School of Public Health, University of Michigan, Ann Arbor, Michigan, United States of America
Substance-abuse treatment should be used to prevent new infections
Treatment for substance abuse should be held to high standards as a health-care intervention
The criminal justice system should be used to prevent and treat disease
For decades, infectious diseases have posed a serious and avoidable threat to the health and survival of injection drug users (IDUs). The most deadly threat currently arises from the human immunodeficiency virus (HIV) and the acquired immunodeficiency syndrome (AIDS). Yet less visible infectious diseases are also significant. Data from local and population-based surveys indicate that the great majority of IDUs in the United States of America are infected with hepatitis B or hepatitis C virus [1, 2]. A smaller but still significant proportion is infected with endocarditis, bone and joint infections , tuberculosis  and other infectious ailments. As the average age of IDUs increases, the prevalence of infectious disease may increase .
What can be done to address those threats? Can epidemiological models play a useful role in public health prevention efforts? Can treatment for substance abuse slow the spread of disease, and is treatment cost-effective when compared with other public health efforts? Can hepatitus C virus or other highly infectious diseases be controlled within drug-using populations? How can clinicians and policy makers target IDUs at greatest risk of getting and spreading blood-borne disease? The present paper offers five tentative principles to help answer those difficult questions.
Many clinicians and policy makers are skeptical about the merits of formal models of the spread of infectious disease. Some skepticism is warranted. Injection drug users (IDUs) are a hidden population whose risk behaviours, and even whose absolute numbers, are only imperfectly known  Many basic parameters, such as the probability of infection from a single exposure, are poorly understood and differ across the drug-using population. The baseline rate of new infections is difficult to estimate within many high-risk groups. Rigorous evaluations of prevention are rare. When such evaluations do exist, they may not be applicable to local conditions. Even when sound epidemiological and programme evaluation data are available, this information may not be easily translated into politically and administratively feasible policy guidance.
Thus, the first question any policy analyst confronts may well be the hardest: Why bother? Although that question may be answered in many ways, the simplest and best answer is that models reveal and scrutinize the implicit assumptions already being used to fight infectious disease . Rules of thumb often reflect unexamined assumptions that turn out to be implausible .
Kaplan and Pollack reviewed procedures used to allocate resources for human immunodeficiency virus (HIV) prevention in the United States [8-9]. Many United States policy makers try to allocate resources on the basis of the number of individuals in each risk group. That method of allocation is sensible when individual risk groups are similar and when the available prevention procedures are similarly effective across risk groups. When those assumptions go wrong, population-based resource allocation fails to provide an appropriate means of targeting resources to those in greatest need. Moreover, such allocation fails to prevent the largest number of new infections, given the resources available for interventions aimed at prevention . In the real world, the incidence of HIV and the effectiveness of available programmes across risk groups vary greatly.
Worse, the political and organizational realities of group decision-making foster arbitrary decision-making. Altman, Greene and Sapolsky note that health planners respond to technical and political uncertainty by seeking “convenient proxies for need to be applied in allocation decisions”. Wary of debating the merits of specific facilities, health system planners often resort to drawing up need-assessment formulas to evaluate proposed capital-intensive services. Economists have never doubted that such formulas provide poor guidance in evaluating the impact or cost-effectiveness of proposed expenditures. However, such methods find wide appeal as planners seek credible focal points to resolve internal disputes and justify controversial policies. For many of the same reasons, Kaplan and Pollack found the widespread application of complex and analytically ungrounded formulas to allocate funds for HIV prevention.
Models also help policy makers understand the linkage between readily observed data and the less visible, underlying pattern of the spread of disease. In facilitating the process of understanding, explicit models indicate which data have the most important influence on policy and how such data might be interpreted.
Evaluations of syringe exchange programmes (SEPs) illustrate the importance of such models. Although SEPs have not been scrutinized through prospective randomized trials, an impressive literature documents the effectiveness of such interventions [12-17]. A United States Secretary of Health and Human Services recently acknowledged that point, although she was not able to lift the current ban on United States federal funding [18, 19]. Although many studies have examined the impact of SEPs on self-reported behaviour change, such studies did not exa mine the relationship between SEPs and the rate of new HIV infections. Kaplan and Heimer addressed this critical policy concern by developing an elegant mathematical model to interpret HIV test data from exchanged needles .
Unfortunately, transmission modelling also indicates that the success of SEPs in curbing HIV transmission may not be replicated for other diseases. Short-term reduction in HIV transmission is sufficient to reduce long-run incidence and prevalence because HIV, though deadly, is difficult to transmit. The analytic work of the author of the present paper and existing programme data suggest that the same programme performance is not enough to prevent the spread of more efficiently transmitted infectious diseases [20-22].
Apart from HIV, hepatitus C virus (HCV) is the most serious blood-borne epidemic among IDUs. So far, no intervention for prevention or treatment has been consistently found to slow the spread of the disease. Most discouragingly, many IDU populations display endemic HCV prevalence despite well-implemented SEPs and methadone maintenance treatment that successfully contain HIV [23-25]. Policy analysis has not reached a stage where it can offer reliable programme recommendations. In contrast to HIV, for which successful targeting of high-risk individuals may offer the greatest long-term benefit, the most effective prevention strategies for HCV might include more substantial education and outreach directed at low-risk IDUs.
For analytic models to be of use, accurate and pertinent information is needed concerning both the spread of infectious disease and the performance of available interventions.
One of the most important but difficult tasks is to obtain usable (if imperfect) estimates of disease incidence, that is, the number and overall pattern of new infections in the drug-using population. Those estimates are important, because existing prevention measures are often based upon current disease prevalence, that is, the number of IDUs known to be infected. Ignoring many complexities, researchers can easily measure prevalence when infected individuals seek medical attention. In many settings, disease prevalence is estimated through the use of existing clinical data systems at hospitals, public clinics and other sites where health care is provided.
Unfortunately, this method of resource allocation is often inappropriate. Current prevalence indicates the past history of the spread of disease. In a dynamic blood-borne epidemic, prevalence estimates can provide a misleading guide to the specific risk-groups that are currently experiencing high rates of new infection. In the area of HCV, incidence analysis indicates that young and inexperienced IDUs are experiencing remarkably high rates of new infection [1, 22].
Many policy makers in public health are well aware that disease incidence is the touchstone of prevention efforts. The practical challenge is to estimate disease incidence in a hidden population frequently estranged from the health-care delivery system. Both technology and improved public management can play an essential role in meeting this challenge. In the area of HIV prevention, new statistical and chemical testing methods have become available to enable the rate of new infection in specific populations to be directly estimated [26, 27]. Many of these methods are easily transferred to other blood-borne ailments. Other advanced techniques such as molecular tracing are also helpful in tracing outbreaks and documenting novel modes of transmission.
Apart from technological innovation, the most important challenge is to treat the surveillance of infectious disease as a central tool of public health policy rather than as the stepchild of existing convenient data systems . Because many IDUs have limited contact with the health-care delivery system or with public health services, incidence and prevalence estimates may be biased if they are drawn from HIV/AIDS treatment, voluntary testing and other services that reach small, often self-selected, populations. Such data must be supplemented by new sources that include more hidden populations of IDUs.
Careful sampling of IDUs in other contexts is therefore essential to provide an accurate picture of disease incidence and prevalence in the drug-using populations. Drug-treatment clients and SEP clients are critical populations that bear careful study in population-based surveillance. Out-of-treatment IDUs may also be identified through epidemiological studies of prison inmates [28, 29]. Perhaps most important is to explore the rate of new infections among out-of-treatment IDUs who do not use other services . Many of them are probationers or are under other forms of judicial supervision .
Treatment for substance abuse plays a critical role in improving the well-being and social performance of drug users. Many studies document the effectiveness, and cost-effectiveness, of methadone maintenance and other forms of treatment in reducing criminal activity, improving health and increasing employment among treatment clients [32-36].
The value of treatment was underscored by a randomized study of methadone maintenance among Swedish IDUs [32, 37]. Within two years of the inception of the study, 2 out of 17 members of the non-methadone control group died from apparent overdose. One other member of the non-methadone group suffered a leg amputation, while two others suffered severe infection. Two of the remaining members of the non-methadone group were incarcerated, and 9 of the remaining 10 continued illicit drug use. Only one was healthy and off drugs. Over the same period, none of the methadone treatment group suffered major health problems. Thirteen of the original 17 were no longing using illicit drugs. Three more members of the non-methadone group died in the following three years, in a study completed before the era of HIV/AIDS .
More recently, treatment for substance abuse appears to lead to a significant reduction in HIV seroconversion among treatment clients. Metzger and colleagues document a sixfold difference in HIV incidence between methadone clients and out-of-treatment IDUs . More recently, analytic models also indicate that methadone maintenance treatment would be highly cost-effective even if its only benefits would be to prevent HIV infection [20, 39, 40]. Using analytic models that examine highly imperfect treatment programmes with high relapse rates, the author of the present paper found that methadone maintenance treatment would cost between 150,000 and 300,000 United States dollars ($) per averted HIV infection in a high-risk population. These estimates are below lifetime estimated treatment costs  and compare favourably with widely accepted interventions to improve and to extend human life .
Methadone maintenance treatment and other approaches perform three different and complementary functions to prevent the spread of HIV and other blood-borne diseases.
First and foremost, drug treatment reduces the likelihood and the frequency of injecting drug use. Those modalities are hardly perfect. Substance abuse is a chronic, relapsing condition. Not surprisingly, many clients relapse or fail to comply with recommended treatment. Yet because HIV is difficult to transmit, modest and imperfect programmes can have a large impact on the spread of disease . Though beyond the scope of the present paper, best-practice treatment includes sufficiently high methadone dosage  and the provision of accompanying social services to reduce behavioural risks [45, 46].
In addition to reducing drug use, treatment for substance abuse can also include harm reduction elements such as instruction on the proper use of bleach. Because most treatment clients will engage in some level of future drug use, such instruction may have a substantial impact on the spread of infectious disease. These approaches are controversial. Some treatment professionals view instruction on “safer injection” as implicitly condoning injecting drug use . The author of the present paper considers that harm reduction services are essential to protect treatment clients, given the realities of widespread relapse and non-adherence in most treatment settings.
Treatment for substance abuse plays a third role in providing a venue for identifying and delivering medical treatment for infected IDUs. Most substance-abuse treatment facilities in the United States currently provide routine HIV counselling and testing . Many, though by no means all, provide effective linkages to required medical care . Although the impact of such interventions on the spread of disease is not well known, it is plausible that such linkages reduce the severity of the disease and the likelihood of further transmission.
Despite the above-mentioned achievements, treatment for substance abuse would be much more effective if its core mission included the provision of high-quality health care. Many treatment providers describe substance abuse as a treatable medical condition. Yet few drug treatment facilities provide high-quality care when judged by best-practice standards of medical care applied in other settings. Many facilities do not provide basic physical examinations, reproductive health services or infectious disease screening for highly prevalent conditions among drug users.
The author of the present paper recently visited a large and reputable United States provider of methadone maintenance. The facility used a high-technology management system to provide accurate billing and to meet the exacting requirements of government regulators. Yet the same facility dispensed methadone in assembly-line fashion with rudimentary medical and social services to a high-risk population. Quite openly, the facility provided minimal services to uninsured adults for whom it received low rate of reimbursement.
A critical problem is to provide effective treatment for those drug users at greatest risk of spreading and contracting infectious disease. That is most obvious in the area of HIV prevention, where IDUs with dual-diagnosis psychiatric disorders, unmarried young men, those with intense criminal justice involvement, those in poverty and polydrug users are more likely to contract or to transmit HIV . Precisely the same characteristics are correlated with non-adherence and premature exit from treatment for substance abuse . Treatment providers—especially those evaluated on the basis of mean relapse rates and other measures—have strong incentives to select the most cooperative and successful clients and to avoid those most at risk of HIV infection.
More generally, public policies must recognize the unique role of injecting drug use. Although marijuana users and other, relatively low-risk, drug users require appropriate care, resource allocation for substance-abuse treatment must reflect the public health reality that curbing injecting drug use is essential to control blood-borne disease.
Drug treatment facilities are probably much more effective in dealing with HIV than they are in addressing prevention and treatment of other blood-borne diseases. For example, IDUs face high risks of contracting hepatitis B virus (HBV). Many treatment facilities provide rudimentary screening and treatment, despite strong evidence of the efficacy and effectiveness of HBV vaccination among IDUs [51, 52].
Paradoxically, services to address other blood-borne ailments may be especially important in areas of low or declining rates of new HIV infection. Because HIV provides unique motivation to avoid needle-sharing and other high-risk behaviours, additional services may be required to address behavioural risks involving other serious blood-borne ailments that are more infectious, if less prominent, than HIV itself. For that reason, policy makers must be open to the possibility that there is some trade-off between policies that minimize the spread of HIV and those that minimize the spread of other infectious diseases.
SEPs reflect some of the same limitations that are common within substance abuse treatment. Many SEPs have important limitations that hamper both their effectiveness and the political sustainability of this controversial intervention.
For SEPs to curb more infectious agents successfully, programme quality must be improved. Some SEPs have already added drug treatment referrals, case management for HIV prevention and other services that go beyond the distribution of sterile equipment . Such linkages would provide a valuable means of bringing high-risk individuals into treatment. In that connection, analysis by Kim Blankenship of Yale University (United States) indicates that HIV-negative IDUs are ineligible for some of the most important preventive services.
Finally, efforts to control infectious disease must find more effective ways to identify and to serve out-of-treatment IDUs who face the greatest risk of disease. As first documented by Hammett and others for the United States, a large proportion of IDUs can be found in prisons and other correctional settings . Seventeen per cent of all AIDS cases in the United States, 36 per cent of tuberculosis cases and 29 per cent of HCV cases occur within the incarcerated population . Many of the diagnosed inmates are IDUs.
Despite the well-documented need for substance-abuse and health-care services, correctional health-care and substance-abuse treatment are frequently unavailable or sub-standard. A 1991 report of the United States General Accounting Office indicates that federal prisons provide appropriate treatment to only 1 per cent of inmates with significant drug problems . Dramatic incidents of medical misconduct highlight the darker possibilities of correctional care [31, 55, 56].
Even larger numbers of IDUs are under supervision of the criminal justice system in the general community—on probation, parole or pretrial release or in other arrangements. More rigorous supervision, referrals and more effective social and medical services are essential to reduce substance use and to improve the well-being of this population . Research by Pollack, Khoshnood and Altice proposes a multifaceted strategy of entitlement security, case management, outreach and prison discharge planning to address those concerns [28, 31]. Recent innovations such as “coerced abstinence” and graduated sanctions may also be effective in deterring relapse among criminal offenders .
The AIDS epidemic provides the inevitable context in evaluating public health interventions for IDUs. Like any powerful historical experience, HIV presents powerful lessons that should be heeded in confronting other blood-borne epidemics.
One clear lesson is that disease prevention requires effective public management and carefully applied epidemiology. The late and inadequate United States response to HIV reflected public indifference to a problem afflicting IDUs and other stigmatized groups . Yet that policy failure also reflected the failure of disease surveillance systems to detect HIV prevalence among IDUs quickly, the failure of policy makers to fund and implement substance abuse treatment adequately and the failure of health-care delivery and criminal justice systems to implement best-practice HIV prevention among men and women who face the greatest behavioural risks .
Many of the policy failures could have been avoided through improved public management and through the decisive political leadership required to enact controversial policies. One task of such leadership is to teach a skeptical electorate that public health measures for IDUs are sound public investments. When methadone maintenance treatment or SEPs are evaluated by widely accepted standards, they compare quite favourably with other public health interventions . It would be naive to think that academic studies are sufficient to mobilize political support for unpopular policies. But it would also be wrong to discount the importance of judicious analysis in providing policy makers with the data—at times the political cover—to take unpopular steps. Scientific consensus supporting the effectiveness of SEPs has been cited by a former Mayor of New York City, a former United States Secretary of Health and Human Services and other policy makers. More recently, politicians and commentators across the political spectrum cite evidence that supply-side cocaine enforcement is less cost-effective than interventions for prevention and treatment interventions .
The HIV epidemic also serves as a reminder that IDUs are members of the broader community, and are not some incorrigible group beyond help or justified concern. Although IDUs engage in illegal, and sometimes destructive, behaviours, HIV prevention efforts indicate that IDUs possess unexpected capacities to help themselves and others in avoiding deadly disease. Those capacities are an important asset for future interventions [12, 60].
One of the most important lessons is that effective interventions must include elements of both prevention and treatment to be fully effective. To attract clients and to reduce long-term risk, harm-reduction interventions must include conduits to substance-abuse treatment and medical and social services. To provide lasting protection for a client population prone to both relapse and non-adherence to treatment, traditional treatment for substance abuse must include preventive measures such as instruction in safer injecting. For diseases that can be prevented by vaccination, such as HBV, SEPs and substance-abuse treatment also provide opportunities for more permanent prevention. The prevention component of treatment will be equally important for ailments that involve the prospect of recovery and subsequent infection.
Like any powerful experience, HIV teaches some lessons that may prove misleading in addressing other epidemics. Like generals preparing to fight the last war, prevention specialists who uncritically generalize from HIV may therefore be cruelly disappointed. HIV is a uniquely feared infectious agent, is currently impervious to cure or vaccination, and is rather difficult to transmit. Infectious diseases that do not share those characteristics are likely to require different interventions for treatment and prevention.
Opponents of SEPs have long argued that harm reduction measures to make substance use safer are an inadequate response to the individual and social harms associated with injecting drug use. Those critics are wrong about HIV, but they have a more persuasive case concerning more easily transmitted agents that are more difficult to control through imperfect behavioural interventions.
Put differently, the spread of HCV and other infectious agents provides a painful reminder of both the necessity and the limitations of harm reduction. Measures such as SEPs reduce the risk of disease and provide critical outreach to engage the hidden population of IDUs. Those modest interventions are not enough to protect men and women in an environment of persistent and frequent needle-sharing and other behavioural risks. Future harm-reduction measures may therefore require greater and more explicit use reduction than is currently implemen ted in SEPs. Measures such as SEPs must be further expanded to address other concerns, including overdoses and the management of chronic disease in an aging population of IDUs.
Clinicians and policy makers must also realize that the unique nature of the HIV epidemic created new opportunities that are unlikely to be replicated in the fight against other diseases. Among those at personal risk, fear is a powerful motivator. HIV therefore produced remarkable behaviour change among many IDUs who are probably less fearful of other blood-borne diseases . A similar pattern has been observed among gay and bisexual men, some of whom have resumed high-risk behaviours in response to declining local HIV incidence and the overall development of more effective HIV therapies [10, 62].
Within the broader society, HIV prevention is a major and highly visible social issue, one that links the interests of IDUs with those of other citizens at risk of disease. In many countries, the HIV epidemic framed prevention and treatment measures for IDUs as part of a broader national effort to contain a new and frightening infectious disease. HIV greatly increased the amount of public attention and public resources devoted to substance-abuse policy. More obscure and less lethal blood-borne ailments are unlikely to produce such an expensive and focused policy response.
The spread of blood-borne disease is a serious public health threat among IDUs. The threat will remain, even in the happy event that HIV is ever eradicated within that population. Even if the right lessons are learned from HIV, not all problems associated with the spread of blood-borne disease will be solved. But it will be possible to confront them more effectively, and thereby reduce the avoidable suffering being experienced by those who continue to inject illicit drugs.