harm reduction, rational addiction, and the optimal prescribing of illegal drugs

8
HARM REDUCTION, RATIONAL ADDICTION, AND THE OPTIMAL PRESCRIBING OF ILLEGAL DRUGS RICHARD STEVENSON* Harm reduction (HR) policy is designed to reduce the social, medical and economic cost of illegal drug use to users and to society at large. It is the most important recenf development in international drug policy. However, in the United States and elsewhere, some fear that H R facilities such as needle exchanges and methadone maintenance programs may increase drug consumption and the risk of addiction. This is most likely if users are rational, forward-looking utility maximizers as in the Becker-Murphy model (1988). The tendency for HR policy to increase drug consumption may be strong or weak, depending on the type of drug and the elasticity of demand with respect to the full cost of drug use. The model predicts that the effects will differ between social groups and that there is an optimal prescribing policy for addicts who legally receive drugs in treatment programs. Evidence suggests that clinicians tend to prescribe sub-optimally. I. INTRODUCTION:HARM REDUCTION POLICY Harm reduction (HR) policy is de- signed to reduce the private and social cost of illegal drug use within the frame- work of existing law. It is distinct from decriminalization or legalization, which would require de jure or de facto changes in national and international law. HR pol- icy also differs from coercive demand and supply-side policies, which indirectly aim to reduce drug damage by suppressing il- legal markets. HR policy tackles public health issues directly by seeking to improve the welfare of individual users. Many strategies have been proposed, but the principal ingredi- ents of most programs are educational and advisory services, syringe exchanges, and treatment and maintenance services. Some *Lecturer, Department of Economics, Liverpool University, England. This is a revised version of a paper presented at the Western Economic Association 68th International Conference, Lake Tahoe, Nev., June 24,1993 in a session organized by Jack W. Osman and C. Daniel Vencill, San Francisco State University. The author is grateful to referees for comments and suggestions. Contemporary Economic Policy Vol. XU, July 1994 programs also employ outreach workers to contact drug users most at risk from HIV infection and other health hazards. Demand reduction is one HR policy ob- jective, but abstinence is not regarded as a realistic short-term goal for most heavily dependent users. The policy therefore pro- ceeds in a pragmatic fashion through a hi- erarchy of more achievable objectives. Non-users are urged to abstain. Users are advised to reduce doses and to avoid the most dangerous substances and the riskier means of ingestion. A user who insists on injecting is offered advice on safe tech- niques. Needle sharing is strongly dis- couraged, but those who persist are taught to clean equipment and urged to reduce the number of people with whom they share equipment (Newcombe, 1992). At all stages, users are encouraged to enter treat- ment programs in which clinicians offer addiction therapy and may prescribe a maintenance allowance of drugs. Abbreviations I RA: Rational addiction HR: Harm reduction 101 @Western Economic Association International

Upload: richard-stevenson

Post on 30-Sep-2016

213 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: HARM REDUCTION, RATIONAL ADDICTION, AND THE OPTIMAL PRESCRIBING OF ILLEGAL DRUGS

HARM REDUCTION, RATIONAL ADDICTION, AND THE OPTIMAL PRESCRIBING OF ILLEGAL DRUGS

RICHARD STEVENSON*

Harm reduction (HR) policy is designed to reduce the social, medical and economic cost of illegal drug use to users and to society at large. It is the most important recenf development in international drug policy. However, in the United States and elsewhere, some fear that H R facilities such as needle exchanges and methadone maintenance programs may increase drug consumption and the risk of addiction. This is most likely if users are rational, forward-looking utility maximizers as in the Becker-Murphy model (1988). The tendency for HR policy to increase drug consumption may be strong or weak, depending on the type of drug and the elasticity of demand with respect to the full cost of drug use. The model predicts that the effects will differ between social groups and that there is an optimal prescribing policy for addicts who legally receive drugs in treatment programs. Evidence suggests that clinicians tend to prescribe sub-optimally.

I. INTRODUCTION: HARM REDUCTION POLICY

Harm reduction (HR) policy is de- signed to reduce the private and social cost of illegal drug use within the frame- work of existing law. It is distinct from decriminalization or legalization, which would require de jure or de facto changes in national and international law. HR pol- icy also differs from coercive demand and supply-side policies, which indirectly aim to reduce drug damage by suppressing il- legal markets.

HR policy tackles public health issues directly by seeking to improve the welfare of individual users. Many strategies have been proposed, but the principal ingredi- ents of most programs are educational and advisory services, syringe exchanges, and treatment and maintenance services. Some

*Lecturer, Department of Economics, Liverpool University, England. This is a revised version of a paper presented at the Western Economic Association 68th International Conference, Lake Tahoe, Nev., June 24,1993 in a session organized by Jack W. Osman and C. Daniel Vencill, San Francisco State University. The author is grateful to referees for comments and suggestions.

Contemporary Economic Policy Vol. XU, July 1994

programs also employ outreach workers to contact drug users most at risk from HIV infection and other health hazards.

Demand reduction is one HR policy ob- jective, but abstinence is not regarded as a realistic short-term goal for most heavily dependent users. The policy therefore pro- ceeds in a pragmatic fashion through a hi- erarchy of more achievable objectives. Non-users are urged to abstain. Users are advised to reduce doses and to avoid the most dangerous substances and the riskier means of ingestion. A user who insists on injecting is offered advice on safe tech- niques. Needle sharing is strongly dis- couraged, but those who persist are taught to clean equipment and urged to reduce the number of people with whom they share equipment (Newcombe, 1992). At all stages, users are encouraged to enter treat- ment programs in which clinicians offer addiction therapy and may prescribe a maintenance allowance of drugs.

Abbreviations I RA: Rational addiction HR: Harm reduction

101

@Western Economic Association International

Page 2: HARM REDUCTION, RATIONAL ADDICTION, AND THE OPTIMAL PRESCRIBING OF ILLEGAL DRUGS

102 CONTEMPORARY ECONOMIC POLICY

In these ways, HR policy aims to reduce the private cost of drug use. This cost in- cludes poverty, social dislocation, ill- health, and the risk of legal penalties. It is expected that improvements in users’ wel- fare also will reduce the costs of illegal drug use, which spill over to society at large in the form of risks to public health and drug related crime.

The ideas are not new, but they were presented as a coherent philosophy to- wards the end of the 1980s as an alterna- tive to the “War on Drugs.” The HR move- ment is strongly influenced by drug policy in The Netherlands and by experience on Merseyside in northwest England. Merseyside owes its reputation in drug policy to a remarkable record of containing some of the worst problems associated with drug abuse. The region has the largest concentration of notified addicts in the United Kingdom-about nine times the national average. An exceptionally high proportion of users-perhaps 50 percent- inject intravenously. Liverpool, the largest city on Merseyside, is a port served by perhaps 1,000 prostitutes. One might ex- pect that this would be a fertile environ- ment for the spread of HIV infection. How- ever, the region has the second lowest number of notified HIV+ injecting users and the second lowest number of cases of AIDS in the country. Moreover, in 1990, Mersey was the only region to report a reduction in reported crime (OHare et al., 1992).

Until the discovery of AIDS, HR policy was resisted on the grounds that it might seem to condone or actually encourage drug use. Of particular concern was the dependence of some HR strategies on po- lice cooperation. It would not, for in- stance, be possible to operate needle ex- changes if police forces were to insist on enforcing the law strictly in the vicinity of clinics. These concerns remain, and many HR strategies are controversial, or illegal, in the United States and in other countries.

HR policies that reduce drug-related crime and risks to public health are unam- biguously beneficial provided that pro- gram costs do not exhaust the expected benefits and that they are politically ac- ceptable. However, HR policies aim to im- prove users’ health, and drugs may be provided free or at a reduced price. The fear is that HR policy may increase the demand for drugs and the risk of addic- tion by reducing the pecuniary and non- pecuniary cost of a drug habit.

The plausibility of this claim depends on preferences between models of addic- tion. If drug users are thought to be ill-in- formed and irrational, the availability of HR services would seem unlikely to influ- ence their consumption decisions. Some economists claim that the disease model of addiction implies irrationality:

The current approach to the drug wars is imperilled by its refusal to acknowl- edge addiction as a disease rather than a crime ... Addicts (and alcoholics) sim- ply don‘t respond to incentives the way rational people do, because they are not-in this dimension-rational. By the very definition of the term, an addict has “given himself up” to his drug (Benjamin and Miller, 1991, p. 54).

A more conventional economic approach supposes that users are well-informed and take account of the risks of addiction in choosing to use drugs. If drug users are in this sense rational, HR policy will tend to increase drug consumption.

11. THE RATIONAL ADDICTION MODEL

The theory of rational addiction (RA) proposed by Becker and Murphy (1988) applies to legal and illegal drugs. It as- sumes that drug users are rational, for- ward looking utility maximizers who base consumption decisions on full knowledge of the consequences of addiction. Addic- tion in the rational model is defined as a state characterized by reinforcement and tolerance. The reinforcement effect is re-

Page 3: HARM REDUCTION, RATIONAL ADDICTION, AND THE OPTIMAL PRESCRIBING OF ILLEGAL DRUGS

STEVENSON: HARM REDUCTION AND OPTIMAL PRESCRIBING OF ILLEGAL DRUGS 103

sponsible for the “craving for drugs.” Cur- rent consumption increases the marginal utility of future use. Tolerance to drugs increases with past consumption and re- duces the marginal utility to be expected from a given dose in a future period. The rational user takes account of the addictive properties of drugs and their implications for future health and wealth. If the short- term benefits exceed the discounted cost of the long-term consequences, a rational user may choose to become addicted.

The RA model supposes that for given prices and income, consumers with a con- stant rate of time preference will allocate income to maximize an intertemporal util- ity function of the form:

where U(t) is utility at time t, C(t) is the quantity consumed of an addictive good at time t, y(t) is the quantity consumed of a non-addictive good, and S(t) is a stock of ”addictive capital.” A decision to con- sume a unit of C in the present period adds to the stock of addictive capital and increases demand in future periods through the reinforcement effect. The size of the stock grows with current consump- tion but decays at a natural rate, 6, since after a period of abstinence, withdrawal symptoms eventually will disappear. Life- cycle experiences also influence the size of the stock of addictive capital. Association with drug-using friends may increase the propensity to become addicted, as might disagreeable experiences. Happy experi- ences-a new job, marriage, or the birth of child-may reduce the stock of addic- tive capital.

The RA model predicts that user’s be- havior will depend on the relation be- tween the size of the stock of addictive capital and the level of current consump- tion, as figure 1 shows (Becker et al., 1991). The ray from the origin shows all points where C(f)=6S. These are steady states

where current consumption exactly matches the rate of depreciation of S, so that the user remains in equilibrium on a stable dose.

Curve X, shows the relation between C(t) and S. Where X, is above the steady- state line, consumption and the stock of addictive capital rise over time. If X, is below the steady-state line, current con- sumption is insufficient to offset the natu- ral depreciation of the stock, so both C(t) and S fall. The propensity to become ad- dicted depends on a user’s initial endow- ment of addictive capital. If C = 0 and S is less than or equal to So, a life-cycle event or curiosity may induce a non-user to con- sume the addictive good, but consump- tion is insufficient to maintain the capital stock, and the user eventually will return to abstinence. This is the experience of most people who try drugs but do not be- come addicted.

Other possibilities exist. Considering a user in equilibrium at A or B. At A, a user consuming a dose C, is in unstable equi- librium. An event that reduces S to less than S, causes the user eventually to ab- stain. However, an exogenous event that increases the stock above S, will cause consumption to rise. This increases S, and C rises still further until the higher steady- state equilibrium is reached at €3. The equi- librium at B is stable. Current consump- tion might fall from C , but unless it is reduced to below C,, consumption will rise back to equilibrium at B. If consump- tion temporarily rises above C,, the user will return over time to high level equilib- rium at B.

The existence of unstable states seems to model addicts’ unstable behavior quite accurately. As most slimmers and cigarette smokers will be aware, attempts to reduce consumption gradually from high levels frequently fail. Similarly, users of illegal drugs often go through periods of absti- nence, but if the stock of addictive capital

Page 4: HARM REDUCTION, RATIONAL ADDICTION, AND THE OPTIMAL PRESCRIBING OF ILLEGAL DRUGS

104 CONTEMPORARY ECONOMIC POLICY

is not sufficiently depleted, a small ”cele- bratory’’ dose can turn into a ”binge,” and the user can become addicted again.

These broad features of the RA model seem applicable to a wide range of behav- ior, including addiction to legal sub- stances (alcohol and tobacco), some activ- ities (gambling and perhaps eating), and most illegal psychoactive substances. However, illegal addictions differ from legal addictions in ways that complicate the model. In the general RA model, the budget constraint depends on the prices of addictive and non-addictive goods and on income derived from an initial stock of assets and discounted expected earnings (Becker and Murphy, 1988, p. 677). In the decision to consume illegal drugs, further distinctions are important.

For illegal drug users, transaction costs are likely to be highly significant. The ra- tional addict takes account of the “full market price” of drugs-that is, “street price” plus the extra cost and risk associ- ated with dealing in illicit markets. These include search costs, uncertain product quality, and the risk of legal penalties. 11- legal drug users also run greater than nor- mal risks of personal violence involving for drugs or money. Furthermore, illegal addictions are much more likely than legal addictions to be financed from illegal earnings and assets, so a further refine- ment would separate earnings and assets according to their sources, whether legal or illegal.

These considerations do not necessarily alter the RA model’s predictions but widen the model to demonstrate that the pharmacological steady-state condition, C(t) = bS, is a partial equilibrium likely to be disturbed by unpredictable variations in the quality of drugs and periods of im- prisonment. For an addict dependent en- tirely on street drugs, a closer approxima- tion to full equilibrium requires mainte- nance on a stable dose with illegal income and assets equal to zero.

111. HARM REDUCTION IN THE RA MODEL

The rational addict will take account of the existence of HR services that reduce the morbidity and premature mortality as- sociated with illegal drug use. The result is a reduction in the full cost of drug use. Figure 1 represents this reduction by a shift in the relationship between C(t) and S from XI to X,.

Users who were initially in the region between A and B are propelled by the re- inforcement effect to high level steady- state at D. Similarly, some users with a lower addictive stock who previously were not prone to addiction will become dependent on drugs and move to stable or unstable equilibrium at C or D.

The tendency for HR policy to increase drug use may be strong or weak according to the properties of individual drugs and the elasticity of demand with respect to a change in full cost. If the demand for drugs is highly inelastic, HR policy will have relatively small influence on the amount of drug use. However, the RA model suggests that demand may be more elastic than is sometimes supposed due to the complementarity that exists between the demand for current and future drugs through the reinforcement effect.

A user, stable at A on X,, initially re- sponds to a reduction in full cost by mov- ing vertically from A to A, . This short-run effect could be small, but in the longer term, the reinforcement effect moves the user along X, to steady-state at D. The combination of the moves from A to A, plus the move from A, to D could repre- sent a large increase in drug use.

IV. THE ROLE OF INFORMATION

Information services are an important ingredient of HR programs, but the effect of improved information is as uncertain in the RA model as it is in real life. Govern- ments have supposed that the provision of information on the hazards of drug use

Page 5: HARM REDUCTION, RATIONAL ADDICTION, AND THE OPTIMAL PRESCRIBING OF ILLEGAL DRUGS

STEVENSON: HARM REDUCTION AND OPTIMAL PRESCRIBING OF ILLEGAL DRUGS 105

FIGURE 1 Relation between Drug Consumption and the Stock of Addictive Capital

CONSUMPTION 1 Cc=6S

STOCK s 2

will tend to reduce demand. However, drug educationalists are increasingly skeptical (Pearson et al., 1990; Dorn and South, 1990). Mass advertising, blanket condemnations, and simple exhortations ("Just say, No!") are now believed to be ineffective and even counterproductive. Drug professionals prefer to provide non- judgmental, factual information to indi- viduals and small groups. The impact of this sort of information, if any, is hard to assess. Some purely factual information could persuade a prospective user that il- legal drugs are not as dangerous as might be supposed. Injecting need not be dan- gerous if it is performed skillfully in hy- gienic conditions. Heroin, unlike some legal drugs, is not known to damage vital organs. Intravenous heroin users have been maintained on stable doses for more than 40 years without serious effects on their general health. One clinician has stated that opiates are not intrinsically dangerous but become dangerous when

prohibited by law (Ellard, 1989). Long- term addiction to cocaine is relatively rare and less is known about its consequences. However, a patient in Britain was main- tained in good health on legally prescribed cocaine for 55 years (Brown and Mid- dlefell, 1989). Therefore, rational consum- ers might choose to increase drug con- sumption in response to improved infor- mation flows.

The impact of information might be ex- pected to differ between social groups ac- cording to time preferences and levels of income and education. Studies of cigarette smoking suggest that younger, worse ed- ucated social groups discount future costs heavily and respond more to changes in prices than to health warnings. Better ed- ucated groups with higher expected future earnings and lower rates of time prefer- ence are more sensitive to information on the health hazards of smoking (Townsend, 1987). If these findings are transferable to illegal drug users, the knowledge that

Page 6: HARM REDUCTION, RATIONAL ADDICTION, AND THE OPTIMAL PRESCRIBING OF ILLEGAL DRUGS

106 CONTEMPORARY ECONOMIC POLICY

drug use need not be so very dangerous would have the strongest influence on bet- ter educated, higher income users and po- tential users.

V. PRESCRIBING

If an addict depends on the street mar- ket for supplies, the pharmacological steady state is a precarious, partial equi- librium prone to disturbance by changes in income, prices, product quality, and pe- riods of imprisonment. One way of stabi- lizing users pharmacologically and legally is for clinicians to prescribe maintenance doses of illegal substances.

The controlled availability of clean drugs to dependent users has seemed at- tractive to economists and others as an in- stance of reverse price discrimination (Moore, 1972). Prescribing relieves addicts from the need to commit crime, while the high price in the illegal market deters new users. Competition between maintenance programs and the illegal market also has beneficial supply-side implications. A small program on Merseyside that pre- scribed heroin and other addictive drugs (rather than methadone) is said to have displaced the illegal market entirely (Marks, 1991). In The Netherlands, the quality of street drugs is said to have im- proved in response to the competition from free methadone.

Maintenance prescribing has been a part of the British system of drug control since 1926 when the Rolleston Committee recommended that users unable or un- ready to abstain should receive free drugs under medic a1 supervision . Approxi- mately 25,000 users are registered as ad- dicts and are eligible to receive free drugs from the British National Health Service. This medical approach to illegal drugs by and large has prevailed in the United Kingdom, but controversy remains over the quantities and types of drugs that should be prescribed. In the United King- dom, clinicians in receipt of a Home Office license can prescribe heroin, cocaine,

crack, amphetamines, and other drugs of addiction in smokeable and injectable forms but rarely do so. Approximately 98 percent of registered users receive oral methadone (Griffin, 1992).

It is improbable that a simple model can fully capture the subtleties of treatment and prescribing. However, figure 1 sug- gests that addiction therapy should aim to reduce the stock of addictive capital to a level below the unstable equilibrium at A so that the user ultimately will abstain. If treatment fails, a second-best strategy might be to prescribe the quantity and type of drug that will stabilize the user at an equilibrium such as B.

The effectiveness of prescribing de- pends on the type and dose of the drug and on the terms and conditions of offer. The provision of free drugs modifies but does not break the income constraint on the choice between addictive and non-ad- dictive goods. If prescription drugs are un- attractive, some users will stay out of treatment. If doses are inadequate, the ad- dict will benefit from having more income to spend on non-addictive goods but will need to allocate some income to street drugs to remain in pharmacological steady-state. However, even if the addict receives sufficient doses of desired drugs, prescription drugs remain expensive in terms of transactions costs. In some pro- grams, clients must consume the drug on the premises, so frequent attendance at clinics is required. Drugs may be pre- scribed on a take home basis, but prescrip- tions for more than two weeks' supply rarely are issued. In all cases, the user faces significant travel and time costs that reduce earnings or employment prospects. Furthermore, the user must bargain for ad- equate doses and is subjected to the indig- nity of urine tests, close questioning on private matters, and lectures on the merits of abstinence.

All of these factors may play a part in determining low take-up and retention rates of drug programs in the United King-

Page 7: HARM REDUCTION, RATIONAL ADDICTION, AND THE OPTIMAL PRESCRIBING OF ILLEGAL DRUGS

STEVENSON. HARM REDUCTION AND OPTIMAL PRESCRIBING OF ILLEGAL DRUGS 107

dom and elsewhere. U.K. community studies show that only about 10 per cent of seriously dependent users are regis- tered with the Home Office and eligible to receive controlled substances on prescrip- tion (Hartnoll et al., 1985; Parker et al., 1987). The other 90 percent of users prefer the hazards and high price of the illegal market to free drugs offered by the Na- tional Health Service. One must presume that most users judge the costs ("hassle," waiting time, travel costs, and loss of earn- ings) to exceed the expected benefits (probability of effective therapy plus oral methadone).

In a small study in Nottingham, all users in treatment expressed dissatisfac- tion with the service received from clinics. Of those not in treatment, 76 percent cited the type and quantity of drugs prescribed as reasons for not registering as dependent users (Whynes et al., 1989). Oral metha- done, described in Liverpool as "divvy gear" -suitable only for children or those of weak intellect-is not acceptable to many users, especially injectors who are not attempting to abstain.

Clinicians also may systematically under prescribe. In the United Kingdom, analyzing prescriptions of controlled sub- stances by dosage is not possible, but the modal dose of oral methadone is not greater than 35mg daily. Many users are known to top-up their legal rations with street drugs, a practice which suggests that prescriptions are too small to achieve a steady-state. Heroin maintenance is un- common, but clinicians are reluctant to prescribe doses in excess of 100mg. Most heavy users are stabilized on a daily con- sumption of 300-400mg, although a study of re-addicted ex-addicts showed that doses leveled off after several months at 600-750mg daily (Parry, 1992; Chapple and Somekh, 1970).

Evidence from the United States is much less impressionistic. Daily metha- done doses in the range 50-100mg are the most effective in reducing intravenous in-

jecting, maintaining addicts in good health, and reducing criminal activity. Clinics offering smaller doses have lower retention rates, and the drop-outs are far more likely to have recourse to street her- oin and to die of AIDS (Schuster, 1989).

VI. DISCUSSION AND POLICY IMPLICATIONS

Harm reduction radically departs from the strict prohibitionist position that takes abstinence as its principal objective. The main advantage is that HR offers a way round the impasse that drug policy has reached. The strict application of national and international law has not been effec- tive in dealing with drug problems, but changes in the law, which would decrim- inalize or legalize psychoactive sub- stances, remain politically unattractive. By concentrating on politically less sensitive public health issues, HR policy makes it possible to improve the management of dependent users.

HR is the most significant recent devel- opment in international drug policy, but the RA model's predictions are not en- tirely favorable to the movement. Some HR strategies have clear benefits to users and non-users, and all programs offer treatment that, to the extent that it is ef- fective, will tend to reduce the stock of addictive capital and drug consumption. The RA model also provides a theoretical basis for maintenance prescribing. How- ever, it predicts that demand elasticities might be higher than has been previously supposed and that drug use could in- crease, especially in higher socio-eco- nomic groups.

How realistic is the RA model? Perhaps most fundamentally, the model will be re- ceived well or skeptically according to perceptions of the nature of drug addicts. Plainly, many addicts live chaotic lives and cannot be supposed capable of the elaborate ratiocination that the rational model implies. However, our knowledge of users is strongly biased by the small sample who are sufficiently incompetent,

Page 8: HARM REDUCTION, RATIONAL ADDICTION, AND THE OPTIMAL PRESCRIBING OF ILLEGAL DRUGS

108 CONTEMPORARY ECONOMIC POLICY

unlucky, or desperate to come to the atten- tion of the medical or legal authorities. Others live relatively normal lives, have jobs, and remain in reasonably good health. Dependent users may not be en- tirely happy in their circumstances, but at any time no more than 5 percent may be trying to give up drugs. Some clinicians have remarked on the high proportion of clients who are intelligent, articulate, and well-informed. In the Nottingham study, for example, users seemed capable of looking ahead and considering alternative expenditure allocations.

By and large, the RA model provides a good basis for theorizing, and its predic- tions need not be very damaging to the HR movement. HR policy may induce some extra drug use. This would be regret- table, but well-managed drug users do not impose large costs on society in the form of criminal activity or public health risks. Furthermore, although most HR programs have not been fully evaluated, they seem to offer substantial benefits.

The high cost of care for AIDS victims suggests that one can expect high benefit- cost ratios for needle exchanges if only a few cases are averted in a year (Decker and Rosenfeld, 1992). More is known about maintenance programs. Experience in The Netherlands suggests that metha- done maintenance has not increased drug use or addiction, and similar claims have been made on Merseyside for both meth- adone and heroin maintenance (Marks, 1991). In Merseyside, special local factors may have been at work, but success in containing the HIV virus suggests that there is merit in attracting users into HR programmes by offering drugs on terms and conditions that are more acceptable to users. In general, some American cities could benefit from a more receptive ap- proach to HR policy.

REFERENCES Becker, Gary S., and Kevin M. Murphy, “A Theory of

Rational Addiction,” Journal of Political Economy,

Becker, Gary S., Michael Grossman, and Kevin M. Murphy, “Rational Addiction and the Effect of Price on Consumption,“ American Economic Re- view, Papers and Proceeding of the AEA Meeting, May 1991,237-241.

Benjamin, Daniel K., and Roger L. Miller, Undoing Drugs: Beyond Legalization, Basic Books, New York, 1991.

Brown, R., and R. Middlefell, ”Fifty-five Years of Co- caine Dependence,” British Journal of Addiction, 84,1989,946.

Chapple, P. A. L., and D. E. Somekh, “Treatment of Drug Addiction,” The Lancet, November 28,1970, 1,134.

Decker, Scott, and Richard Rosenfeld, “Intravenous Drug Use and the AIDS Epidemic: Findings from a 20-City Sample of Arrestees,” Crime and Delin- quency, October, 1992,492-509.

Dom, Nicholas, and Nigel South, “Communications, Education, Drugs and €-IN,” in John Strang and Gerry Stimson, eds., AIDS and Drug Misuse, Routledge, London, 1990.

Ellard, John, ”The Drug Offensive,” Modern Medicine of Australia, December 1989.

Griffin, Jane, Drug Misuse, Office of Health Economics, London, 1992.

Hartnoll, R. L, R. J. Lewis, M. Mitcheson, and S. Bryer, “Estimating the Prevalence of Opioid Depen- dence,“ The Lancet, 26,1985, 203-205.

Marks, John A., “The North Wind and the Sun” Pro- ceedings Royal College of Physicians, Edinburgh, 22,

Moore, M. H., “Policies to achieve discrimination on the Effective Price of Heroin,” American Economic Review, 63:2,1972,270-277.

Newcombe, Russell, “The Reduction of Drug-related Harm: A Conceptual Framework for Theory, Practice and Research,” in P. A. OHare, R. New- combe, A. Matthews, E. C. Buning, and E. Druc- ker, eds., The Reduction of Drug-Related Harm, Routledge, London, 1992.

OHare, I? A., R. Newcombe, A. Matthews, E. C. Bun- ing, and E. Drucker, eds., The Reduction of Drug- Related Harm, Routledge, London, 1992.

Parker, H., K. Bakx, and R. Newcombe, “The New Her- oin Users: Prevalence and Characteristics in Wir-

96:4,1988, 675-700.

1992,13-17.

ral, Merseyside” British Iournal of Addiction, 82, 1987, 147-158.

Parry, A l k ~ , “Taking Heroin Maintenance Seriously: The Politics of Tolerance,” The Lancet, 339,1992, 350.

Pearson, G., M. Gilman, and P. Traynor, ”The Limits of Intervention,” Druglink, May/June 1990,lZ-13.

Schuster, Charles R., ”Methadone Maintenance,” N l D A Notes, Spring/Summer 1989.

Townsend, Joy L., ”Cigarette Tax, Economic Welfare and Social Class Patterns of Smoking,” Applied Economics, 19, 1987, 355465.

Whynes, David K., Philip T. Bean, John A. Giggs, and Christine Wilkinson, “Managing Drug Use,” Brit- ish Journal of Addiction, 84, 1989, 533-540.