General Discussion

9.1 Introduction

Prior to the AIDS epidemic, many IDUs all over the world shared needles and syringes. The present research was directed at particular aspects of needle sharing in relation to HIV prevention, over the years 1985-1991. Before discussing the present findings, it seems important to focus on possible reasons for the high level of needle sharing before AIDS.

The sharing of drugs and other resources forms an important part of the drug subculture1 ,2. This sharing should be seen in the context of the marginalization and criminalization of drug users, which forces members of the 'deviant' group to share in order to cope. Thus, sharing of specific items may be primarily related to their availability. It may also have become a ritual (i.e. a planned behavioral sequence invested with a special meaning for the person performing the sequence3 ,4) or it may be a pleasurable experience in itself. When sharing is chiefly dependent on the availability of resources, one would expect an increase in available resources to lead to a decrease in sharing. However, if sharing is mainly a ritual, or experienced as pleasurable in itself, no such relation will be found. These possible causes of sharing also apply to the subject of this paper: needle sharing.

In general, post-AIDS studies have found strong links between needle sharing and limited equipment availability. Needle sharing, in one ethnographic study of drug taking rituals in Rotterdam', appeared to be determined by a) the availability of needles and syringes, b) experience with the injecting ritual and c) drug craving. Needle sharing as a ritual was not observed. In Edinburgh, a policy of restricting injecting equipment from 1982 to 1985 led to high levels of needle sharing, which was followed by a Hepatitis B and HIV epidemics. On the other hand, Glasgow and other areas, which did not restrict equipment availability, had much lower levels of needle sharing and a smaller HIV problem6. In the U.S., IDUs with diabetes were found to have a significantly lower HIV seroprevalence than nondiabetic IDUs. Despite a similar duration and intensity of drug use, diabetic IDUs tended not to share injection equipment and were less likely to attend shooting galleries than nondiabetic IDUs. This was ascribed to the more ready access that diabetic IDUs have to sterile injection equipment7.

Further support for the link between sharing and availability of resources is that, since 1984, IDUs all over the world have reduced their risky injecting practices in response to health education and provision of means8 ,9. When there is no structural scarcity of means, IDUs themselves tend to see needle sharing as a deviant act2 ,3. Finally, unlike high risk sexual behavior for homosexual men (unprotected anal sex), injecting with a needle and syringe used by somebody else is not pleasurable in itself. Injectors generally prefer to inject with new sterile injection equipment, which functions better than used equipment, does not get clogged as easily and has a sharp needle. Adaptive behavior thus seems more pleasurable and rewarding than maladaptive behavior. For homosexual men, the opposite appears to be true: adaptive behavior (condom use during anal sex or abstinence of anal sex) is generally experienced as less pleasurable than maladaptive behavior.

All together, this suggests that the high prevalence of unsafe injecting at the start of the AIDS epidemic was mainly the result of scarcity of means, together with a lack of awareness, or a low level of perceived threat, of the known medical complications of unhygienic injecting procedures among injectors.

Against this background, the findings of the present research will be discussed in five sections. First, in section 9.2, an overview will be provided of the samples studied as part of the present research, and the generalizability and reliability of the collected data will be discussed. Section 9.3 discusses the contribution of the low threshold methadone programs to HIV-prevention, and section 9.4 compares syringe exchange attenders to non-attenders with regard to injecting risk behavior. The findings with regard to determinants of needle sharing are discussed in section 9.5. Section 9.6 focuses on HIV-seropositive drug users.



9.2 The present research: generalizabiIity and reliability of data

In general, little is known about the representativeness of study samples of drug users or IDUs; the present research is no exception. Therefore, one should be careful in generalizing the present findings. Furthermore, participants are self-selected, and self-selection occurs again with respect to participation in the follow-up study. When compared to IDUs in low threshold methadone programs, IDUs who enter the cohort-study appear to have a slightly increased risk of being HIV-positive'°. On the other hand, there is evidence that behavioral risk reduction occurred especially among drug users participating in the follow-up study".

The samples studied in the present research differ from each other on a number of characteristics, for example with regard to seroprevalence, knowledge of serostatus, history of injecting, kind of visit in the epidemiological study (for the studies reported in chapters 3 to 8) and year of data collection. Table 1 summarizes these differences. The table suggests a decrease in borrowing and lending over time. However, chapters 3 to 6 concern only or partly intake visits, while chapters 7 and 8 concern only follow-up visits. As evidence exists that participation in the follow-up study is associated with behavioral changes", it seems likely that this apparent decrease is an artefact of sample selection.

Table 1 also shows that, apart from the study described in chapter 2, lending is in general less often reported than borrowing. Selfreports of injecting risk behavior may be unreliable due to memory problems or to answering biases. Long-term drug users may have impaired memory; apart from that, it may be extremely difficult to recall complex patterns of drug use accurately". Memory may be more impaired for some groups of users, e.g. drug users with coexisting psychopathology or HIV-infected drug users1 3,14. Nevertheless, studies trying to validate self-reports of drug use with data obtained from other sources, mainly urinalysis, have generally found reasonably close agreement13 The discrepancies found in the present research between self-reports of borrowing and lending suggest an answering bias rather than memory problems. Lending may be underreported, possibly because reporting lending is more threatening to one's selfesteem than reporting borrowing. As borrowing concerns potentially self-destructive behavior, it may be less embarrassing for drug users to admit to this. However, two alternative explanations are also possible: overreporting of borrowing, or the existence of a relatively small group of frequent lenders in combination with a large group of infrequent borrowers. Unfortunately, no data on frequency of lending are available.

Within the cohort-study, which operates strictly apart from the low threshold methadone program, measures aimed at improving reliability include anchoring to autobiographical events": for follow-up participants questions on behavior concern the period since their last visit. The interviews are done by trained interviewers who are experienced with drug users, and (potentially) threatening material is asked later in the interview. To minimize threat, questions assume risk behavior, instead of safe behavior, in order to give the respondent the impression that risk behavior is not unusual.

Despite these measures, self-reports of risk behavior may still not be reliable. Attempts can be made to check reliability by measuring the tendency to respond in a socially desirable way13,15. However, it is questionable whether such a general inclination gives an indication about a response bias with regard to a specific behavior, i.e., borrowing or lending. Self-reports of injecting can be validated with unobtrusive measures1 6, for example physical evidence of injection", while self-reported needle sharing, as suggested by Samuels et al.1 3, can be validated with measures of immune parameters, or with forensic testing of used syringes".

In order to obtain indications for risk reduction at a behavioral level, it is often thought that "hard" data, like the prevalence and incidence of hepatitis-B or HIV infection, are more objective measures than self-reports of behavior. However, this is a problematic assumption. The occurrence of (new) infections in injectors is related not only to the sharing of needles and syringes or the sharing of other equipment (for example, spoons") but also to other factors, like:

-the social network in which one shares (the chance to borrow from a seropositive person), 20-22

- the quality of the employed risk reduction strategies (e.g. refusing to share if there is visible blood in the syringe),23

- the size and type of injection equipment used (e.g., t ml syringe versus 2 ml), 24,25

- the practice of 'booting' (i.e. the repeated withdrawal and reinjection of blood) by the person from whom a needle and syringe is borrowed, which leaves a blood residue in the syringe, -24,26,27

- the duration since the borrowed injection equipment has last been used, 28

-the efficacy of cleaning borrowed injection equipment (cleaning with water or with a disinfectant like bleach) . 17,21,11

To a certain extent, these factors are related to chance, and therefore not stable over time. Furthermore, several factors are not under control of the person borrowing. Thus, it appears incorrect to conceptualize the prevalence and incidence of infections as more objective measures of behavioral change than self-reports. Instead, they should be seen as different measures, reflecting not only risk behavior itself but also network characteristics, mixing patterns, changes in the social context of drug use and the other factors mentioned above. Blower's argument 31 that stabilization of seroprevalence levels is possible without changes in behavior can be extended to incidence: a decreasing incidence does not necessarily reflect behavioral risk reduction, nor does an increasing incidence necessarily reflect increases in risk behavior. To conclude, although self-reports of injecting risk behaviors may be unreliable and are difficult to validate, they provide indispensible information next to information on the spread of infections transmitted through these risk behaviors. As Blower 31 suggests, mathematical models that are specific for a particular IDU community may be required in order to assess the epidemiological effects of certain behavioral changes.

9.3 The contribution of the low threshold methadone program to HIV prevention

No indications were found that long-term regular methadone users (LTM users) injected less often or had less injecting risk behavior than non-LTM users. Chapter 4 shows that, among drug users with a history of injecting who entered the cohort study through the low threshold methadone program, approximately 80% had injected, 35% had injected daily and 40% had borrowed in the six months preceding entry into the cohort study. When comparing LTM users with non-LTM users on these variables, no differences were found. Controlling for confounders, there was also no significant difference between the two groups with respect to HIV prevalence. Thus, no indication of a protective effect of LTM use on HIV risk was found. The study reported in chapter 2 failed to find a relationship between safe injecting and longterm contact with a methadone program; the study reported in chapter 6 failed to find a relationship between safe injecting and current daily methadone use. A study of incident HIV infections among participants in the cohort-study 32 failed to find a protective effect of daily methadone use on seroconversion rate. In view of the low average daily methadone dose (35 mg.) prescribed by the low threshold methadone program, and the permitted use of other drugs, these findings are not really surprising.

The rationale for the low methadone doses in the Amsterdam program may have been unfulfilled expectations with regard to providing high dose methadone in the seventies 33. 3 4. Through cross-tolerance, an adequate dose of methadone is supposed to block the effect of heroin. This would result in a gradual extinction of the drive to use heroin 35. However, Van Epen 33 reports that Dutch drug users who received 100 mg methadone daily still felt the "flash" of 10 mg morphine. Furthermore, addicts who were prescribed high doses of methadone kept using heroin on the side. Van Epen suggests that these findings were due to the relative high purity (30-50%) of Dutch street heroin, as compared to a lower purity of U.S. street heroin. However, Platt and Labate 34 mention several U.S. studies in the early seventies which also failed to find a sharp reduction of drugcraving and heroin use among high methadone dose maintained addicts. Based on his experience with prescribing high dose methadone, Van Epen decided to give doses so low that his clients stayed just above the level at which withdrawal symptoms would manifest themselves. Other factors which played a role in the Netherlands were resistance to high dose methadone maintenance, both among the staff of the programs and participants 3 6, .37. In the beginning of the 1970s, a number of methadone overdoses had occurred. Furthermore, staff members saw high dose maintenance as a treatment option only for those addicts who were judged hopeless cases with respect to other methods of treatment, and it was often felt that high dose maintenance in fact replaced one addiction by a more severe one. Many addicts did not want to see themselves as lifelong methadone addicts.

In the AIDS era, the main goal is to reduce injecting, rather than to stop or reduce heroin use. A study in the U.S38 which compared different methadone maintenance programs, found that methadone dose level has a strong negative relation to current injecting. Furthermore, an adequate methadone dosage was a much better predictor of current injecting than patient characteristics like length of addiction. Duration and quality of treatment were also important factors in reducing intravenous use. Other studies have confirmed these findings 39.40. With regard to the question of how methadone maintenance effects a reduction of injecting, Ball et al38 stress that both the short-term pharmacological and the long-term rehabilitative aspects of methadone maintenance (i.e., counseling and other support services) are important, with high methadone doses being a necessary, but not sufficient factor, whereas rehabilitation is the sufficient factor. It can therefore be expected that a program, which provides high doses of methadone, prohibits the illicit use or 'injecting of drugs and provides - or intensifies - long-term counseling and other support ser vices, will diminish injecting, and thereby sharing.

It needs to be noted that the findings reported in chapter 4 (and the results obtained in the study of incident HIV infections32) should be interpreted with some caution, since methadone data obtained in these studies are based on self-reports, which may be unreliable. For further studies of the effect of methadone use on injecting, sharing and HIV risk, it seems important to validate selfreport data, for example by comparison with data from the Central Methadone Registration. However, these data may also be biased, as methadone pills obtained through and registered by the program are sometimes sold in the streets41. To test the validity of the above formulated conclusion, a possible study design would be random allocation of drug users applying for a new or a reintake in the low threshold methadone program to either the regular scheme or to a scheme such as described above. However, this would neglect the role of motivation. Another option is to offer the high dose methadone maintenance scheme to those addicts who wish to stop injecting: in the sample of HIV-negative current IDUs studied in chapter 7 this wish was reported by 16%.

9.4 Exchange program attenders versus non attenders

The first study (chapter 2) indicated that especially frequent long-term injectors participate in the exchange program. It was hypothesized that participation in the exchange program may not be exclusively dependent on the motivation to reduce one's risk. For example, as the cost of injection equipment is higher for frequent injectors, financial reasons may induce frequent injectors to participate in the exchange program more often than infrequent injectors. A later study (chapter 6) confirmed the earlier findings: especially frequent long-term injectors participated in the exchange program.

Contrary to common stereotypes, heroin users do not use heroin continuously4 2. The present research (chapter 6) corroborates this. Long-term heroin users reported a longest continuous period of nondependence (outside an institution) of, on average, 13 months. Results of the study reported in chapter 2 suggested that persons who are in a stopping or starting phase of their (intravenous) drug use may have objections against saving used equipment or against having a supply of new equipment. This was later partly confirmed: 16% of HIV-negative injectors reported the wish to stop injecting, and this wish was especially present among persons not participating in the exchange program (see chapter 7). These findings thus support the earlier formulated hypothesis: participation in the exchange program is not only dependent on a motivation to reduce one's risks, but also on financial motives and drug use characteristics, or on combinations of all three.

In chapter 2, exchange program attenders were found to inject more safely then nonattenders. However, in the later studies, these differences became less clear: in one study (chapter 7), a relation was found between participation in the exchange program and safe injecting, while another study (chapter 6) failed to find a significant relation, although the differences were in the expected direction. In an earlier study of risk reduction among participants in the cohort-study, Van den Hoek et al.", conclude that the exchange program is an important starting point, but in itself not enough to produce the necessary drastic change in risk behavior; intensive counseling is also needed. This conclusion was partly based on the finding that, at intake in the cohort study, one form of needle sharing, borrowing used needles and syringes, did not decrease over the years 1985-1988, while the number of needles and syringes spread through the exchange program increased greatly in that same period. However, recent evidence suggests that drug users participating in the cohort study enter because of their increased risk'°. This may go together with a relatively high and consistent level of borrowing reported at intake. Thus, self-selection of participants in the co hort study may have confounded the expected impact of the exchange program on borrowing. Another aspect of needle sharing, lending used needles and syringes to others, decreased over time, while single syringe use, indicating the availability of equipment, increased. These changes were not merely attributable to study participation (i.e., to received counseling), but were also related to time, indicating behavioral changes due to the impact of the exchange program. A recent study of incident HIV infections 3 2 failed to find a protective effect of 100% participation in the Amsterdam ex change program. However, these data suggested a time effect: exchanging of needles and syringes was protective at the beginning of the program in 1986/1987. Time effects were also suggested by two other studies, one in Amsterdam, the other in Great Britain. In Amsterdam, hepatitis B incidence declined rapidly over the years 1986-1989 "3 . However, with regard to HIV incidence, an initial decline was found among injectors from 1986 to 1987, and after that a stabilized incidence until 1989. In England, find ings indicate a general decline in injecting risk behavior since 1987. This decline was strongest among non-exchangers, resulting in less difference between exchangers and non-exchangers over time".

The present findings, in relation to the results obtained in the other mentioned studies, can be explained as follows. In 19841985, health education messages about the severity of HIV and about ways of transmission reached injectors who were most motivated to reduce their risks. These were the first clients of the exchange program. It can be assumed that the provision of health education and the distribution of new needles and syringes by the exchange program also indirectly influenced non-attenders. Although not studied in Amsterdam, English findings suggest that many attenders give new needles and syringes to friends if needed "s. Gradually, health education messages reached all injectors, either through the exchange program or through other prevention activities, while at the same time, the availability of free new needles and syringes increased tenfold from 1985 to 1991. At present, the motivation for risk reduction may not differ between attenders and nonattenders, as opposed to the situation at the start of the program. Furthermore, the findings indicate that frequent injectors generally exchange, while infrequent injectors mainly buy. This means that access to needles and syringes, relative to the amount and cost of those needed, is not greatly different for attenders as compared to non-attenders. To conclude, the exchange program makes one million needles and syringes freely available each year. Through taking away financial barriers, the response costs of adaptive behavior are diminished. The available evidence seems to indicate that the exchange program has contributed to reductions in injecting risk behavior and thus should be considered an important HIV prevention measure. However, the present findings also suggest that the program is especially attractive to frequent, long-term injectors and less attractive to injectors who wish to stop injecting. Since exchange program attenders differ from non-attenders, this implies that the impact of the exchange program on injecting risk behavior can no longer reliably be assessed by studying differences in risk behavior between attenders and nonattenders. It seems at present more relevant to study determinants of injecting risk behavior, both among attenders and nonattenders.



9.5 Determinants of injecting risk behavior

No indications were found for an increase in cocaine injecting or in daily injecting over the years 1985-1989 (chapter 3). Injecting cocaine by itself was rare; most IDUs injected heroin and cocaine together ("speedball"). Heroin smoking became more prevalent over the years, which, with respect to HIV prevention, can be interpreted as a favorable development. Smoking or inhaling of cocaine base (in chapter 3 indicated as cocaine freebasing) is similar to the use of crack. This behavior also became more prevalent over the studied years, but was, contrary to U.S. findings, not related to risky sexual behavior or to an increased HIV prevalence. It can be concluded that, among Amsterdam drug users, needle sharing remains the major risk factor for HIV transmission.

In line with other studies, a high level of (relatively mild) psychopathology was found among subjects. Similar to results of McKegney et al.", HIV-positive drug users did not have higher levels of stress or psychopathology than HIV-negatives. Contrary to expectations, HIV-positives with psychopathology and/or stress did not report higher levels of lending, but lower. This could possibly be related to the fact that the employed GHQ-30 measures whether respondents experience themselves as problematic and instable. Among IDUs with high GHQ-30 scores, self-esteem maintenance may cause underreporting of lending. Among HIV-negative injectors, psychopathology, but not stress, was associated with more HIV-risk injecting behavior. This is in accordance with a study by Metzger et al. 46 among methadonetreated patients, who found that patients who continue to share needles are a more disturbed subgroup of the methadone maintenance population.

The study concerning demographic and drug use factors and HIV risk injecting behavior (chapter 6) focused on HIV-negative injectors. Although most subjects were aware of their serostatus, the main point of this analysis was to find out whether HIV-negatives with certain characteristics, regardless of knowledge of serostatus, were at increased risk. For the study of protection motivation (chapter 7), a sample was selected of HIV-seronegatives aware of their serostatus, since perceived risk can be expected to depend on this knowledge. Results of this last study show that perception of risk varies widely among HIVnegative subjects aware of their serostatus, and is related to perceived self-efficacy and to previous and subsequent risk behavior. Studies of the effect of counseling and testing on HIV risk behavior"' find little substantial behavioral risk reduction, and Higgins et al.47 suggest that this is possibly the result of too little counseling. However, the present findings suggest that risk behavior is also dependent on perception of risk. Therefore, it seems important to take cognitive factors into account in studies of the effect of testing and counseling on HIV risk behavior. Table 2 presents an overview of factors (apart from participation in the methadone or exchange program) studied as potential associates of HIV risk injecting behavior (because of earlier findings or theoretical assumptions) which were not confirmed in the present research. Several of these factors, notably a German nationality, time living in Amsterdam, duration of injecting and mainly speedball injecting, have been found as independent risk factors for HIV infection in IDUs (see Van den Hoek et al4 8 and chapter 4). Time living in Amsterdam is also associated with a higher HIV seroconversion rate 32. This suggests that, rather than being related to higher levels of risk behavior, these factors are related to the earlier introduction of the virus in subgroups of injectors, to network characteristics or to other factors mentioned in section 9.2.

Factors which could be interpreted as predictive of HIV risk injecting behavior, and which were confirmed in at least one study, were the following: male sex and young age (chapter 2), psychopathology (chapter 5), previous borrowing, current cocaine injecting, long term moderate-to-heavy alcohol use, having no permanent housing (chapter 6) and a high perceived vulnerability to HIV infection (chapter 7).

In the first study (chapter 2), an indication was found that especially individuals who like to take risks (i.e., young and male drug users) inject unsafely more often. This was not replicated in the later studies, which may be due to differences in sample selection (see 1.8). Several of the other determinants of injecting risk behavior that were found are interrelated. Psychopathology among addicts has been found to be associated with alcoholism" 9-53. The alcohol use variable employed in the present research cannot be equated with alcoholism. However, it can be assumed that most alcoholics will fall under the category of long term moderate-to-heavy alcohol users. Excessive alcohol use among heroin users is associated with health risks, i.e. with an increased risk for a fatal overdose and with an increased general mortality54. Polydrug use and psychopathology also appear to be related 55,57. Korf and Hoogenhout41 suggest that the increase in polydrug use (especially tranquilizers and sleeping pills) among Amsterdam addicts from about 1984 on, may parallel an increase of these substances in the general population. However, according to van Brussel et al. 5 8 polydrug use existed from the beginning of the heroin epidemic in the Netherlands around 1972, but the abused substances changed over the years (e.g., from amphetamines in the seventies to cocaine in the late eighties). They see polydrug use as an attempt at self-medication"'. Van Brussel et al. report that multiple overdoses have been found to occur especially among polydrug users (of sleeping pills and alcohol) and that in many of these cases, psychiatric diagnosis and treatment seemed indicated. Results of a study by Joseph et alb° among homosexual men indicate that a high perception of risk with regard to HIV infection is associated with unsafe sexual behavior and with psychological and social distress. In the present research (chapter 7), we also found a relation between a high perceived vulnerability and unsafe injecting, which may be the result of feelings of hopelessness and/or fatalism. Finally, homelessness has been found related to psychopathology61.

To summarize, results of the present research indicate that unsafe injectors are especially alcohol and polydrug users, and subjects with (more) psychological and social (i.e. homelessness) problems. The results of a number of studies show that these characteristics are often related. In the introduction (section 1.6), drug use was described as a gradually failing attempt to cope with stressors. Evidence suggested that ineffective coping with the HIV threat could especially be expected among drug users with coexisting psychopathology. The present findings with regard to unsafe injecting behavior confirm this. Polydrug users or drug users with most psychological and social problems may share needles because of limited temptation coping skills, or because they have more problems in coping with general life strains, for example with managing to have new needle and syringes available at all times. Barriers of time and place in relation to the availability of needles and syringes can be expected to be especially problematic for these groups. Findings from chapter 7 suggest that perceptions of the HIV threat are realistic and based on previous risk behavior. However, a substantial number of IDUs appear to have given up coping adequately with the HIV threat, possibly as a result of feelings of hopelessness and/or fatalism. While the present unsafe injector may care about his or her health, the severity of the addiction together with psychological and social problems seem to determine behavior and lead to apparently 'irrational' actions. Thus, the presently existing risk behavior appears to be under limited volitional control. These findings may partly explain that, despite the early introduction of the exchange program in Amsterdam, the HIV incidence rate among participants in the cohortstudy has remained relatively high; in 1989 it was 5.5 per 100 person-years (95% confidence interval 1.7-9.4) 43.

Therefore, parallel to the "harm reduction" approach with regard to hard drug use in general, a similar approach, which makes little demand on drug users, appears at present to be most suitable with regard to HIV prevention. If the main obstacle for some injectors is the carrying around of needles and syringes, or having a supply of them at home, then the provision of bleach may be considered as an alternative, provided that these injectors do not have objections against carrying bleach or having bleach at home. As high levels of perceived risk may lead to maladaptive behavior, too much emphasis on risk seems undesirable. Health education, in particular counseling on safe use, may employ relapse prevention techniques62 to improve temptation coping skills. Other new prevention approaches may need to be developed to prevent the residual HIV risk injecting behavior. These approaches should focus in particular on IDUs who borrowed previously, on IDUs with moderate-to-heavy alcohol use, on cocaine injectors and on IDUs without permanent housing.

IDUs who do not participate in the exchange program appear to have good reasons for not participating, such as injecting infrequently, or the wish to stop injecting. This implies that trying to reach all injectors with an exchange program is not an adequate policy. Rather, facilities for buying needles and syringes, like pharmacies or vending machines, should be well spread throughout the various city districts, and vending machines should be in operation during the hours in which addicts are mainly active. For participants in the exchange program, exchange facilities should be optimized in the same way. It does not seem sensible to take away financial barriers, while at the same time leaving barriers of place and time intact.

9.6 HIV-seropositive drug users

Among HIV-seropositive current injectors, 8% (chapter 5) and 4% (chapter 8) reported lending used needles to others during a period of approximately four months. IDUs with psychopathology or high levels of stress did not report more lending than others. As discussed earlier (in section 9.2), lending may be underreported, due to a tendency to give socially desirable answers. Unsafe vaginal sex (chapter 8) was more prevalent. It was reported by 17% of the studied sample, and occurred both with injecting and non-injecting partners. Unsafe sex was not related to cognitive factors, such as attitude and selfefficacy with respect to condom use. Most unsafe vaginal sex was reported by female prostitutes (not only with clients) and by foreign drug users. Both are groups with a marginal social position. Social and cultural factors have often been found to interfere with the adoption of preventive health behaviors.

In general, the studied HIV-positive drug users reported a strong intention to use condoms and not to share needles and syringes. However, almost half of the sample thought they might infect someone with HIV in the future through sex, and one quarter of the current injectors thought this might happen through needle sharing. This suggests that these HIV-positive drug users expected to engage in unsafe sex and needle sharing in the future, despite their good intentions. While for HIV-negative drug users a motivation for self-protection is important, HIV-positive drug users need also to be motivated to protect others. An altruistic attitude may be unrealistic for drug users in a marginal social position, especially when means are scarce and maladaptive behavior is rewarded, financially or otherwise. For these reasons, it seems important, in addition to programs which aim at HIV-negative injectors or sexual partners of injectors, to provide counseling and means (i.e., needles and syringes and condoms) especially to HIVpositive drug users.

9.7 Concluding comments

The high prevalence of unsafe injecting at the start of the HIV epidemic appears to be caused mainly by scarcity of new needles and syringes, together with a lack of awareness or a small perceived threat among injectors of the infections which could be acquired through needle sharing at that time. The changes in injecting risk behavior which have taken place among injectors since 1984 suggest that a substantial amount of effective coping has occurred. When means became more easily available, together with health education, injecting risk behavior under volitional control seems to have changed rapidly. However, injecting risk behavior which is less under volitional control, and which is associated with addiction and with psychological and social problems, has remained. A wide range of measures to further diminish needle sharing has been discussed. It seems worth trying to offer drug users who wish to stop injecting high dose methadone maintenance facilities, and to develop prevention programs specifically for HIV-seropositive drug users. If a further reduction of HIV risk injecting behavior or of HIV incidence among injectors is wished for, it is recommended, in parallel to the harm reduction approach with regard to hard drug use in general, that both the threshold of health education and of other preventive measures such as the provision of needles and syringes are lowered even more than is presently the case.

References

1. Grund JPC, Kaplan CD, Adriaans NFP, Blanken P. Drug sharing and HIV transmission risks: The practice of frontloading in the Dutch injecting drug user population. J Psychactive drugs 1991;23:1-10.

2. Grund JPC, Stern LS, Kaplan CD, Adriaans NFP, Drucker E. Drug use contexts and HIV-consequences: The effect of drug policy on patterns of everyday drug use in Rotterdam and the Bronx. Br J Addict 1992;87:381-92.

3. Grund JPC, Kaplan CD, Adriaans NFP. Needle sharing in the Netherlands: an ethnographic analysis. Am J Public Health 1991;81:1602-7.

4. Agar MH. Into that whole ritual thing: Ritualistic drug use among urban American heroin addicts, pp. 137-148 in: Du Toit BM (Ed.) Drugs, rituals and Altered States of Consciousness. Rotterdam, Balkema 1977.

5. Robertson R. The Edinburgh epidemic: A case study, pp. 95-107 in: J Strang & G Stimson (Eds.) AIDS and drug misuse. London, Routledge 1990.

6. Brettle RP. HIV and harm reduction foar injection drug users. AIDS 1991;5:125-36. 7. Nelson KE, Vlahov D, Cohn S, Lindsay A, Solomon L, Anthony JC. Human Immunodeficiency Virus infection in diabetic intravenous drug users. JAMA 1991;266:2259-61.

8. Becker MH & Joseph JG. AIDS and behavioral change to reduce risk: A review. Am J Public Health 1988;78:394410.

9. Friedman SR & Des Jarlais DC. HIV among drug injectors: The epidemic and the response. AIDS care; 1991;3:239-50.

10. Hartgers C, Santen GW van, Haastrecht HJA van, Hoek JAR van den, Akker R van den, Coutinho RA. De HIV-prevalentie onder druggebruikers die methadon verstrekt krijgen bij de drugsafdeling van de GG&GD to Amsterdam. T Soc Gezondheidsz 1992;70:275-279.

11. Hoek van den JAR, Haastrecht van HJA, Coutinho RA. Risk reduction among intravenous drug users in Amsterdam under the influence of AIDS. Am J Public Health 1989;79:1355-7.

12. Bradburn NM, Rips U, Shevell SK. Answering autobiographical questions: The impact of memory and inference on surveys. Science 1987;236:157-61.

13. Samuels JF, Vlahov D, Anthony JC, Chaisson RE. Measurement of HIV risk behaviors among intravenous drug users. Br J Addict 1992;87:417-28.

14. McKegney FP, O'Dowd MA, Feiner C, Selwyn P, Drucker E, Friedland GH. A prospective comparison of neuropsychologic function in HIVseropositive and seronegative methadone-maintained patients. AIDS 1990;4:565-9.

15. Crowne DT & Marlowe D. The approval motive. New York, Wiley 1964.

16. Webb EJ, Campbell DT, Schwartz RD, Sechrest L. Unobtrusive measures: Nonreactive research in social sciences. Chicago, Rand Mcnally, 1966.

17. Anthony JC, Vlahov D, Celentano DD, Menon AS, Margolick JB, Cohn S Nelson KE, Polk BF. Self-report interview data for a study of HIV-1 infection among intravenous drug users: description of methods and preliminary evidence on validity. J Drug Issues 1991;21:739-57.

18. Gibson DR, Guydish JR, Wraxall BGD, Blake ET, Clark G. Using forensic techniques to verify self-reports of needlesharing. AIDS 1991;5:1149- 50.

19. Loimer N, Werner E, Presslich O. Sharing spoons: A risk factor for HIV- 1 infection in Vienna. Br J Addict 1991;86:775-8.

20. Des Jarlais DC, Friedman SR & Strug D. AIDS and needle sharing within the intravenous drug use subculture, pp. 111125 in: D Feldman & T Johnson (Eds.) The social dimensions of AIDS: Methods and theory. New York, Praeger 1986.

21. Schoenbaum EE, Hartel D, Selwyn PA, Klein RS, Davenny K, Rogers M, Feiner C, Friedland, G. Risk factors for human immunodeficiency virus infection in intravenous drug users. N Eng J Med 1989;321:8749.

22 .. Vlahov D, Munoz A, Anthony JC, Cohn S, Celentano DD, Nelson KE. Association of drug injection patterns with antibody to human immunodeficiency virus type 1 among intravenous drug users in Baltimore, Maryland. Am J Epidemiol 1990;132:847-56.

23. Chitwood DD, McCoy CB, Inciardi JA, McBride DC, Comerford M, Trapido E, McCoy V, Page B, Griffin J., Fletcher MA, Ashman MA. HIV seropositivity of needles from shooting galleries in South Florida. Am J Public Health 1990;80: 150-2.

24. Gaughwin MD, Gowans E, Ali R, Burrell C. Bloody needles: the volumes of blood transferred in simulations of needlestick injuries and shared use of syringes for injection of intravenous drugs. AIDS 1991;5:1025-7.

25. Grund JPC, Stern LS. Residual blood in syringes: size and type of syringe are important. AIDS 1991;5:1532-3.

26. Vlahov D, Polk B. Perspectives on infection with HIV-1 among intravenous drug users. Psychopharm Bull 1988;24:325-9.

27. Hoffman PN, Larkin DP, Samuel D. Needlestick and needleshare - The difference. J Inf Dis 1989;160:545.

28. Resnick L, Veren K, Salahuddin SZ, Tondreau S, Markham PD. Stability and inactivation of HTLV-III/LAV under clinical and laboratory environments. JAMA 1986;255:1887-91.

29. Siegel JE, Weinstein MC, Fineberg HV. Bleach programs for preventing AIDS among IV drug users: Modeling the impact of HIV prevalence. Am J Public Health 1991;81:1273-9.

30. Kaplan EH. Needles that kill: Modelling Human Immunodeficiency Virus transmission via shared drug injection equipment in shooting galleries. Rev Infect Dis 1989;11:289-98.

31. Blower S. Behaviour change and stabilization of seroprevalence levels in communities of injecting drug users: correlation or causation? J Acq Immun Def Synd 1991;4:920-4.

32. Ameijden van EJC, Hoek van den JAR, Haastrecht van HJA, Coutinho RA. The harm reduction approach and risk factors for HIV seroconversion in injecting drug users, Amsterdam. Am J Epidemiol July 1992; in press.

33. Epen van JH. De drugs van de wereld de wereld van de drugs. Alphen aan den Rijn/Brussel, Samson Stafleu, 1988.

34. Platt JJ & Labate C. Heroin addiction, Theory, research and treatment. New York, John Wiley & Sons, 1976.

35. Dole VP, Nyswander ME, Kreek MJ. Narcotic blockade. Arch Intern Med 1966;118-304-309.

36. Geerlings PJ. Methadonverstrekking: Hoe en waarom. Pharmaceutisch Weekblad 1982;117:507-13.

37. Noorlander E. De ontwikkeling van de methadonverstrekking in Nederland. MC 1987;42:402-4.

38. Ball JC, Lange WR, Myers CP, Friedman SR. Reducing the risk of AIDS through methadone maintenance treatment. J Health Social Behav 1988; 29:21426.

39. Williams AB, McNelly EA, Williams AE, D'Aquila RT. Methadone maintenance treatment and HIV type 1 seroconversion among injecting drug users. AIDS care 1992;4:35-41.

40. Cooper JR. Ineffective use of psychoactive drugs: Methadone treatment is no exception. JAMA 1992;267281-2.

41. Korf DJ, Hoogenhout HPH. Slaap- en kalmeringsmiddelen op de Amsterdamse drugsmarkt. Pharmaceutisch Weekblad 1989;124:95-100.

42. Robertson JR, Bucknall ABV, Skidmore CA, Roberts JJK, Smith JH. Remission and relapse in heroin users and implications for management: Treatment control or risk reduction. Intern J Addict 1989;24229-47.

43. Haastrecht van HJA, Hoek van den JAR, Bardoux C, Leentvaar-Kuypers A, Coutinho RA. The course of the HIV epidemic among intravenous drug users in Amsterdam, The Netherlands. Am 1 Public Health 1991;81:59-62.

44. Donoghoe MC, Dolan K, Stimson GV. Changes in injector's HIV risk behaviour and syringe supply in UK 1987-1990. Presented at the Vllth International Conference on AIDS, Florence, June 1991.

45. Donoghoe M & Stimson GV. Encourage passing on sterile syringes. (Letter) Druglink 1992;xx(1an/Feb):17.

46. Metzger D, Woody G, De Phillippis D, McLellan AT, O'Brien CP, Platt JJ. Risk factors for needle sharing among methadone-treated patients. Am J Psychiatry 1991;148:636-40.

47. Higgins DL, Galavotti C, O'Reilly KR, Schnell DJ, Moore M, Rugg DL, Johnson R. Evidence for the effects of HIV antibody counseling and testing on risk behaviors. JAMA 1991;266:2419-29.

48. Hoek van den JAR, Coutinho RA, Haastrecht van HJA, Zadelhoff van AW, Goudsmit J. Prevalence and risk factors of HIV infections among drug users and drug using prostitutes in Amsterdam. AIDS 1988;2:55-60.

49. Kosten TR, Rounsaville BJ. Psychopathology in opioid addicts. Psych Clin North America 1986;9:515-32.

50. Roszell DK, Calsyn DA, Chaney EF. Alcohol use and psychopathology in opioid addicts on methadone maintenance. Am 1 Drug Alcohol Abuse 1986; 12:269-78.

51. Barr HL, Cohen A. Abusers of alcohol and narcotics: Who are they? Internat J Addict 1987;22:525-41.

52. Weiss RD, Mirin SM, Griffin ML, Michael JL. A comparison of alcoholic and nonalcoholic drug abusers. J Studies on Alcohol 1988;49:510-5.

53. Ross HE, Glaser FB, Germanson T. The prevalence of psychiatric disorders in patients with alcohol and other drug problems. Arch Gen Psychiatry 1988;45:1023-31.

54. Brinkman N. Over dood en dosis bij heroinegebruikers. MC 1987;42:399-401. 55. Gove WR, Geerken M, Hughes M. Drug use and mental health among a representative national sample of young adults. Social Forces 1979;58:572-90.

56. Strain EC, Brooner RK, Bigelow GE. Clustering of multiple substance use and psychiatric diagnoses in opiate addicts. Drug Alc Depend 1991;27:127-34.

57. Magura S, Siddicli Q, Freeman RC, Lipton DS. Cocaine use and help seeking among methadone patients. J Drug Issues 1991;21:617-33.

58. Brussel GHA van, Zadelhoff AW van, Sluys TA. Polydruggebruik - een nieuw verschijnsel? T Alc Drugs 1987;13:198-203.

59. Khantzian EJ. The self-medication hypothesis of addictive disorders: Focus on heroin and cocaine dependence. Am J Psychiatry 1985;142:1259-64.

60. Joseph JG, Montgomery SB, Emmons CA, Kirscht JP, Kessler RC, Ostrow DG, Wortman CB, O'Brien K, Eller M, Eshleman S. J Appl Soc Psychol 1987;17:231-50.

61. Kahn MW, Hannah M, Hinkin C, Montgomery C, Pitz D. Psychopathology on the streets: Psychological assessment of the homeless. Prof Psychology Res Prac 1987;18:580-6.

62. Marlatt GA & Gordon JR. Relapse prevention: Maintenance strategies in the treatment of addictive behaviors. New York, The Guilford Press 1985