counselling projects, and information campaigns) have been responsible for the decline in high-risk injecting behavior.

However, it must be realised that a considerable number of drug users from time to time borrow an used needle/syringe and that transmission of HIV among drug users still occurs.

MARGARET MILLSON, Department of Preventive Medicine and Biostatistics, University of Toronto, Canada, and City of Toronto Department of Public Health; and CATHERINE HANKINS Centre for AIDS Studies, Montréal-Centre Regional Public Health Team, Canada

The first official HIV prevention program for injection drug users in Canada which included needle exchange opened in Vancouver in January 1989. It was quickly followed by programs in Montreal, Toronto, and several other major cities, so that by the end of 1990 there were 8 such publicly-funded programs involving needle exchange in operation. These programmes had not seen the light of day spontaneously but rather, with the exception of the Vancouver needle exchange, were part of a federal government strategy to stimulate the development of pilot intervention programmes for injection drug users.

The overall objectives for this initiative on the part of Health Canada were twofold. First, the federal government aimed to collaborate with provincial governments in the support of pilot prevention programmes designed to reduce the transmission of HIV among injection drug users and their sexual partners. The second objective was to acquire national data concerning the risk of HIV infection among injection drug users and the efficacy of prevention strategies.

The funding criteria for the pilot prevention programmes were clearly spelled out. In order to be eligible, pilot studies had to take place within the context of a collaborative multidisciplinary network involving local public health officials, addiction treatment agencies, law enforcement, community groups and, where possible, academic or research institutions. A variety of collaboration mechanisms were considered to be acceptable, ranging from inter-agency committees to direct multi-agency programme delivery.

In addition to this multidisciplinary context, pilot studies had to embody a multifaceted approach to prevention.¹ Proposed programs had to include risk reduction education and counselling, as well as linkage to addiction treatment services and to other existing health and social services. Prevention programs could opt to include a carefully monitored needle-syringe exchange component.

With respect to evaluation, in order to be eligible for funding, pilot studies were required to include a comprehensive and methodologically sound as well as ethically approved evaluation component. The evaluation research protocols had to receive the approval of a constituted ethics review committee and follow the Medical Research Council of Canada's guidelines on research involving human subjects, 1987.²

With respect to the funding strategy, the federal government of Canada offered to cost-share the service delivery component on a 50-50 basis with the provinces during the pilot study period, up to a maximum of two years. The federal share of the programme funding was completely administered through those provinces agreeing to participate. The federal government bore the full cost of the programme evaluation component during the same pilot period.

What was the outcome of Health Canada's initiative to support pilot intervention programs for injection drug users? Five provinces participated with federal funding for the last of the pilot projects ending in March of 1993. Recognizing the importance of a long term integrated HIV prevention strategy in the injection drug using population, many provincial and local governments not only assumed the responsibility for ongoing funding but proceeded to increase the number of outreach programmes. More than 30 intervention projects are now operating and an additional 10 projects are slated to come on stream by the spring of 1994.

Publically funded programmes involving needle exchange in Montréal, Toronto, Vancouver and several major cities were opened in 1989 and 1990, eight of which form the basis for this national evaluation. Although a series of meetings was held to discuss evaluation of the projects and promote the use of common methodology and instruments in order to seek comparability, there were significant differences in the evaluation approaches utilized, and to some extent these have limited the ability to compare and attempt to generalize findings. This paper will provide a brief overview of key program features, evaluation methodologies employed, their strengths and limitations, and the major findings so far. This overview will be followed by a more detailed description of the evaluations carried out in Montreal and Toronto.

The programs which were established in different cities varied substantially in their organizational features. Although all were required to be multiagency collaborations, some programs were established within agencies already serving the target population, with pre-established credibility with their clients, while others were established as new services which needed to become known and accepted by potential clients. The cities involved also varied in the degree of acceptance by the police, politicians, and the public of the services, particularly of needle exchange. There were also important differences in the size and characteristics of the client population and in the drugs being used. For example, at the outset of these programs, injection of a combination of talwin and ritalin (''T and Rs") was very common in Western Canada, but was almost unheard of in Toronto and Montreal, where heroin and especially cocaine were the drugs of choice. There were police reports of shooting galleries in Montreal, but these were said to be quite uncommon in Toronto. Therefore each city represented a unique set of circumstances which influenced both the program and the evaluation undertaken.

An important difference among different services was the mix of outreach services undertaken. Some services relied heavily on a mobile van to reach potential clients, others conducted considerable street outreach on foot, while others relied primarily on a fixed site in a location judged to be appropriate for attracting potential service users. In

some cities these fixed site(s) were in pre-existing agencies, in others they were newly established locations.

A common feature of all services was the effort to provide a range of services, including counselling and referral to drug treatment. In some sites where the volume of clients seen was high, the ability to carry out in-depth counseling may sometimes have been compromised.

All the evaluations undertaken as part of these projects attempted to collect process measures documenting numbers of clients served, numbers of needles given out and exchange rates, client characteristics including age, sex, drugs used and frequency, length of time injecting, needle-sharing frequency, reasons for needle-sharing, disinfection techniques used when taking used needles and their frequency of use, as well as frequency and gender of sexual partners, condom use, how the client heard of the service, and distance traveled to the service. Evaluators agreed to ask questions in a standardized way as much as possible, based on questionnaires developed for the World Health Organization/European Economic Community collaborative studies, which have been used in several cities, including New York, for research on the epidemiology of HIV and injection drug use.

Several of the evaluations undertaken attempted to measure impact through tracking individuals in some fashion to allow for one or more follow-up comparisons over time looking for reported behaviour change as well as repeated measures of HIV status using saliva +/fingerprick blood samples. Generally speaking these approaches included a comparison group of injection drug users who were not in treatment and not attending needle exchange. Most of the evaluations using this approach experienced difficulty following up individuals, due in part to the policy of allowing participants to remain anonymous, which made repeated participation dependent on the individual to return, or workers/interviewers to recognize previous participants and encourage them to participate again. From the beginning, there was agreement that studies requesting lengthy interviews would need to pay participants; most studies such as the one in Toronto have paid $20 for a personal interview requiring 45 minutes or more; the payment is given to those completing the interview regardless of whether they agree to provide samples for testing. Despite this incentive, this study and some others had difficulty obtaining follow-up data on the same individuals; in the first year of the study, 3 month follow-up rates were only 20%, so that it was necessary to modify the design to focus on repeated cross-sectional measures of the population as a whole rather than individual follow-up. The implications of this will be discussed further below.

In general, researchers who were external to the program and trying to collect impact measures in some cases experienced difficulties in collaborating effectively with project staff who were asked to collect process measures. In some instances, staff considered collecting even minimal information such as age or drug of choice to be too intrusive to clients, and feared that clients would be driven away be being asked any

questions. In many instances this problem was more acute at the beginning when programs were being launched, and became less problematic once client comfort with the service was established.

Table 1 summarizes the findings of a key impact measure, numbers of needles distributed annually by some of the larger programs. It should be noted that each of these cities was estimated to have several thousand injectors at the time these programs were undertaken.

In addition to needles, most programs provided alcohol swabs, sterile water for injection, and condoms; many also provided bleach kits and instructions on needle disinfection for use in situations where sterile needles were not immediately available. Some programs found a very high demand for condoms, in some cases by clients who did not request needles, reflecting the mix of sex trade workers attending services. Another key measure was demand for referral to other services, in particular to drug treatment. Some programs, for example Vancouver, reported that requests for treatment outstripped availability, and pressed for more treatment to be made available. In Toronto, experiences of the needle exchange and other community services in seeking treatment for their clients lead to concern about difficulty in finding treatment slots. The Ministry

of Health responded with development of a registry of available treatment slots to allow counsellers to more readily refer their clients.

Over time, most services reported attempts to improve client access and to reach more clients through approaches such as alteration of hours, street outreach, and use of vans to provide a mobile service. Several cities identified an over-representation of males, especially older, long-term injectors, among their IDU clients, and in some cases also absence of certain ethnic groups known to have injectors. The response was generally to develop special outreach strategies, in some cases involving health, social service, or AIDS prevention organizations dealing with these populations in promoting or delivering services.

Most evaluators found high reported client satisfaction with the services, except in some cases for a desire for different or longer hours of service, a more convenient location, or, in a few cases, a more liberal exchange policy. Programs differed in exchange policy, with some adhering more strictly to 1:1 exchange or to limitations of the total number of needles provided at one visit; clients tended to prefer less strict policy in these matters.

Most programs demonstrated declines in needle sharing and increases in use of bleach to clean needles. Generally the improvements in safe needle use were greater among users of needle exchange than non-users; however in some studies, including the one in Toronto, there is some difficulty with interpreting trends in non-attenders of needle exchange, because trends in the population as a whole may be influenced by the work of the needle exchange, as well as by other services doing HIV education and counselling in the same population, so that the non-attenders are not truly without services, and the impact of the needle exchange in isolation is very difficult to determine. Indeed our own outlook has been that needle exchange should not be considered in isolation, but as part of a multi-service strategy; the reductions seen in risky needle use behaviour in the overall population suggest that the program as a whole is having beneficial impact.

The evidence for reduction in sexual risk behaviour is less clear-cut; although there is some suggestion of increased condom use with casual partners, most programs have not demonstrated statistically significant changes. This is a finding which has been identified elsewhere in the world, and requires further examination and program focus.

With the possible exception of Montreal, which will be discussed further below, most cities in Canada have demonstrated HIV infection rates under 10%; some cities have been able to demonstrate maintenance of steady seroprevalence rates over two or more years.

The programs which began as pilots have now been continued, usually with provincial funding from AIDS prevention budgets, and the provinces of British Columbia and Ontario in particular have funded several more programs in smaller cities; the province of Quebec has made a particular effort to enhance needle availability in pharmacies as a strategy to complement ongoing needle exchange activities. It is highly desirable that ongoing core measures of process and impact be collected in order to assess the longer term results of these programs, and modify them as needed in response to changing circumstances.

The support and assistance of the National Health Research and Development Program and Ms. Betsy MacKenzie, AIDS Education and Prevention Unit, Health Canada, as well as the principal investigators of the evaluations mentioned are gratefully acknowledged.

MARGARET MILLSON, Department of Preventive Medicine and Biostatistics, University of Toronto, and City of Toronto Department of Public Health, Canada TED MYERS, City of Toronto Department of Public Health and Department of Health Administration, University of Toronto, Canada JAMES RANKIN, Department of Preventive Medicine and Biostatistics, University of Toronto, and Addiction Research Foundation, Toronto, Canada WILLIAM MINDRELL, City of Toronto Department of Public Health, Toronto, Canada BERNADETTE MCLAUGHLIN, CAROL MAJOR, and MARGARET FEARON, Ontario Ministry of Health, Toronto, Canada JANET RIGBY and STEFFANIE STRATHDEE, Department of Preventive Medicine and Biostatistics, University of Toronto, Canada; RANDALL COATES, Department of Preventive Medicine and Biostatistics, University of Toronto, Canada (deceased)

From October of 1989 to October 1990, we conducted anonymous interviews with out-of-treatment injection drug users in Toronto as part of an evaluation of the City of Toronto Department of Public Health's HIV prevention program for injection drug users. A key element of this program was the needle exchange called "The Works." In May 1991, we began another year of interviews with out of treatment IDUs recruited in the same sites as in the first year of study; this research represented the first year of a three year study in collaboration with the WHO multicentre study of HIV and injecting drug use. This drug user research has been funded by the National Health Research and Development Program of Health Canada.

This report, therefore, represents a comparison of the results of these repeated cross-sectional surveys conducted over two and a half years among injection drug users who had injected within 2 months of the time of interview, and who had not received treatment for drug problems within the preceding 3 months. Subjects were recruited from agencies serving IDUS in the downtown area of Toronto in both years; in the second year, recruitment was expanded in one particular area which is well known for prostitution and drug use. Recruitment was through posters and word of mouth, and subjects were paid $20 for the interview. HIV testing was conducted in an anonymous fashion using saliva and finger prick blood samples; results were not provided to interviewees, who were instead encouraged to seek testing with counselling at appropriate clinics.

Fewer young subjects were interviewed in the second year; this was attributed to a combination of another study going on among street youth during part of our study period in the second phase, as well as a reported drop in injecting as a mode of drug taking among some young drug users. As a result of this age variation, we controlled for age in the analyses reported.

In both study periods the majority of interviewees were male, but there was a small (non-significant) increase in the proportion of females in the second year.

The drug of choice most commonly reported in both years was cocaine, but there was an increase in the proportion reporting heroin as their drug of choice in year 2.

Of particular note in our study population is the high rate of reported incarceration. In both years, about 80% reported having been in jail at least once since they began to use drugs. We also documented reported needle sharing while in custody.

1. There was a statistically significant (p<.05) move away from sharing needles, for both giving and receiving used needles, in this population between 1989 and 1991, such that in the second year of the study, 65% of IDUs reported never using someone else's needles in the preceding 6 months; 68% reported never giving someone else their used needles. A small but fairly stable percentage of 10-14% reported sharing at least weekly in the preceding 6 months in both years, with the remainder reporting infrequent sharing (i.e. monthly or less).

2. There was also an increase in cleaning of used needles with bleach reported by year two, with 75% of all those who took used needles reporting some use of this cleaning method, up considerably from the 51% reported for the 1989 sample.

3. Sources of new, sterile needles also shifted between the two study periods, with needle exchange being reported as the most important source by 46% in year two as compared with 18% in year one, with a drop occurring between the years in the proportion reporting pharmacies or friends as their most important source. This was the case even though approximately equal proportions of those interviewed mentioned pharmacies and needle exchange as one source they used (69% and 63% respectively). This may reflect the degree of access perceived to the two sources by IDUs. It should be pointed out that by the second year of study, the City of Toronto Dept. of Public Health had begun funding needle exchange in some of the community agencies in which we were doing recruitment, in addition to its own service, so these figures may reflect to some extent an increase in access within these agencies.

1. For males, consistent use of condoms with casual partners did rise from 30 to 40%, mainly associated with a drop in those reporting never using condoms with casual partners. However, for those with a regular partner, the proportion always using condoms dropped slightly over the two years, with a corresponding increase in those who never used condoms. This finding requires further detailed analysis to attempt to determine the factors involved in this behaviour.

2. For females, although the numbers are smaller, the trend with regular partners is similar to that for males; reporting of consistent use of condoms with casual partners did increase, but the numbers were small and this was not statistically significant.

Self-report By the second year, nearly three-quarters of our interviewees reported at least one previous HIV test; 5.2% of those tested reported that they were positive.

Anonymous, Unlinked Testing We have found very good compliance with being tested in our studies, such that in year two, only 1.5% of those interviewed failed to provide any specimen for testing. There has also been a drop in the proportion providing only saliva or only blood, although 5% continue to refuse to provide a dried blood spot. We have found saliva to be safe, accurate, and convenient as a tool for these seroprevalence studies in the field, provided it is tested by a lab able to carry out the techniques required.

The results of testing showed a small increase in seropositivity in the year two sample when compared to year one, from 4.3% in year one to 5.7% in year two, however the confidence limits for the second year are 2-10%, so there is clear overlap between the two years, and no statistically significant increase. We have completed testing on specimens collected during our third year of study, ending in May 1993, and although we have not completed detailed analysis, the crude seroprevalence rate appears to be approximately the same as in years one and two.

We conclude that there has been significant decline in needle use related risk behaviour in this population, at the same time that ongoing HIV prevention efforts have been introduced and expanded. There is also some modest improvement in condom use with casual partners, although not with regular partners. These patterns and their significance require further exploration.

Seroprevalence in this population is high enough to be worrisome, and mandates ongoing intensive prevention efforts. We are encouraged by the finding that there is no

increase apparent in seroprevalence in our ongoing studies; we intend to continue to monitor the situation as our city's prevention program continues to be active.

The accompanying copy of the poster presented at the IXth International AIDS Conference in Berlin illustrates the findings presented here.

CATHERINE HANKINS and SYLVIE GENDRON, Centre for AIDS Studies, Montréal-Centre Regional Public Health Team, Canada; JULIE BRUNEAU, Detoxification Unit, St.-Luc Hospital, Montréal, Canada; and Élise Roy, Centre for AIDS Studies, Montréal-Centre Regional Public Health Team, Canada

Several criteria can theoretically be used to evaluate the success of needle exchange programmes. Among them are measures of awareness of the service in the target population of injection drug users (IDU), utilization rates, and satisfaction with the overall service among the clientele. Changes in injecting behaviour can be documented cross-sectionally or, better yet, followed over time in a cohort of IDU to provide a measure of the behavioural impact of prevention-education activities in the injecting drug using population. Once a baseline HIV prevalence proportions has been established, it can be followed prospectively to assess the stability of prevalence and document any increases or declines. The gold standard criterion for success, which has appears to have achieved general consensus, is the estimation of baseline HIV incidence with subsequent documented declines in incidence among injection drug users who are attending a needle exchange programme regularly.

From a more general perspective, the implementation and maintenance of needle exchange activities should ideally have positive influences on policy and programme development, both at the local and national level. Since the opening of CACTUS-Montréal (Centre dAction Communautaire auprès des Toxicomanes Utilisateurs de Seringues), there have been eight other projects initiated across the province of Québec in both urban and semi-urban settings. Among them is a project for pharmacy exchange currently in the pilot phase in Montréal and two projects involving community clinic and hospital-based exchange services in semi-urban areas. Finally, needle exchange programmes are perceived to have the potential to increase demand for methadone maintenance, detoxification services and rehabilitation programmes and therefore should result in expansion of these services although such an expansion has not been documented thus far in Québec.

The CACTUS-Montréal fixed site is located at 1209 Saint-Dominique Street in downtown Montréal. It has been operating seven nights a week from 21:15 to 04:00 since July 9, 1989. Clients enter the site directly from the street. There is no exterior sign identifying the building as a needle-exchange site other than several cactus plants in the window. Two teams, each composed of a male and female nurse, alternate to provide seven day coverage. In addition to disseminating information on AIDS and the prevention of HIV transmission, the nurses distribute condoms, alcohol swabs, lubricant, gloves and travel size bottles containing bleach or distilled water. They also exchange

sterile needles and syringes for used ones. First time clients receive a pamphlet which demonstrates how to clean needles and how to use condoms. Additional print materials regarding AIDS and other STD as well as business cards which indicate resources and the range of treatment options available in Montréal for drug users are available on site. Anonymous testing for HIV antibodies accompanied by pre and post-test counselling is also offered at scheduled times in a closed private room.

The exchange policy is a one-for-one exchange to a maximum of 15 needles with one extra needle always provided. Individuals who arrive with no needle are provided with one needle with no examination for needle tracks. The exchange rate has averaged a relatively stable rate of 75-80% over the period 1991 to 1993. The total number of visits per week stabilized by one year of operation at approximately 1200 visits per week.

Two components of the CACTUS-Montréal evaluation will be discussed in this paper. The first concerns the findings of the awareness, utilization, and satisfaction study conducted in the first two years of operation and the second concerns data from the prevalence-incidence study underway at the CACTUS-Montréal fixed site.

To determine the degree of awareness, utilisation, and satisfaction with the services offered at the exchange's fixed site, injection drug using CACTUS-Montréal attenders were approached in three correctional medium security institutions and at a detox unit drop-in clinic for individual interviews of five to ten minutes duration¹. Injection drug users from the correctional setting were study volunteers for an ongoing research project on risk factors for HIV infection among inmates in medium security correctional institutions². This study involved a standardized nurse-administered interview, as well as personalized counselling and HIV antibody testing³. At the drop-in clinic the first five IDU presenting to the clinic per week were asked to participate. A preliminary plan to administer the questionnaire at the fixed site was abandoned because the physical setting was inappropriate and because the validity of data collected in the presence of service providers was questionable.

Most information (70%) concerning levels of awareness, utilisation, and satisfaction with the services offered at the CACTUS-Montréal fixed site was obtained from incarcerated IDU, the remainder being provided by IDU attending the detox unit drop-in clinic near the fixed site. Comparing the prison subjects with the detox clinic clients, a greater proportion of women (p=.001), exclusive cocaine users (p=.0005), and people who had attended CACTUS-Montréal more than 15 times (p=.03) were represented among prison subjects. Overall, seventy-three per cent (73%) of those interviewed had heard about CACTUS-Montréal and respondents readily identified that needle-exchange services and free condoms were available at CACTUS-Montréal. Publicity for the fixed site mainly occurred by word of mouth, with 57% of aware respondents having heard about the programme from another IDU. Fifty-five per cent (55%) of those who were aware of the existence of CACTUS-Montréal had visited the fixed site, and 76% of these had actually been to CACTUS-Montréal more than five

times. Attenders generally expressed satisfaction with the fixed site personnel (88% liked the staff) and 98% of attending IDU indicated that they had received the services they asked for when they visited the site. Location and opening hours did not appear to suit everyone however. When asked what they liked about CACTUS-Montréal, the most frequent responses were ''the staff" and "it's free." None of the respondents suggested that CACTUS-Montréal dispense or exchange more needles than current quotas.

Within the context of a seroepidemiological study conducted in the three correctional institutions, an analysis of seroprevalence levels in injection drug users by attender or non-attender status was conducted. One hundred and sixty-three (163) men and 130 women participated in the study. Among male injection drug users, 8.6% (14/163) were seropositive with a marked discrepancy noted between rates in attenders and non-attenders. Among male attenders of CACTUS-Montréal, 20.5% (8/29) were positive versus 4.8% (6/124) of non-attenders (p=0.002). This contrast was not observed among the 130 women, 15 of whom were seropositive (11.5%). The rate among female attenders was 11.6% (8/69) and among non-attenders it was 11.5% (7/61). Comparing all attenders and non-attenders, attenders were twice as likely to be HIV-positive with 14.8% of attenders (16/108) and 7% of non-attenders (13/185) found to be infected (p=0.05). With respect to the behavioural profiles of male inmates, attenders (n=39) and non-attenders (n = 126) were equally likely in the six months prior to incarceration to have injected with a syringe from a dealer (p=0.24) and to have given, shared or sold a used syringe to close friends (p= 0.23) or to people in shooting galleries (p=0.97). In prison, attenders and non-attenders were equally likely to have injected drugs (p= 0.35). However, attenders were significantly more likely to have had bleach in their possession (p < 0.0005) and to have injected with a needle from someone they did not know well (p=0.001). This preliminary analysis suggests that CACTUS-Montréal may be attracting a particularly high risk population, at least with respect to male participants.

Volunteers were enlisted at the fixed site of CACTUS-Montréal to provide anonymous samples for HIV antibody detection. On one randomly chosen evening per week during a three-hour period IDU were approached as they were leaving the site and asked to provide a blood specimen obtained via fingerprick or a saliva specimen collected in a sputum collector or Omni-sal. The sensitivity of saliva testing using the dried blood spot testing as a gold standard was 92.7% with a specificity of 100%⁴. Since no false positive results were detected, data from individuals with only saliva specimens were combined with data from individuals with dried blood spot specimen results for analysis. Filter paper and saliva specimens were identified by a random bar code number, ensuring complete anonymity for all samples at the laboratory. This code was also affixed to a service delivery/behavioural form to permit the linkage of serology results to information concerning behaviours in the previous 7 days as well as to basic information on age and sex found in the CACTUS-Montréal client code. Such a link allowed for the characterization of participants and refusers and served to identify repeat

samples for the estimation of overall seroprevalence and seroincidence. No information was collected which could be used to personally identify any one individual. Participants who desired knowledge of their HIV status were referred to clinical testing services at CACTUS-Montréal or at anonymous testing sites.

Overall, 25% of individuals who were approached agreed to provide a specimen. Most refusers seemed to refuse because they were in a hurry. Seroprevalence data were culled by year so that each person was represented only once per twelve month period. Individuals who had never obtained or returned needles and who had never responded to questions concerning injection drug use practices were excluded from the analysis.

Analyses for the period January 1990 to January 1993 yielded the following seroprevalence proportions:

Using the CACTUS-Montréal client code, repeat tests on the same individual were identified to assess seroconversions and to calculate estimates of incidence for the same 36 month period. The date of seroconversion was calculated as the mid-point between the last negative and first positive result. Incidence was calculated as a proportion (%) and as a rate in person-years (incidence density).

In 36,805 person-days of observation, 13 of 136 (9.6%) individuals, for whom at least two test results were available, seroconverted. The overall incidence rate was 12.9 per 100 person-years of observation (95% CI: 6.9; 22.0). Preliminary analyses revealed that factors such as age, gender, cocaine use, condom use, and the sexual orientation of males were not significantly associated with seroconversion. Among those who had borrowed needles in the previous seven days, the incidence rate was 18.7 per 100 person-years versus 2.5 for those who had not (p =.02). Likewise, among those who had loaned needles in the previous seven days, the incidence rate was 20.3 per 100 person-years versus 2.2 for those who had not (p=.01). Further analyses are being conducted to examine these and other sociodemographic and behavioral factors which may be associated with seronconversion and higher incidence rates.

The challenge in evaluating any prevention programme lies in the attempt to determine direct programme effect by measuring the gap between what would have likely happened had there been no programme and what actually has occurred in terms of behaviour change and seroprevalence/seroincidence. With respect to CACTUS-Montréal, this challenge is heightened by the finding that the programme, which has a greater than 2 to 1 ratio of male to female attenders, appears to be attracting a higher risk male clientele. This may account for the fact that prevalence has not yet completely stabilized in this population and that incidence rates remain unacceptably high. However, findings from behavioural studies at CACTUS-Montréal⁵ have revealed that the loaning of needles has declined from 31% to 20% since the beginning of operation and that 62% of those who inject with a used needle do so after having cleaned with bleach, in combination with another method or alone, compared with 30% in the first two months of operation.⁶ This suggests that this core group of injection drug users may possibly be contributing less now to HIV transmission among injection drug users that they were two years ago. The fact that these individuals are attracted to the site and are participating in risk reduction activities should be viewed as having positive implications for the eventual reduction of HIV transmission in Montréal's injection drug using community.

This study was supported by grant #6605-3463-AIDS from the National Health Research and Development Program, Department of Health, Canada.

ALEX WODAK, Alcohol and Drug Service, St. Vincent's Hospital, NSW, Australia

There is a large and consistent body of HIV seroprevalence data from a variety of sources which indicates that HIV infection is still relatively uncommon (< 5 %) in Australian injecting drug users (IDUs). AIDS data are consistent with this. The epidemic thus appears to be well under control at present among IDUs. HIV infection is, however, prevalent among homosexual/bisexual males. Substantial overlap between these populations provides the preconditions for potential rapid spread of infection. National estimates of the number of IDUs range from 100,000 to 200,000.

The course of the HIV epidemic in Australia differs from that of many other western countries as an IDU component has not yet occurred. HIV infection now appears to be spreading slowly with the number of new cases of HIV infection nationally estimated to be about 600 per year.

AIDS was identified as a priority issue soon after the epidemic was first recognised. The role of prevention was given particular emphasis and there was no disagreement about the critical role of IDUs in the epidemic. It was accepted in the early 1980s that it would be necessary to implement a range of strategies including some sensitive measures to maintain control of HIV among IDUs in order to contain the epidemic in the general community. Consequently, prevention strategies relevant to IDUs were identified early, adopted with broad support and implemented vigorously. IDUs (and other high-risk groups) were involved in the identification and implementation of HIV prevention policies. Adoption of HIV prevention measures within prisons has however, been slow, difficult and incomplete. A national drug policy of minimisation of harm was declared in 1985 at a meeting of senior political leaders. This policy facilitated the adoption and later implementation of pragmatic HIV prevention strategies.

Sterile injecting equipment has been readily available in all major cities and large towns since 1988/1989 and in critical areas beginning in 1986. Drug store sales and needle and syringe exchange programmes (NSEPs) have played a major role in increasing the availability of injection equipment. Drug stores and NSEPs service different populations of IDUs. Several years were required for most needle and syringe exchange outlets to reach maximum throughput and begin to achieve efficiencies. Unit cost of exchange and distribution appears to be decreasing. After six years of NSEP, unintended negative consequences have been relatively minor and consisted of littering of public places with discarded equipment and two detected uses of NSEP facilities for drug dealing. Littering was quickly overcome by a series of measures including special collection arrangements, disposal bins and the marketing of specially developed plastic boxes ("Fitpack") which contain sterile equipment and a tamper resistant device for

retaining used equipment. Approximately three to four million needles and syringes are distributed in Australia each year. Availability is still being expanded.

Explicit education campaigns directed to IDUs began in 1987. Expansion and improvement of drug treatment began in 1985. National methadone capacity increased more than six times between 1985 and 1993 and is now over 13,000. The national methadone guidelines have been revised and liberalised several times in recent years. Community bleach distribution programmes began in 1988. However, demand for bleach was never impressive. Decontamination practices with bleach were often deficient. Consequently, even if there was good evidence that bleach effectively destroyed HIV in vitro, there would be some doubt about the likelihood of this benefit being achieved in vivo because decontamination procedures are often deficient. Organisations of drug users began to be formed in 1985 but were established with government support in all jurisdictions by 1988. Research has played an important role in defining and evaluating strategies.

There is some evidence that high risk injecting practices are becoming less common in Australian IDUs. As in other countries, baseline levels of unsafe injecting practices were very high and unsafe sexual practices appear to have declined only slightly. HIV prevention policies are widely regarded in the community, among politicians and health policy makers as having successfully prevented spread of HIV among IDUs although rigorous proof of the effectiveness of strategies is unavailable, probably unobtainable and was wisely never regarded as a prerequisite for adoption and expansion of prevention programmes. HIV prevention policies for IDUs continues to have very strong community and political support. Community support for NSEP and methadone has been shown in a telephone poll to be 80-90%. Media coverage for AIDS and HIV prevention strategies for IDUs is (with very rare exceptions) understanding and supportive. Bleach programmes are still supported although their future appears somewhat uncertain following recent and accumulating evidence of lack of efficacy.

There is recent concern that the magnitude and consequences of IDU-related epidemics of other blood borne viruses including Hepatitis C (HCV) have been seriously underestimated. Continuing high incidence levels of HCV, especially among young IDUs, continuing high levels of risk behaviour and international spread of HIV among (and from) IDUs in an increasing number of countries suggest that complacency about the possibility of future spread of HIV related to IDU in Australia is unwarranted. There is some recent evidence that HIV spread in prisons has also been underestimated both in terms of documented seroconversions and public health impact. Evidence has emerged recently of an unofficial needle and syringe exchange programme successfully conducted for almost a year in several prisons in NSW.

The Australian experience with HIV prevention among IDUs has stimulated a more critical appraisal of prohibition. There is increasing support for the view that prohibition is expensive, ineffective, counter-productive and impairs the effectiveness of efforts to control the spread of HIV.

These conclusions may be of interest to other countries—especially those which share many similar characteristics such as the United States. However, the many major

differences between Australia and the United States must also be borne in mind when attempting to extrapolate from the Australian experience to North America.

HIV/AIDS presents a major challenge to the international community. There is much to be learnt from both positive and negative experiences of other countries. This paper describes the current state of the HIV epidemic among IDUs in Australia. The Australian response to the HIV epidemic has been deservedly praised. Favourable comments have been made by international experts as well as local, independent, public health practitioners. The aim of this paper is to present the Australian experience to an American audience to assist efforts to slow the spread of HIV infection among and from IDUs in the United States. Comparisons between the two countries have been drawn even if these may appear to be unflattering to the nation generously hosting this meeting. The temptation to echo anodyne and diplomatic platitudes has been resisted. It is hoped that critical comments will be accepted as being offered in good faith and in the interest of protecting public health.

There are at least two very important connections between Australia and the United States that are relevant to this paper. All of the earliest cases since the first case of AIDS was diagnosed in Australia in 1982 were homosexual/bisexual males who had lived for some time in the United States. Also, New York was the first city in the world to attribute a large proportion of AIDS cases to IDUs. From an Australian perspective, the response to this challenge seemed curiously minimal. The notion of "preventing another New York" occurring in an Australian city was often discussed when responses to the threat of an IDU related epidemic in Australia were being developed. Australia learnt from the negative US experience to the extent that Sydney has been included in a study of "prevented" HIV epidemics among IDUs together with Glasgow (Scotland), Lund (Sweden), and Tacoma (Washington State) (Des Jarlais, submitted for publication).

There is a large and consistent body of evidence which indicates that HIV infection is still relatively uncommon in Australian IDUs. These data include published and unpublished HIV seroprevalence surveys of IDUs. A recent review (Kaldor, 1993) of all published HIV seroprevalence surveys indicates that prevalences of less than 5% have been found consistently in female and heterosexual male IDUs although prevalences are far higher among homosexual/bisexual male IDUs. Seven studies published between 1985 and 1991 reported prevalences of HIV of 1%, 4%, 5%, 1%, 3%, 2% and 4% among IDUs after excluding homosexual/bisexual males (Kaldor, 1993).

These findings are supported by HIV test results of prison entrants in several jurisdictions which continue to be well under 1%. HIV seroprevalence among prison entrants in New South Wales was only 0.59% despite an AIDS incidence of 357.9 per million. In Victoria, HIV seroprevalence among prison entrants was 0.47%. It is generally assumed that more than 50% of prison inmates in Australia are serving time for drug-related offences and an even higher percentage are IDUs. Anecdotal information from testing laboratories is also consistent with a low prevalence of HIV infection among IDUs. A low prevalence of HIV infection has also been reported in a number of studies of prostitutes.

The number of new diagnoses of HIV infection nationally was estimated to be about 1,201 in the year to 31 March, 1992 and 1,031 in the year to 31 March, 1993 (National Centre in HIV Epidemiology and Clinical Research, 1993). Of cases with a documented known exposure category among the 17,068 new diagnoses of HIV infection cumulative to 31 March, 1993, 81.9% were attributed to male homosexual/bisexual contact, 2.8% to male homosexual/bisexual contact and IDU and 4.9% to IDU alone. It is estimated that there were only 600 new HIV infections nationally each year in the period 1989-90 (National Centre in HIV Epidemiology and Clinical Research, 1992). These trends are shown in Figure 1.

A cumulative national total of 4,102 AIDS cases (including 4,073 persons over the age of thirteen) had been reported up to March 31, 1993 representing a case load of 23.6 per 100,000 (National Centre in HIV Epidemiology and Clinical Research, 1993). Of the cases in adults, 3,462 (84.4%) were homosexual/bisexual males, 128 (3.1%) were homosexual/bisexual male IDUs and only 78 (1.9%) were female or heterosexual male IDUs with an additional 405 (9.8%) consisting of heterosexuals, haemophiliacs, recipients of blood transfusions or blood products or undetermined. These data are consistent with seroprevalence data suggesting that HIV has not yet become established in heterosexual IDUs in Australia.

The low prevalence of HIV infection in IDUs in Australia allows multiple interpretations including late entry of HIV into the IDU population, limited pool of infection in other risk groups, limited overlap between risk groups, substantial spread remaining undetected and serendipity.

All of these possibilities can be effectively discounted except the last. HIV infection has been present in the heterosexual IDU population in Australia since at least 1985 when one of 200 Sydney IDUs in drug treatment, a heterosexual male from the United States (resident in Australia for about a decade), tested seropositive but several of his sexual and needle contacts were HIV infected (Blacker, 1986). The area most associated

with drug use in Australia is in eastern Sydney and is also associated with a very large community of homosexual/bisexual males constituting the national epicentre of the HIV epidemic. The geographical overlap of these two populations makes the low prevalence of HIV among Australian IDUs all the more remarkable. The gay community in Sydney has long maintained close links with the gay community in San Francisco. Not surprisingly, HIV entered the gay population of Sydney relatively early.

A steep gradient of HIV infection has been observed in a study of 1,245 Sydney IDUs in 1989 (Ross, 1992, AIDS Care 139-48) with 3.2% of heterosexual male, 12.1% of homosexual/bisexual male and 35.4% of homosexual male IDUs infected with HIV suggesting that HIV entered the IDU population from homosexual/bisexual male IDUs. As 5.6% of male IDUs were homosexual and 13.1% were homosexual/bisexual (ANAIDUS, 1991), there was clearly considerable overlap with non-drug using homosexual populations.

The course of the HIV epidemic in Australia differs from that of many other western countries in that a significant IDU component has not yet occurred. The first case of AIDS was diagnosed in Australia in 1982. In the early 1980s, Australia had a high per capita incidence compared to other OECD countries. In 1983, Australia ranked fourth among developed countries in terms of AIDS cases per capita with 1.1 cases of AIDS per 100,000 population. By 1991, Australia had slipped to sixth place with 16.6 cases per 100,000 being overtaken by Spain (23.3) and Italy (17.2). Both Italy and Spain have experienced an explosive spread of HIV infection among IDUs who now represent over 60% of all AIDS cases in those countries. If Australia had still retained fourth highest ranking of AIDS cases per capita among developed countries, this would mean that instead of 4,073 cases of AIDS as of the March 31, 1993, 5,717 AIDS cases would have been expected. Of all known AIDS cases in Australia at present, 64% have died. There are therefore an estimated 1,052 Australians alive today because the epidemic appears to have run a different course in Australia than some other countries. The major difference has been the absence of an epidemic in IDUs. If the medical management of each AIDS case in Australia cost A$50,000, and a figure of A$100,000 is far more realistic, this represents a saving of 1,644 AIDS cases, 1,052 lives and A$53,000,000.

It is reasonable to conclude that HIV entered the Australian IDU population early, that substantial HIV infection is present in other risk groups, that there is substantial overlap between IDUs and other risk groups and that substantial undetected spread of HIV among IDUs can be discounted. Accordingly, the most parsimonious conclusion is that the course of the epidemic has been altered compared to other countries.

AIDS was identified as a priority issue early in the epidemic. A National AIDS Task Force was established rapidly. A highly controversial national advertising campaign in 1987 on a ''Grim Reaper" theme succeeded in its aim of raising awareness of AIDS as an issue. The Health Minister at the time described AIDS as "the greatest threat to public health in Australia since Federation" (i.e. 1901). The role of prevention was given

particular emphasis and the need to maintain control of HIV among IDUs was well accepted (Wodak, 1992). Prevention strategies were identified early, adopted with broad support and implemented vigorously. IDUs (and other high-risk groups) were involved in the identification of prevention policies and their implementation. Parliamentary all party AIDS committees were established at Federal and State levels with explicit agreement to refrain from party political conflict. A national drug policy of minimization of harm had been declared in 1985 at a meeting of the Prime Minister and State Premiers and facilitated the adoption and implementation of sensitive HIV prevention strategies even though AIDS had not influenced consideration of the original drug policy. The Prime Minister's wife set a prominent example in the mid-1980s indicating that discrimination against persons with HIV infection was unacceptable.

The Prime Minister's wife also launched a "Never ever share needles" pamphlet in December, 1987. This campaign made no mention of abstinence from drug use. It was well known that one of her daughters had been an IDU and this endorsement of harm minimisation by such a prominent member of the community was a critical development in the nation's response to the impending threat of an epidemic. IDUs became aware of the hazards of needle sharing before HIV had gained a substantial foothold in this population.

In 1989, a National HIV/AIDS Strategy was agreed following lengthy consultation (Department of Community Services and Health, 1989). This document endorsed needle and syringe exchange and distribution programmes and resulted in policy and financial commitments covering a three year period. In 1991 it was noted that "programs for IDUs receive a larger share (37.4%) of total education and prevention funding ... than programs for any other target group. Funding grew substantially in all States and Territories from 1989-90 to 1990-91 with a 36% increase overall." (Inter Governmental Committee on AIDS, 1992)

Adoption of HIV prevention measures within prisons however, has been slow, difficult and incomplete. A communique covering strategies to prevent the spread of HIV in prisons was unanimously endorsed at a national conference in November 1990 (Douglas, 1991) but has had little effect.

A vigorous and at times acrimonious debate about needle and syringe availability as a prevention strategy took place in 1985-6 with covert support for implementation coming from senior political and Health Department figures. A pilot (illegal) programme was established in Sydney in November, 1986. In December, 1986, the New South Wales Department of Health established a drug store based needle and syringe distribution scheme with NSEPs set up from 1987. All other jurisdictions (except one) rapidly established drug store schemes and NSEPs. The last jurisdiction to introduce

needle and syringe exchange has permitted NSEPs to be established unofficially while waiting for some years for the appropriate legislation to be passed. Sterile injecting equipment has been readily available in all major cities and large towns in Australia since 1988/1989 (see Table 1). Providing NSEP services to rural IDU populations has often been difficult because of logistical problems and the more conservative nature of small towns.

Several years were required for most NSEPs to reach maximum through-put and begin to achieve efficiencies. For example, a NSEP in western Sydney recently increased its throughput more than ten times (from 3,000 to 40,000 units per month) over a two year period within a constant budget by building up a fixed outlet at the expense of a mobile unit. A survey in October, 1992 of 43 clients of this service found that 84% rated the service as excellent with the remaining 16 rating it as good (Duckett et al., 1993). The cost of distribution per unit declined from A$2.86 in 1991 to A$1.04 in 1992 and is expected to decline to A$0.96 in 1993/94. Similar trends are occurring elsewhere in the country. In some places, mobile units operate from cars with paging devices or portable telephones. This presents a difficult balance between the higher unit cost of providing mobile NSEPs and the need to service more vulnerable and less mobile populations. Mobile units were recently scrapped in one state as a cost cutting exercise.

It is difficult to estimate the quantity of injection equipment made available in Australia each year but it is likely that at least three to four million sterile needles and syringes are distributed or exchanged each year. In 1991, New South Wales (population 6.5 million) had 32 primary and 90 secondary outlets while Victoria had 102 outlets (Inter Governmental Committee on AIDS, 1992).

After six years of NSEP in Australia, there have been relatively few unintended negative consequences. Littering of public places with discarded used equipment was briefly an issue which threatened to jeopardise (then) fragile public support. This problem was overcome by a series of measures including specialised collection of used equipment, special disposal bins and the marketing of specially developed plastic boxes ("Fitpack") which contain sterile equipment and a tamper resistant device for retaining used equipment. The Fitpack was designed with the assistance of members of a government funded drug users organisation. Staff of NSEPs have been detected using these facilities to also distribute illicit drugs on two occasions. In one state, a NSEP was operated by a government funded organisation of drug users. Following allegations of embezzlement, the operation was handed over to a government agency (with subsequent decline in throughput). A study of urine analysis specimens obtained from two methadone units, one of which was immediately adjacent to a pilot NSEP, concluded that sterile needle and syringe availability did not appear to increase the frequency of drug use in patients of methadone programmes (Wolk, 1990).

The mean frequency that sterile needles and syringes where used by IDUs (n=2,451) when injecting drugs ranged in four cities in 1989 from 69.1% to 79.6% with the overall mean for respondents being 72.6% (ANAIDUS, 1991). Respondents were also asked "how easy do you think it is to obtain new (sterile) needles and syringes at the present time?" with answer options being that it was easy 0-25%, 26-50%, 51-75%, 76-98% or 99100% of the time. The mean percentage of the time respondents in three cities in 1989 reported finding it easy to obtain new injecting equipment ranged from 72.0 to 84.3%.

The Queensland government introduced exchange and distribution schemes after the other states and thus availability in Brisbane lagged behind other cities in the study. Male respondents found it significantly easier to obtain equipment than female respondents (77.2 +/-26.8; 73.5+/-26.9; p<0.01) (ANAIDUS, 1991).

The importance of drug stores as outlets was emphasised in responses to the question "where do you get your new needles and syringes?" Drug stores, NSEPs and after hours drug stores were the most important sources (see Table 2).

The mean number of needles and syringes obtained in the most typical using month (n=2,422) in 1989 was 54.2 +/-91.6. The most important time to obtain equipment was mid-afternoon to midnight with the period 9 pm to midnight being especially important (ANAIDUS, 1991). When asked to indicate how needles and syringes could be made more available, a range of responses was obtained but more drug stores selling needles and syringes, vending machines and special needle and syringe exchanges were the most common answers received (ANAIDUS, 1991).

Availability of injection equipment had further improved in two of the cities twelve months after the original data collection (ANAIDUS, 1992) with the mean number of respondents reporting that it was easy to obtain new equipment in Sydney increasing from 84.3 to 93.0% (see Table 3). Drug stores were still the major source of new equipment but NSEPs were nominated as a more important source than twelve months earlier.

Needle and syringe exchange schemes have broadened the scope of their activities without diminishing the enthusiasm of their staff. They now often prefer to be known as HIV prevention units and in some areas have become involved in AIDS coordination and HIV prevention advocacy. Vending machines have been introduced in small numbers but are still undergoing evaluation. Their introduction took much longer than anticipated. Vending machines are unlikely to ever replace NSEPs but are being used to provide an affordable 24 hour service in areas with a particularly high incidence of drug use.

IDUs who usually attended NSEPs or drug stores to obtain sterile injecting equipment were compared (Wodak, submitted for publication). Discriminant function analysis correctly allocated 75% of respondents. Those whose usual source of supply was NSEPs obtained more than twice as many needles and syringes per month, injected alone less frequently, were more likely to reside in the inner city than outer suburbs, injected with new needles and syringes more frequently than those whose usual source was drug stores, and were less likely to be in drug treatment. NSEP attenders reported that they

found it more easy to obtain injection equipment a significantly greater percentage of the time, were younger, less likely to inject with a needle and syringe used by someone else, and spent more money on drugs per week than respondents whose usual source of supply was drug stores. Almost twice as many drug store as NSEP attenders were employed (although unemployment was common in both groups). NSEPs attracted a greater proportion of women. These data suggest that some IDUs were deterred from obtaining injection equipment from drug stores by modest prices, while others in drug treatment prefer to obtain their equipment from the (presumably) more anonymous setting of a drug store. Different kinds of IDUs appear to utilise different kinds of needle and syringe outlets

Legal impediments to HIV prevention have recently been comprehensively reviewed including obstacles to NSEP operation (Inter Governmental Committee on AIDS, 1992). The Legal Working Party made recommendations covering repeal of self administration and other offences, the need for more non-custodial sentencing options and supported further research including the investigation of drug policy reform.

During the 1980s, considerable efforts were made to develop effective policies and review progress. In more recent years, the attitude that the epidemic of HIV among Australian IDUs has been prevented has become commonplace and is often accompanied by a sense of complacency.

Explicit education campaigns directed to IDUs began in 1987. A number of education campaigns have been conducted including mass campaigns which may have also contributed to the broad support existing for HIV prevention activities. Homosexual and homosexual/bisexual male IDUs have been specifically targeted in education campaigns. Social marketing approaches have been used to raise and maintain a high level of awareness about HIV/AIDS in sub-populations of IDUs. A low level campaign targeting homosexual male IDUs has been running in NSW for some years using the slogan "FIT FOR A QUEEN. NEW OR CLEAN."

Expansion and improvement of drug treatment began in 1985 with national methadone capacity increasing in the last eight years more than six times. The most rapid expansion of methadone capacity has occurred in NSW (Gaughwin, 1993) which has 34% of the national population and almost 60% of the nation's AIDS cases. The national methadone guidelines have been revised a number of times in recent years in an effort to liberalise programmes to increase their attractiveness, improve retention rates, reduce costs and assist national efforts to contain HIV infection in IDUs. The unit cost of providing methadone programmes has been falling in real terms and is now about A$1,200 per person per year. Other modalities of drug treatment have also been expanded and improved. AIDS research is regarded as a priority area and is still funded separately.

Organisations of IDUs have been established in all jurisdictions with government support and funding. The Australian Prostitutes' collective was established in 1985.

IDU representatives are invited to actively participate in policy development and implementation.

National efforts were made to increase bleach availability and utilisation. Many NSEPs had difficult distributing bleach to IDUs as sterile injection equipment was so available there was little interest in obtaining alternatives. Since the biological effectiveness of bleach as a decontamination agent began to be questioned publicly in the US in 1993 (National Institute on Drug Abuse. Community Alert Bulletin, March 1993), the wisdom of advocating bleach decontamination has been reviewed. Decontamination practices of IDUs in Australia fall far short of acceptable and are a further reason for reviewing exhortations to IDUs to use bleach.

There is some evidence that high risk injecting practices are becoming less common but as in other counties, baseline levels of unsafe injecting practices were very high and unsafe sexual practices appear to have declined only slightly. HIV prevention policies are widely regarded as having successfully prevented the spread of HIV among IDUs although rigorous scientific proof of their effectiveness is unavailable, probably unobtainable and was wisely not regarded as a prerequisite for adoption and expansion of prevention programmes. HIV prevention policies for IDUs continue to have very strong community and political support.

In 1989, one sixth of a Sydney sample (17%) were at low risk of HIV infection as they had never shared injection equipment, cleaned injecting equipment effectively 100% of the time it was shared, and were celibate, monogamous, or had not had unsafe sex in the past six months (Wodak, in press). Half (51%) had either unsafe injecting or sexual behaviour, with the remaining third (33%) engaging in both unsafe injecting and sexual practices. Comparison of two large and consecutive (1989 and 1990) samples of Sydney IDUs recruited in non-treatment settings were consistent with major risk reduction (Ross, in press) (see Table 4).

These data are drawn from two cross-sectional samples and therefore the possibility that the differences observed were due to sampling cannot be discounted. Nevertheless, the strength and consistency of the behavioural differences, the similarities in demographic characteristics and drug use of the two samples, and the similarity of findings in comparable studies in other countries suggest that these behavioural differences are real. They are all the more remarkable when it is considered that the two samples were recruited less than twelve months apart.

Attributing benefit to any single intervention is impossible when multiple strategies have been implemented at about the same time, the intensity of implementation is difficult if not impossible to measure, and the effect of interventions is in all likelihood synergistic. In a categorical sense, these methodological problems can not be resolved without a controlled trial of communities randomly allocated to a single intervention or no intervention. The ethical, logistic, financial and public health problems of attempting such a study are such (Des Jarlais, 1993) that there is no alternative, especially given the

urgency of the epidemic, to making a judgement on the grounds of plausibility, feasibility, cost and international experience. At issue is whether authorities in a particular country prefer to be roughly right or precisely wrong. Australian authorities, perhaps reflecting a characteristically pragmatic and non-ideological national approach, preferred the former option. So too did most other developed countries.

Continuing high levels of unsafe injecting practices and international spread of the HIV epidemic within and to an increasing number of countries suggest that complacency about the possibility of future spread of HIV among (and from) Australian IDUs is unwarranted. There is increasing concern in Australia that the magnitude of the Hepatitis C (HCV) epidemic and its consequences has been seriously underestimated.

High incidence levels of HCV in Australian IDUs, especially young IDUs, evidence of annual incidence rates for Hepatitis B and C of between 10 and 20 % in Victorian inmates with more than one occasion of prison entry (Crofts, 1993a), and recent estimates that there are five times more people in Australia infected with HCV than HIV and at least a fifteen times higher incidence of HCV than HIV (Crofts, 1993b) may lead to a reassessment of the public health threat of IDU-related blood borne viruses. Evidence of continuing high incidence levels of Hepatitis B and C among IDUs in Australia also indicates the potential for spread of other blood borne viruses including HIV.

Although the morbidity and mortality associated with HIV exceeds that of HCV, the much larger pool size and higher incidence of HCV in Australia suggests that far greater attention needs to be directed in the future to the containment of blood borne viruses in addition to HIV. It is estimated that at least 20% of HCV infected individuals will develop cirrhosis within 5-10 years with up to 5% developing a hepatocellular carcinoma.

There is some evidence that HIV spread in prisons has been underestimated in terms of documented seroconversions. The relatively short mean duration of imprisonment in relation to the "window period" for seroconversion suggests that entry-exit testing of HIV infection underestimates the extent of the problem. A network of IDUs who became infected with HIV while in prison is currently being investigated with at least one of these cases being virtually certain (K. Dolan, personal communication). There are about a dozen published cases of HIV infection in prison world-wide but none of these cases can be regarded as definite.

IDUs enter correctional facilities from diverse geographic and social backgrounds. They usually remain within correctional facilities for relatively brief periods during which time they are often moved frequently mixing with many other prisoners. In contrast, unsafe injecting practices among IDUs in the community are increasingly restricted to small social networks. HIV infection within such a network has far less public health impact than infection of a prison inmate with the potential for wide dissemination of HIV within correctional facilities and subsequently following release.

Unsafe injecting practices in prison are less frequent than in the community but may be more hazardous for several reasons. Injecting equipment is less available in prison and is therefore likely to be shared between a larger number of partners. Bleach is also less available inside prison. Drug injecting and equipment decontamination is also likely to be more furtive with less opportunity to carefully decontaminate injecting equipment.

Evidence of an unofficial needle and syringe exchange programme successfully conducted for almost a year in several prisons in NSW has emerged recently and has been presented to a committee of inquiry. This unofficial strictly "one for one" trial involved over 100 prisoners and was conducted by an HIV infected prisoner with covert assistance of some health professionals and possibly with the knowledge of some correctional staff. The existence of a prison NSEP lasting almost twelve months raises the possibility of considering this intervention more widely at a time when the effectiveness of current decontamination strategies relying on bleach have been called into question.

In 1990, an HIV infected prisoner stabbed a NSW prison officer with a syringe filled with the inmate's blood. The prison officer seroconverted over the next months becoming the first documented case of occupational exposure of a prison warder. It is anticipated that recommendations to consider a pilot NSEP in an Australian prison will therefore meet much resistance and accordingly should not be supported at the risk of endangering more likely interventions.

The Australian experience with HIV prevention among IDUs has stimulated a more critical appraisal of prohibition. There has been increasing support for the view that prohibition is expensive, ineffective, counter-productive and impairs the effectiveness of efforts to control the spread of HIV. The number of influential members of the community calling for a review of drug policy in recent years has been steadily growing. The Australian Parliamentary Group for Drug Law Reform was established in 1993 and includes representatives of the major parties with members drawn from Commonwealth, State and local levels of government.

A debate about the fundamental nature of drug policy has been slowly intensifying with fundamental reform strongly supported by many leaders of the medical and legal professions and most doctors involved in delivering HIV prevention or treatment services. The National Centre for Epidemiology and Population Health, Canberra, is investigating the possibility of a trial of controlled availability of currently illicit drugs

and was awarded a grant of A$0.5 million by the Australian National University to further this work. This trial followed from a recommendation made by an Australian Capital Territory Legislative Assembly official enquiry into HIV, illicit drugs and prostitution.

Whether a debate about drug policy reform would have developed in the absence of adoption of needle and syringe exchange programmes is arguable. However, the fact that liberalisation of availability of injection equipment has been so beneficial and virtually unaccompanied by unintended negative consequences has certainly drawn attention to the lack of evidence of effectiveness, high costs and major adverse consequences associated with prohibition.

Prohibition is seen by some to keep street drug prices high and purity of street drugs low and thus reduce the possibility of IDUs making a transition to non-parenteral modes of administration. Prohibition also delayed the adoption and slowed expansion of NSEPs and also impeded the implementation of many other programmes needed to control the HIV epidemic.

These conclusions may be of interest to other countries especially those which share many similar characteristics such as the United States. However, the many major differences between Australia and the United States must also be borne in mind.

The United States and Australia were both colonised by Great Britain. When the United States declared independence, Great Britain lost a favoured storage site for surplus prisoners. The search for a new gulag lead to the establishment of a colony in Australia. The two countries have similar political systems and the curses and blessings of a federal system of government. The Australian political system is sometimes referred to as Washminster reflecting the debt to both Westminster and Washington. The United States and Australia have been parliamentary democracies without interruption. English is the main language spoken in both countries. The area of the continental United States is only fractionally larger then Australia. Both countries have populations drawn from diverse cultures. US drug policies have been a dominant influence on Australia drug policies since soon after the turn of the century. In both countries, the drug policy is essentially prohibition although this takes a kinder and gentler form in Australia.

However there are many differences between Australia and the United States. Australia only has about 6% of the population of the United States. Like most other western countries, Australia has a universal health care system. A policy of harm minimization for illicit drugs has been adopted at the highest political level in Australia. This policy states that the aim "is to minimize the harmful affects of drugs on Australian society". In contrast, US drug policies has been dominated (explicitly) by attempts to reduce drug use. Although poverty and severe social disadvantage exist in both countries, Australia has never had a large urban under-class as is seen in a number of major US cities. Furthermore, policy makers and injecting drug users in Australia come from the same racial group and speak the same language. Moreover, injecting drug

users and policy makers in Australia are in contact with one another. In recent years, some injecting drug users have been employed on Government projects because they were currently injecting drug users.

Social policies are markedly different in Australia and the United States. In Australia, abortion is readily available and ceased to be an issue several decades ago. Capital punishment was last carried out decades ago. In 1992, it was decided that sexual orientation of male and female military recruits would no longer be taken into account (despite strenuous opposition of the military). All Australian jurisdictions have laws restricting the availability of guns which are much stricter than in the United States. Crime rates and rates of imprisonment are much lower in Australia than the United States.

In most western countries including Australia, the questions policy makers ask about the availability of sterile injection equipment for injecting drug users is not whether this is a legitimate strategy to prevent the spread of HIV infection or even whether this policy contributes to improved control of the epidemic. Improving the availability of sterile injection equipment is assumed to make a most important contribution to the control of the epidemic even in the absence of categorical proof. In Australia, the standard of proof required for evaluating the contribution of needle and syringe exchange programmes to improve control of the epidemic was commensurate with the speed of spread of the epidemic and the magnitude of the consequences of an uncontrolled epidemic.

The critical question for policy makers in Australia regarding needle and exchange is how needle and syringe availability can be implemented more effectively and at lower cost. There is a strong desire to focus on improving availability particularly for groups of major public health importance such as homosexual/bisexual male injecting drug users. After seven years of needle and syringe availability, political, bureaucratic and community support is still overwhelming.

PETER SELWYN

Peter Selwyn noted that a large number of studies from the United States and many other countries throughout the world have provided generally consistent data about needle exchange programs. Nevertheless, underlying social and political dynamics in the United States seem to be requiring a standard of scientific proof for such programs that is much higher than that usually required of other preventive, or even clinical, interventions. The more important underlying dynamic appears to be the social context for policy concerning drug injection and related preventive interventions. The tendency to respond to drug injection within a criminal justice as opposed to a medical public health context inhibits responses that are truly preventive and educes instead responses that are more proscriptive and punitive. Thus, he observed, there is the irony that the United States, which is both heavily affected by injection drug use and by the AIDS epidemic, has served as an alarming negative example to other countries around the world. Other countries have gone on to develop comprehensive AIDS prevention programs targeted on drug users; in the United States, however, similar steps have not been taken in any systematic way, notwithstanding the many important steps undertaken in selected individual cities and communities.

He commented that it may be helpful to separate the issues that surround needle exchanges into three categories: (1) the data that exist, (2) the limitations of the data, and (3) the political, cultural, and often emotional issues that underlie the data. As for the data that exist, he said, it is clear that virtually all of the research that has looked at outcomes of needle exchange programs around the world points in the same direction-that is, toward no evidence of significantly increased use of drugs, initiation of drug use, or drug injection at an earlier age. There is also evidence of reductions in needle sharing among participants in needle exchange programs and lower or stable HIV seroprevalence. The importance of increased access to drug treatment and other medical interventions for active injectors is also a consistent finding.

Selwyn noted the methodological limitations of the data that have already been discussed—sampling frames, nonrandomized designs, inability to separate needle exchange effects from other program effects and secular changes over times, validity of self-report data, and so on. An important point that bears repeating is that use of HIV seroconversion as an outcome measure may not be feasible or even relevant to the evaluation of needle exchanges. In most of the studies that were presented, needle exchanges were being introduced into populations that were either already heavily saturated with HIV infection or still at a very low level of seroprevalence. Thus, statistically demonstrating a reduction in seroconversion would have been a priori difficult, just from the standpoint of sample size. Another important limitation is the inability to separate easily the effects of other behavioral factors, such as sexual risk behavior, from the effects of needle exchanges. Finally, not so much from a data

standpoint, but perhaps more conceptually, there is the difficulty of proving that something did not happen. As the presentation on Australia's experience made clear, one cannot say that something would have happened otherwise had it not been for the programs that were implemented.

Turning to the underlying political and cultural framework, he noted, the social acceptability of a behavior, in general, lends itself proportionately to preventive, as opposed to punitive or proscriptive, approaches to that behavior. Thus, there has been less reluctance in the United States to introduce behavioral interventions aimed at reducing heterosexual transmission of HIV than transmission among homosexuals or bisexuals and injecting drug users.

In conclusion, Selwyn observed, based on the existing data and leaving aside the policy and underlying social dimensions for the moment, needle exchange programs do not make things any worse and there are some data—e.g., from Kaplan in New Haven-that they may make things better. They do not cause people to use drugs, they seem in many instances to promote positive behavioral change, they definitely help provide access to vulnerable and sometimes otherwise inaccessible populations, and they may reduce the risk of transmission of HIV. But, as all of the preceding presenters noted, needle exchange programs should not be looked at as a single or simple solution. The data from Amsterdam that were presented are an interesting reflection of this point. The fact that cocaine use, alcohol use, and homelessness were the factors that predicted HIV seroconversion, even among people who used the needle exchange, suggests that it is not simply a matter of handing out clean needles. The behavioral and social factors underlying drug use must also be addressed.