U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) Committee for the Substance Abuse Coverage Study; Gerstein DR, Harwood HJ, editors. Treating Drug Problems: Volume 1: A Study of the Evolution, Effectiveness, and Financing of Public and Private Drug Treatment Systems. Washington (DC): National Academies Press (US); 1990.

Cover of Treating Drug Problems

Treating Drug Problems: Volume 1: A Study of the Evolution, Effectiveness, and Financing of Public and Private Drug Treatment Systems.

Show details

5The Effectiveness of Treatment

The question that people ask drug treatment experts most often and most insistently is a simple one: Does treatment really work? In the committee's judgment, and that of most experts, the available clinical experience and research data add up to a similarly short and pointed answer: It varies. This answer should be no surprise, as the question is naive. Virtually everything in Chapters 2, 3, and 4 of this report leads one to expect the effectiveness of treatment to be a complicated matter to understand and assess. Drug treatment is not a single entity but a variety of different approaches to different populations and goals. Response to treatment is not a matter of all or nothing, complete success versus total failure, but of degrees of improvement. Moreover, the setting for evaluation is not the quiet purity of a controlled laboratory experiment but the tangled complexity of real lives and programs under pressure from many directions.

The committee's strategy under the circumstances has been to put forward a line of questioning that is straightforward but somewhat more elaborate and revealing than "Does treatment really work?" These questions, which are listed below, cannot all be fully and confidently answered at present. Consequently, they must continue to be asked about each king of treatment.

  • What are the basic concepts or modalities of treatment? That is, what are the underlying designs or theories of treatment, what specific types of drug problems or population groups are being addressed by each design, and what are the best results that have been obtained under ideal conditions?
  • How well does each modality work in practice? How adequate in terms of methodology are the evaluations of real programs, and what do the best of these evaluations reveal?
  • If a modality is not working as well as might be expected, what are the reasons? For example, is the implementation or replication of the modality flawed or incomplete? Are the wrong kinds of clients being treated? Are there unexpected side effects? Does the environment interfere with the effectiveness of the treatment?
  • Do the benefits of the treatment justify its costs? In other words, is treatment a sound investment of scarce public and/or private resources?
  • In addition to these questions about treatment as it presently exists: How might further research help to improve treatment?

In responding to the first of these questions, this chapter considers serially the four major types or modalities of drug treatment: outpatient methadone maintenance, residential, therapeutic communities (TCs), outpatient nonmethadone (OPNM) treatment, and inpatient/outpatient chemical dependency (CD) treatment. As indicated in the brief description of these modalities in Chapter 2, each type of drug treatment has developed since the 1950s. TCs derived largely from Synanon, which began in California in 1958. Methadone maintenance developed from studies on a hospital ward in New York in 1964; CD programs grew out of hospital-based approaches to treating alcoholism in Minnesota in the 1960s. Outpatient nonmethadone treatment1 goes back at least to psychoanalytic treatment of "toxicomania" in the 1930s, but community mental health movement, youth crisis counseling, "drop-in centers," and "free clinics" of the 1960s adopted quite different orientations that have substantially shaped the OPNM programs seen today. Although every modality has specific roots, all have continued to evolve since their introduction.

The most extensive usable results of research on the effectiveness of drug treatment are from several moderately sized clinical experiments and natural or quasi-experiments and from prospective longitudinal studies involving thousands of clients. There have been two large-scale, multisite, federally sponsored studies of publicly supported programs: the 12-year follow-up of a 1969–1971 Drug Abuse Reporting Program (DARP) national admission sample cohort and the Treatment Outcome Prospective Study, or TOPS, which involved a 10,000-person national sample of 1979–1981 admission to 41 drug treatment programs in 10 cities. The Drug Abuse Treatment Outcome Study (DATOS), a third large-scale national prospective study, is scheduled to begin in 1990.

The committee address the paradigmatic questions separately within each modality. Although many treatment seekers try more than one treatment modality over the course of their drug careers (they build up a "treatment career" as well), the average profiles of clients admitted to the major modalities are quite different. Both treatment seekers and treatment programs engage in a great deal of individual selection into which many factors enter. For example, programs are geographically and economically differentiated in their accessibility to various types of potential clients; methadone clinics are relatively low in cost and typically located in innercity areas; chemical dependency units are generally expensive and found in affluent suburbs. The typical demographic and drug-taking patterns of the different modalities populations (a reflection of who stays in treatment from among those who are admitted) are quite distinctive. As a result, one cannot simply compare the performance or results of each modality with the others as if their client populations were interchangeable. Moreover, because some clients move between programs and there is evidence that treatment effects may, in part, be delayed and cumulative, it is hazardous to ascribe all the effects of a treatment episode to that episode alone; adjustments must be made to take prior treatments into account.

The most extensive and scientifically best-developed evidence concerns methadone maintenance. A lower although still suggestive level of evidence is available concerning therapeutic communities and outpatient nonmethadone treatment. The lowest level of evidence is available for chemical dependency. Where the evidence on treatment effectiveness approaches adequacy, its overall tendencies are clear.

  • Treatment reduces the drug consumption and other criminal behavior of a substantial number of people. Clients exhibit their best behavior while activity enrolled in treatment; their behavior is often poorer following treatment than during it, although still better than before admission.
  • There are large variations in effectiveness across programs, which seem to be related to the varying quality of clinical management and competence. Practices in methadone maintenance dosing are a clear instance of this variation; there is also variance owing to differences in the characteristics of the populations being treated, such as the severity of their problems at admission.
  • The length of time in treatment is a very important correlate of outcome; that is, longer treatment episodes yield better outcomes than shorter ones. Retention is presumably related to general program quality and specific client motivation to remain in treatment; however, no predictive treatment motivation test is available, and the role of treatment in facilitating motivation or averting impulsive decisions to "split" from treatment is not yet well understood.
  • The benefits of treatment programs on the whole outweigh their costs, but variations in cost-benefit methodologies and results are great.

It should be noted that except to describe the model, there are virtually no data to answer critical questions regarding independent self-help fellowship groups such as Narcotics Anonymous and Cocaine Anonymous or the Oxford Houses. Although the ideas underlying the Anonymous fellowship were incorporated at the outset into the clinical approaches 2 of TCs and CD programs and clients in these modalities are encouraged to participate in Anonymous meetings, the fellowship have shied away from involvement in formal evaluation protocols. Because drug-related Anonymous groups have been meeting in most cities longer than drug treatment programs have been present, and because they generally welcome individuals who are in treatment as well as those who are not (except that many Anonymous groups are antipathetic to individuals in methadone maintenance), they are in essence a part of the environmental baseline over which the incremental effects of the more formal treatments must be measured.

Two special topics are set slightly apart from the main lines of the chapter: the role of detoxification, which is often carried out in hospital settings, and the effects of treatment that occurs within correctional institutions. In the committee's view, it is not tenable to consider detoxification a treatment modality for the rehabilitation of drug abuse and dependence. Rather, it is a way of moderating some of the effects of overdose or withdrawal, and it may serve as a gateway to treatment. Correctional programs seem to fall largely into one of three types: they are either therapeutic communities, outpatient-type programs whose clients happen to live in prison, or drug law education programs carrying the name of treatment.

The committee considers the need and opportunity for research relevant to treatment effectiveness to be so important that this chapter presents several recommendations for research on treatment methods and services. With recent budget increases for research, there is no overall lack of resources that could be devoted to such studies. Rather, the challenges of treatment-oriented research are arduous and demand certain kinds of commitments that are altogether too easy to slight in the rush to distribute cascades of research funding to more glamorous (e.g., high-technology) research ventures.

Methadone Maintenance

What Is Methadone Maintenance?

Methadone maintenance is a treatment specifically designed for dependence on narcotic analgesics, particularly the narcotic of greatest concern in the United States, heroin.3 The controversies surrounding methadone maintenance4 have made it the subject of literally hundreds of studies. From these studies, including a few vitally important clinical trials, strong evidence has accumulated about the safety and effectiveness of methadone.

The idea is not unfamiliar that a treatment for a chronic health disorder could involve long-term, even permanent pharmacological maintenance using a powerful drug that is nevertheless safe if properly administered. Perhaps the most obvious examples are treatments for endocrine problems: insulin for diabetes, thyroxine for thyroid deficiency. A treatment for chronic mood disorders (manic-depressive cyclothymia) using lithium chloride for long-term maintenance is a psychiatric example. Although methadone maintenance was viewed as revolutionary when it was first developed in the United States, the historical sketches in Chapter 2 and in Courtwright (1990) point toward early twentieth century instances in U.S. cities of morphine maintenance as a treatment for opiate dependence. In Great Britain, heroin maintenance was also practice, although it has largely been replaced there by methadone maintenance. The application of maintenance concepts to the treatment of drug dependence therefore is not medically unusual. But to understand how methadone maintenance operates as a treatment for heroin dependence, three aspects must be stressed: the significance of clinically defined goals, the pharmacological basis of drug substitution, and the embedding of substitution in a broader clinical behavior strategy.


Methadone maintenance cannot be understood apart from the correct stipulation of the major goals of treatment, primarily to reduce illicit drug consumption and other criminal behavior and secondarily to improve productive social behavior and psychological well-being. It is critical that methadone is a legally prescribed drug for the purpose of treating dependence.5 Yet even more critical is that individuals who receive methadone maintenance treatment should reduce their use of illicit drugs and their commission of other crimes (e.g., selling drugs, stealing money, using weapons to obtain funds to support their drug consumption) ideally to zero but at least by an appreciable amount. Improved social productivity and well-being would be important further measures of the effectiveness of methadone maintenance. The goal of ending the licit dependence on methadone itself is well down the list—so that the risk of increased crime or illicit drug use weighs heavily against arbitrary limitation on the duration of methadone maintenance. Nevertheless, this goal has been given much higher priority in many programs, as discussed later in the chapter.


At the base of methadone maintenance is an empirical observation that was made before the biological reasons for it were well understood: all of the effective narcotic analgesics may be substituted for one another with adjustments in dose and route of administration. Substitution is possible because there are similarities in their objective and subjective effects; in particular, in dependent individuals there is parallel or cross-tolerance to elevated doses and cross-suppression of respective withdrawal effects. Key differences involve how quick, how strong, and how long-lasting these actions are; they are also apparent in the precise mixture of effects for each drug.

Cross-dependence is particularly important in detoxification. Most drugs of widespread abuse and dependence (heroin, cocaine, alcohol) act quickly and dramatically and wear off in a matter of hours. By the same token, the associated primary withdrawal syndromes tend to be striking but short; there is usually, however, a somewhat more protracted but less dramatic phase of sustained withdrawal symptoms such as sleep disturbance, agitation, or mild depression. The general approach to detoxification is to moderate the more severe symptoms, often by substituting a long-acting drug, which can then be tapered down to zero, leaving only the lesser symptoms.

Methadone may be prescribed not for maintenance purposes but for a shorter period—three weeks was once standard, although the period may legally extend up to sex months—to moderate withdrawal symptoms. Detoxification generally begins with an escalating dosage to reach a point such that the patient stops using other opiates and withdrawal symptoms are not evident. Then the methadone dose is tapered down to zero. Individual responses vary, but usually this method does not completely suppress withdrawal symptoms during and after the tapering period; rather, it keeps them mild for a time—until the tapering procedure does not provide enough methadone to prevent the more discomfiting withdrawal symptoms. It is common for individuals to drop out of methadone detoxification some time during the second week of a typical three-week planned detoxification period. Sometimes other medications are given during methadone detoxification to manage particular symptoms.

As shown by the long record of experience with detoxification of heroin dependence, those detoxified were universally found to have a very high susceptibility to relapse—usually well in excess of 90 percent of followed cases (see Vaillant, 1973). After detoxification, and often before its procedures had been completed, there was a resumption of craving for opiates. Dole (1988) and others have theorized that the extensive use of opiates may bring on alterations in the brain neurotransmitter/receptor systems affected by opiates, leaving many individuals with a virtually permanent craving that can only be assuaged by drugs of the opiate family.

Methadone has several unusual pharmacological properties that have made it especially suited to a maintenance approach. Unlike many opiates, it is effective orally, a significant advantage in that oral dosing is more hygienic than the needle and more easily titrated than smoke. Because of Methadone's particular pattern of absorption, metabolism, and elimination, a single dose within a train of level doses, in the typical maintenance range of 30 to 100 milligrams per day (mg/day), takes effects gradually and wears off slowly, yielding a fairly even effect across a period of 24 hours or longer. methadone is thus conducive to a regime of single daily maintenance doses, eliminating dramatic subjective or behavioral changes and making it easy for clinician and client to fit into a routinized clinic schedule.6 This pattern is very different from the shorter action and more dramatic highs and lows of heroin, morphine, and most other opiates. The long-term toxic side effects of methadone, as of other opiates if taken in hygienic conditions in controlled doses, are notably benign.

The short-term clinical effects of methadone were first studied at the Lexington addiction research center in the 1950s, and research continued there and elsewhere into the 1960s. Since the mid-1960s, about 1.5 million person-years of methadone maintenance have accumulated in the United States. Not all clients have been closely observed for medical side effects, but the thousands of research cases that have been carefully observed yield a well-documented conclusion:

[P]hysiological and biochemical alterations occur, but there are minimal side effects that are clinically detectable in patients during chronic methadone maintenance treatment. Toxicity related to methadone during chronic treatment is extraordinarily rare. The most important medical consequence of methadone during chronic treatment, in fact, is the marked improvement in general health and nutritional status observed in patients as compared with their status at the time of admission to treatment. (Kreek, 1983:474)

The most common physical complaints during methadone maintenance are insomnia and weight gain, but these are clinically related both to the consumption of other drugs and alcohol (consumption that continues and sometimes increases among a fraction of clients, the size of which varies from program to program) and to preexisting or coexisting abnormalities common in this population and in the general population.

Clinical Behavioral Strategy

In terms of the social history and individual model of drug-seeking behavior reviewed in Chapters 2 and 3, a program of controlled methadone maintenance at an appropriate dose level could have recovery-inducing effects on heroin dependence. These effects may be felt through two paths corresponding to the two most common motivational processes that operate during heroin dependence: pleasure seeking and withdrawal avoidance.

With regard to pleasure seeking, methadone is an effective analgesic. Yet the effect of an accustomed (tolerated) dose is merely a dim echo or reminder of heroin's most intense effects, not so much a "rush" as a reassurance—which may wear better in the long run and is certainly less disruptive in the short run than the euphoric heroin high with its associated itchiness and dreamy nods. There is also a more subtle and perhaps equally valuable effect: if heroin and methadone are both in the body, their active metabolites compete with each other for access to sites of action in the brain. If the methadone dose is high relative to the heroin dose, the latter will not have a vary distinctive effect, and the individual taking methadone will find heroin less rewarding. As a result, the shooting of heroin "over" the methadone may become self-extinguishing.

On the other side of the pharmacological fence, methadone maintenance prevents symptoms of heroin withdrawal, which, although not life-threatening or excruciating, are immiserating (a good parallel is a head cold or a bout of influenza). The critical condition is that the dependent person feeling withdrawal symptoms knows that all of these unwelcome sensations can be banished within minutes with a dose of an opiate. Recurrent withdrawal symptoms stimulate drug-seeking during heroin dependence, and the ability of methadone maintenance to keep them at bay is a major attraction and benefit.

In its initial clinical trials, which began in inpatient settings and then were extended to outpatient sites, methadone maintenance proved capable of stabilizing the psychological functioning of the heroin-dependent individual at a near normal state. Methadone in effect eliminated the alternating phases of euphoria, somnolence, and agitated concern characteristic of the incipient stage of withdrawal from heroin dependence. the clinicians conducting the trials observed that clients on methadone were not obsessed with acquiring the next dose, became interested in the prospects for improving the conventional strands of their lives, and were generally functioning without notable drug impairment or side effects. An individual on methadone was capable of participating in counseling, psychotherapy, and remedial education and training (most of the same rehabilitative services delivered in therapeutic communities and outpatient treatment). This capability was partly the result of the intrinsic pharmacological effects of methadone and partly because, unlike street heroin, it was provided reliably, in legitimate clinical settings, and in reliable doses.

Methadone maintenance was originally defined as the administration of methadone together with rehabilitative and counseling services, and this definition, along with many detailed specifications about facilities and staffing, was built into federal regulations as a required protocol for a licensed methadone maintenance program. These regulations permit methadone to be dispensed only by licensed maintenance or detoxification programs or by hospital pharmacies. (In hospitals, methadone is prescribed mainly for severe postoperative or cancer pain and occasionally for short-term inpatient detoxification.)

Methadone programs are nearly always ambulatory, with daily visits to swallow the methadone dose (usually provided in a 3- to 4-ounce plastic bottle of sweetened, orange-flavored water) in the clinic, except for the traditional Sunday take-home dose. After several months in the program with a "clean" drug-testing record and good compliance with other program requirements such as counseling appointments, clients may regularly take-home one or more days' doses between every-other-day, twice-weekly, or even weekly visits—a revocable range of privileges. Some methadone clients voluntarily reduce their doses to abstinence and conclude treatment after some time; others remain on methadone indefinitely.

The role of counseling is multifold. In the first instance, the design of methadone maintenance programs includes numerous monitoring and adjustment features that stress the need for clients to wean themselves away from street drug-seeking. Program clinics have specific hours for dispensing, counseling, and medical appointments; there are codes of proscribed behavior (e.g., no violence or threats of violence), and monitored drug tests are conducted at random intervals—at least monthly and as often as weekly, although the cost of the tests have led financially strained programs to cut them back to the minimum. Counseling includes the assessment of client attitudes and appearance (important in themselves and as clues to drug behavior) and the gathering of information about employment, family, and criminal activities; counselors offer psychotherapy and individualized social assistance and recognition, depending on their caseloads and their training for such tasks.

In most clinics, counselors participate in staff decisions with regard to changing dose levels, requirements for therapeutic contacts, award and revocation of take-home privileges, and decisions regarding termination from the program. Clinical experiments have studied methadone dosage and behavioral techniques (contingent rewards and sanctions for "dirty" urines and missed and late counseling appointments) as part of the modality's repertoire. The clinical trial literature has demonstrated important success in the use of methadone dosage supplements or decrements and take-home privileges to punish or reward clients for noncompliance with such clinical rules as the proscription on continued drug use and the requirement of cooperation by timely attendance for dispersing, participating in counseling, and paying required fees (Stitzer et al., 1983).

The drawbacks to methadone maintenance have been well recognized since its inception: the client is still at least mildly dependent; the drug reduces heroin craving and stabilizes the individual psychologically but does not necessarily modify or rehabilitate other behavior; clients often still use or abuse and sometimes become dependent on other drugs including alcohol; and it is possible for take-home methadone to be diverted from therapeutic uses and sold to permit the client to buy heroin or other drugs. Moreover, methadone has no direct pharmacological bearing on abuse or dependence on alcohol or other drugs, especially cocaine, which has become such a serious and widespread problem in the 1980s. The important question is this: Does the modality reach its primary goals in enough cases to outweigh these limitations and drawbacks?

How Well Does Methadone Work?

The goals of methadone maintenance—to reduce illicit consumption of heroin and other opiates, to reduce other criminal activity, and to help clients become more socially productive and psychologically stable—constitute a continuum that can be cut at various points to designate "success" versus "failure." At the outset of its use, the modality was specifically targeted toward those who were most severely dependent, as judged by substantial histories of relapse from earlier detoxification episodes (frequently in jail); this commitment was built into the early regulations requiring documentation of at least two years of heroin use and two prior relapses.

Early trials of methadone maintenance in New York (Dole and Nyswander, 1965, 1967; Dole et al., 1966, 1968, 1969) noted two striking findings: the majority of clients would remain in treatment for as long as it was available to them, in substantial contrast to the usual experience in out-patient psychotherapy; and methadone maintenance significantly improved the condition of clients as revealed by studies that considered behavior in the community for periods of several months to several years. Although there was some use of other drugs, including heroin, especially in the first few weeks after admission, such use generally fell off over time, contrasting sharply with the increasing return over time to heroin dependence that was the norm after detoxification or other typical medical or psychiatric treatments. The steadiness of employment increased somewhat, but a much more dramatic change was the sustained reduction in criminal behavior, especially drug trafficking crimes.7

The most convincing results about the efficacy of methadone maintenance—the capacity of the treatment to induce client changes independent of initial selection or motivational effects—come from a handful of clinical experiments that are widely separated in time and place but that consistently yield very distinctive findings. In these studies, heroin-dependent, heavily criminally involved populations who were randomly assigned to methadone maintenance or a control condition (an outpatient nonmethadone modality) demonstrated clinically important and statistically significant differences in favor of methadone on the gauges of drug use, criminal activity, and engagement in socially productive roles such as employment, education, or responsible child rearing.

In the landmark experiment, Dole and colleagues (1969) randomly assigned 32 well-motivated criminal addicts to either a methadone treatment group (N = 16, whom 4 declined treatment before program initiation) or a year-long control/waiting list group (N = 16). Out of the combined control and refuser group (N = 16 + 4 = 20), every individual became re-addicted to heroin soon after release, with 18 individuals returning to jail and the other 2 being lost to the study. At 7 to 10 months after initiation of the study, only 3 of the 12 addicts in the experimental group had been reincarcerated. Furthermore, although 10 of these 12 individuals had used heroin since the program was initiated, for 6 of the 10 this use was limited to the first 3 months of the program.

Gunne and Gronbladh (1984) have also reported a small but persuasive study (Figure 5-1). Thirty-four heroin-dependent individuals applied for admission to the only methadone clinic in a Swedish community; 17 were randomly assigned to methadone maintenance, and 17 were assigned to outpatient nonmethadone treatment (these individuals could not apply for admission to the methadone clinic again for 24 months). Two years later, 12 of the 17 clients on methadone were no longer using illicit drugs; 10 were employed, and 2 were in school. Five still had drug problems, and of these, 2 had been discharged from treatment for severe abuse of sedative-hypnotic drugs. Of the 17 individuals who went into the outpatient nonmethadone program, only 1 was doing well; 2 were dead, 2 were in prison, and the rest had returned to taking heroin. After two years, then, 71 percent of methadone clients were doing well, compared with 6 percent of controls. Five years after the study began, 13 of the methadone clients remained in treatment and were still not using heroin, and 4 had been excluded from treatment because of unremitting drug problems. Among the controls, 9 had applied for and entered methadone maintenance; of these, 8 individuals were not using drugs and were socially productive. Of the 8 controls who did not apply for methadone when eligible, "five are dead (allegedly from overdose), two in prison and one is still drug free" (Gunne and Gronbladh, 1984:211).8

FIGURE 5-1. Clinical trial of methadone maintenance versus outpatient nonmethadone for heroin addiction conducted through the Swedish Methadone Maintenance Program.


Clinical trial of methadone maintenance versus outpatient nonmethadone for heroin addiction conducted through the Swedish Methadone Maintenance Program. Source: Gunne and Gronbladh (1984).

Another perspective on the effectiveness of methadone treatment is offered by the results of several "natural experiments." In one such study, Anglin and McGlothlin (1984) examined the introduction of the methadone maintenance modality to California in 1971, viewing it as a quasi-experimental intervention. They had previously begun a long-term observational study of heroin-dependent individuals who had been apprehended by law enforcement agencies in 1961–1963 (McGlothlin et al., 1977). Drug consumption and criminal involvement in this study population were high just prior to the introduction of methadone, despite the fact that all members of the population had been incarcerated and supervised for several years in the 1960s by the state's Civil Addict Program (CAP). Some of the study population had already stopped using heroin before the introduction of methadone; this group was termed the inactive user sample. Some of the remainder entered methadone treatment when it became available (the methadone sample), and the rest did not (the active user sample; Figures 5-2a, 5-2b, and 5-2c). McGlothlin and colleagues (1977) had found during their earlier study that the active user and methadone samples had reduced drug and crime activity while under CAP supervision but had quickly resumed their prior high activity levels once CAP supervision ended. After the advent of methadone programs in California, a major difference was observed between those in the active user group and the methadone clients, a difference that persisted for at least three years after the introduction of methadone.

Overall, the findings of this natural experiment indicate that a certain proportion of addicts (i.e., the inactive sample) had responded favorably and permanently to a particular form of criminal justice supervision involving specialized prison treatment and intensive parole. Of those who did not, a significant proportion entered methadone maintenance when it became available and responded very well to it (compared with otherwise very similar individuals who did not enter a methadone program): those pursuing methadone maintenance substantially reduced their drug use and criminal activity and (to a lesser degree) increased their employment.

Similar results have been reported in natural experiments involving the limited introduction of publicly supported methadone maintenance programs in a number of California cities and towns in the early 1970s and the subsequent closure of some of these programs for fiscal and political reasons. In cities where methadone maintenance became much less accessible as a result of such closures, former clients as a whole did appreciably less well at the two-year follow-up (in terms of heroin use, other criminal behavior, and, to a lesser degree, employment) than comparison groups in locations where there was continued access to treatment. In cities where public programs closed but private ones opened, those who transferred to the alternative methadone maintenance programs did much better (in terms of staying free of drugs and out of crime) than those who did not or could not continue treatment (McGlothlin and Anglin, 1981; Anglin et al., 1989a). In these as in all other studies, longer retention in methadone was opposed to early attrition from the program was associated with much better results measured by reduced heroin use and other criminal activity.

Why Do the Results of Methadone Treatment Vary?

A significant proportion of methadone maintenance clients do not respond well to treatment, for a variety of reasons relating to the clients themselves and to the programs. This proportion averages about one in four, although there is wide variation from program to program (U.S. General Accounting Office, 1990; Ball et al., 1988). It is clear that some clients who are admitted enter methadone maintenance for purposes other than to receive counseling and other services or to pursue recovery. These clients are not compliant with clinical rules and are less likely than others to be (or become) motivated; most leave treatment after short periods. It is much easier to identify these clients after the fact than before; programs screen out some but not all such clients through pretreatment intake reviews. There are also probably some clients who would recover just as quickly without methadone maintenance, but they choose methadone treatment because it is helpful or attractive in some ways that other treatments (or no treatment) are not. The proportion of such clients is variable—it may be as low as 1 in 20 or as high as 1 in 10. These clients are beneficial in terms of positive program statistics but somewhat exaggerate the degree to which the program is actually generating worthwhile effects.

The largest group of clients is clearly at some point in the middle. The evidence from experimental and quasi-experimental studies clearly points toward the existence of a substantial number of heroin-dependent individuals who perform at least moderately well in response to methadone maintenance and who would do poorly without it, even when other kinds of treatment are available.

There is compelling evidence that program factors such as methadone dosing policies and counselor characteristics affect the behavior of such relatively malleable clients above and beyond any initial differences in motivation. The strongest treatment retention and outcomes (measured as improved social functioning) were seen in the initial methadone clinical trials (Dole and Nyswander, 1965, 1967) and in cohorts admitted to methadone treatment in New York during the pilot stage of developing the treatment (Gearing, 1970, 1974). This phase of history was characterized by careful screening of clients, self-selection by addicts—as a result of admission waiting lists of up to a year—and extensive adjunctive services provided by highly skilled and motivated clinical staff (Lukoff and Kleinman, 1977).

Later evaluations found that retention rates and outcomes were somewhat poorer when the New York programs had reached large-scale operation, were no longer highly selective in admissions, and had reduced their waiting time for admission to a few weeks (Dole, 1971; Dole and Nyswander, 1976; Dole and Joseph, 1978). Some observers attributed this decline to strains on system capacity and the onset of rigid and antipathetic federal regulations in contravention of good clinical practices (Dole and Nyswander, 1976). Kleber (1977:268) has contended that the programs' primary problems were greatly reduced selectiveness in admissions and the shortage of skilled and motivated staff: "it is not surprising that retention rates dropped and the number of urines containing heroin rose. What is surprising is that the figures were not worse than they were."

Program performance (in terms of client retention and continued use of drugs) has also been observed to vary across programs at the same point in time. The Treatment Outcome Prospective Study, for example, showed a large degree of variation in clinically important client outcomes across nine methadone maintenance programs. Twelve-month retention rates averaged 34 percent of admissions, but five programs had low rates of 7 to 25 percent, whereas two programs had rates greater than 50 percent. Regular heroin use by clients at follow-up (approximately three years later) was reported by 21 percent of the entire follow-up sample, but two programs had rates greater than 30 percent, and three had rates of 11 to 14 percent (Hubbard et al., 1989).

FIGURE 5-2. Effects of methadone maintenance in a sample of California ex-parolees who participated in the Civil Addict Program (CAP) measured on three parameters: (a) percentage of time reported as daily narcotic use; (b) percentage of nonincarcerated time spent in criminal activity; and (c) percentage of nonincarcerated time the individual was employed.


Effects of methadone maintenance in a sample of California ex-parolees who participated in the Civil Addict Program (CAP) measured on three parameters: (a) percentage of time reported as daily narcotic use; (b) percentage of nonincarcerated time spent (more...)

Variation in performance has been linked most strongly to variations in methadone dosage policies. Programs that are committed to maintaining low average doses (30-50 mg/day) as a virtual goal of treatment—because of therapeutic philosophy or because state regulators strongly discourage higher doses—are less tolerant of occasional client drug use, missed counseling appointments, and other such treatment lapses, and have markedly lower client retention rates than more tolerant higher dosage programs. This lower tolerance does not, however, act as a stimulant to better client behavior or as a conveyor to move poorly responding clients out and bring in or keep better ones. There is solid, experimentally grounded evidence (see the major review by Hargreaves, 1983, and the associated conclusions of the expert consensus conference; reported in Cooper et al., 1983) that higher dose levels are fundamentally more successful in controlling a client's illicit drug consumption while he or she is in treatment. Although dose levels must necessarily be adjusted according to individual variations in metabolism and size, programs that maintain an overall average dose of 60-100 mg/day yield consistently better results than those averaging less. Doses in excess of 120 mg/day are seldom needed.

The most recent illustration of the importance of dose levels—and the fact that many programs continue to be committed to low-dose regimes in spite of strong evidence against their relative effectiveness—comes from a study reported by Ball and coworkers (Ball et al., 1988; Ball, 1989; see also Dole, 1989). Dramatic differences in client use of opiates and retention in treatment were found among six methadone clinics in three eastern cities studied in 1985-1986 and selected to begin with as well-regarded programs. In the best clinic, urinalysis revealed that 10 percent of enrolled clients in the sample had used drugs intravenously in the month prior to the one-year follow-up. In the two worst clinics, more than 55 percent of clients had used intravenous drugs in the previous month.

Discriminant function analysis found that the most important factor in predicting intravenous drug use was the methadone dose level (Table 5-1). Among clients in treatment from 6 months to 4.5 years the odds of recent heroin consumption decreased at each higher level of methadone. There was also a dose-related decrease in the chances of cocaine use, although the gradient was less steep. (This trend probably has little direct pharmacological cause but instead arises from the generalized behaviors of drug marketing and drug-seeking: those who are actively seeking heroin are more likely to seek out or at least happen upon cocaine while doing so, and vice versa.) The programs with the highest illicit drug consumption among clients not only had low methadone doses but also had high rates of staff turnover and poor relationships between staff and clients. Knowledge of and sensitivity to the clinical significance of appropriate dose levels is probably one sizable element in a constellation of clinical competencies and strategies that contribute to the greater or lesser effectiveness of methadone maintenance programs. There are only rudiments of standards for training, credentialing, continuing education, evaluation, and clinical performance of counselors and other treatment program staff. It is remarkable how few research efforts have focused on this larger area of competence, appropriate training, and different service arrangements in the clinical management of methadone clients. A serendipitous experimental study by McLellan and colleagues (1988), which demonstrated striking differences in counselor effectiveness within the framework of a large, stable, well-regarded methadone maintenance program,9 is a lonely beacon in the literature.

TABLE 5-1. Heroin or Cocaine Consumption of 338 Methadone Clients (in treatment from 6 months to 4.5 years) in Past 30 Days by Methadone Dose.


Heroin or Cocaine Consumption of 338 Methadone Clients (in treatment from 6 months to 4.5 years) in Past 30 Days by Methadone Dose.

Costs and Benefits of Methadone Treatment

Analyses of the economic costs and benefits of methadone maintenance have been derived from a handful of treatment effectiveness studies, and their results are rather sensitive to how these effectiveness studies are interpreted. An early simulation by Maidlow and Berman (1972), for example, concluded that methadone maintenance could yield lifetime benefits to society of $348,000 compared with average treatment costs of $13,200, a benefit/cost ratio of 26 to 1. However, their assumptions about the effectiveness of methadone were overly optimistic. A simulation by Rufener and colleagues (1977a) was more firmly grounded, yielding a smaller but still quite healthy benefit/cost ratio of 4.4 to 1 for a short period of time. Extended over lifetimes this result would not be too disparate with that of Maidlow and Berman; however, the Rufener team's assumptions about effectiveness also appear to be too optimistic.10

Using more realistic effectiveness data—but from only low-dose programs—McGlothlin and Anglin (1981) compared clients who left methadone maintenance when a community clinic was closed in Bakersfield, California, with clients in another community's program, which remained open. For men, the ratio of crime-related economic benefits to treatment costs was 1.7 to 1, over a short, two-year period. Additionally, the continuous treatment group reported significantly higher rates of employment than those who had been closed out of treatment, although this factor was not formally valued in the study. The results for women were contrary but can be considered little more than a preliminary indication because the sample size was too small for statistical stability. A study of a public clinic methadone program closure in San Diego (Anglin et al., 1989a) showed virtually no net economic loss but also no net gain. In this instance, a private methadone program picked up a large proportion of the clients on a self-pay basis.

The most comprehensive examination of economic benefits and costs of drug treatment was performed with data from the TOPS (Harwood et al.,1988). The data included the average cost of a treatment day in methadone programs in 1979 and detailed interview measures of rates of criminal activities in the TOPS sample in the year before treatment, the period in treatment, and the year after discharge (where applicable). The study also factored in estimates of the average cost to society of particular crimes, based on surveys conducted in 1979 by the Bureau of Justice Statistics. The benefits of methadone maintenance treatment in terms of reduced crime-related costs to law-abiding citizens (including the value of stolen goods) were $13 per day compared with the $6 per day average cost of the program. Moreover, multivariate regression analysis found significant benefits in the year following discharge, such that retention for an additional day in treatment was worth $11 per day in delayed benefits. The final benefit/cost ratio was therefore 4 to 1. An alternative and much more conservative cost/benefit model in which only increases in employment (which were limited) rather than reductions in goods stolen (which were much larger) were valued found a cost/benefit ratio of about 1 to 1. Using either model, methadone maintenance pays for itself on the day it is delivered, and posttreatment effects are an economic bonus.


Methadone maintenance is a treatment that is designed for severe dependence on heroin. Prior to admission to a methadone program, the great majority of clients are consuming large amounts of heroin and other illicit drugs and committing predatory crimes (including drug selling) on a daily basis, a behavior pattern usually extending back several years or more. Although methadone is a relatively long-acting narcotic analgesic and produces dependence symptoms, the consumption of a clinically adjusted oral dose yields a steady metabolic level of the drug, produces little if any behavioral or subjective intoxication, and does not impair functioning or generate appreciably morbid side effects. Once such a solid, comfortable level is reached, suppressing the psychophysiological cues that precipitate and reinforce opiate craving, the client is amenable to counseling and related services that can help shift his or her orientation and lifestyle away from drug-seeking and related crime and toward more socially acceptable behaviors.

Methadone maintenance has been the most rigorously studied of all the drug treatment modalities, and the studies have yielded positive results (although some programs have good and others poor client compliance with rules against illicit drug use and criminal activity). Nevertheless, methadone maintenance is a controversial treatment: its critics contend that methadone clients have "merely" switched their dependence to a legally prescribed narcotic and that many clients continue to use heroin and other drugs intermittently and to commit crimes, including the sale of their take-home methadone. In the committee's judgement, these controversies and reservations are neither trivial nor in themselves compelling. The issues are to what extent undesirable behaviors are reduced and positive behaviors increased as a result of methadone maintenance (in comparison with no treatment or with alternative treatment measures) and whether poorly performing programs can be improved. The extensive evaluation literature on methadone maintenance yields the following conclusions:

  • There is strong evidence from clinical trials and similar study designs that heroin-dependent individuals have better outcomes on average (in terms of illicit drug consumption and other criminal behavior) when they are maintained on methadone than when they are not treated at all or are simply detoxified and released, or when methadone is tapered down and terminated as a result of unilateral client request, expulsion from treatment, or program closure.
  • Methadone dosages need to be clinically monitored and individually optimized, but in general most clients have substantially better responses when maintained at the higher rather than lower end of the dose ranges currently being prescribed (up to 100 mg/day).
  • During and after methadone maintenance treatment, criminal behavior declines and employment increases relative to untreated comparison groups, and the utility of these results substantially exceeds the cost of the treatment, especially when both the crime and employment dimensions are considered over an extended time period.

Methadone maintenance is not the answer for every heroin-dependent individual. At any one time, perhaps one-eighth to one-fifth of all individuals who were recently dependent on heroin can be found in a methadone maintenance program.11 This figure could undoubtedly be increased if program quality were optimized, hostile stereotypes of methadone treatment eliminated, and availability extended. When viewed in terms of lifetime prevalence, the number of current heroin-dependent individuals who will at some time enter the portals of methadone is higher, probably 30 to 40 percent. This range, like the preceding figure, is necessarily only an approximation because the research data that could give more precision to these estimates are inadequate, particularly in light of such recent developments as the AIDS epidemic. Nevertheless, in the committee's judgement, an improved network of methadone maintenance clinics might realistically be capable of reaching and dramatically accelerating the recovery of one-third of all those who become dependent on heroin.

Therapeutic Communities

What Is a Therapeutic Community?

The residential therapeutic community, or TC, is a way of defining the nature of individual drug problems as much as a therapeutic approach to the rehabilitation or, more frequently, the habilitation of drug-dependent persons. It is from this understanding that the TC derives its encompassing and intensive approach.

TCs were originally developed to treat the same problem as methadone maintenance programs: the "hard-core" heroin-dependent criminal. The residential TC has a broader perspective, however; it treats individuals who are severely dependent on any illicitly obtained drug or combination of drugs and whose social adjustment to conventional family and occupational responsibilities is severely compromised as a result of drug-seeking—but who were compromised before drugseeking entered the picture. In this context, the specific drug (or more accurately, combination of drugs) represents a sociological fact more than a pharmacological foundation for treatment. In the 1980s, cocaine dependence has overtaken heroin dependence in the TC population. The profile of TC clients is also more demographically diverse than that of the heroin-dependent population. Generally, on average, TC clients in the early 1970s, when there was a national counting system, were several years younger and predominantly white by a modest margin, a pattern that has continued in later, more partial statistics (e.g., the 1979–1981 Treatment Outcome Prospective Study sample; Hubbard et al., 1989).12

The TC's group-centered methods encompass the following, all of which are grounded in an interdependent social environment with a direct link to a specific historical foundation:

  • firm behavioral norms across a wide range of proscriptions and specifications;
  • reality-oriented group and individual psychotherapy, which extends to lengthy encounter sessions focusing on current living issues or more deep-seated emotional problems;
  • a system of clearly specified rewards and punishments within a communal economy of housework and other roles;
  • a series of hierarchical responsibilities, privileges, and esteem achieved by working up a "ladder" of tasks from admission to graduation; and
  • some degree of potential mobility from client to staff statuses.

Because the therapeutic regimen of TCs has not been uniformly codified—and even if it had, would necessarily still involve substantial clinical discretion and creativity—there are great differences across programs in their recommended lengths of stay, staff-to-client ratios, and types of staff. These differences, which may be determined more by financial realities than by therapeutic philosophies, may have a great deal of influence over the differential clinical effectiveness of TCs.

De Leon (1986:5,7–8) has summarized the approach as follows:

The TC views drug abuse as a deviant behavior, reflecting impeded personality development and/or chronic deficits in social, educational, and economic skills. Its antecedents lie in socioeconomic disadvantage, poor family effectiveness and in psychological factors ... affecting some or all areas of functioning.... Thinking may be unrealistic or disorganized; values are confused, nonexistent or antisocial.

Physiological dependency is secondary to the wide range of influences which control the individual's drug use behavior. Invariably, problems and situations associated with discomfort become regular signals for resorting to drug use.

Thus, the problem is the person, not the drug.... In the TC's view of recovery, the aim of rehabilitation is global.... The primary psychological goal is to change the negative patterns of behavior, thinking, and feeling that predispose drug use; the main social goal is to develop a responsible drug free lifestyle. Stable recovery, however, depends upon a successful integration of these social and psychological goals.

Sugarman (1986:66,69) elaborates further:

All models of the TC involve a set of explicit behavior norms which members support and a set of contingent sanctions, positive and negative.... [H]ierarchical programs have extensive and demanding limits, strictly enforced, on the grounds that addicts need to learn self-control, and to experience the security of a firm framework of order.... Behavioral limits and sanctions plus positive peer pressure engender a short-term process of behavior modification. Even though this changed behavior is dependent upon the external controls of the social setting, still it has a real significance.... The message is: you can change in ways that you would not have thought possible.

The self-sufficient group is a particularly important setting for learning the nature of social responsibility and the interdependence of individual interests. Ideally, the ordinary family and the ordinary peer groups that a child experiences in growing up convey this kind of learning; in practice, the lesson is often missed.

To a significant extent the TC simulates and enforces a model family environment that the client, so to speak, should have had during critically formative preadolescent and adolescent years. The TC tries to make up for lost years of formation in an intensive, relatively short period of time—approximately 6 to 12 months of residential envelopment and an additional 6 to 12 months of gradual reentry to the outside community prior to "graduation." There is encouragement as well of continued alumni involvement for the benefit of role modeling for new residents, recognition and reinforcement for the graduate, and psychological and financial support for the program.

How Well Do Therapeutic Communities Work?

Conclusions about the effectiveness of TCs are limited by the difficulties of applying standard clinical trial methodologies to a complex, dynamic treatment milieu and a population resistant to following instructions. Randomized trials or natural experiments in the community, which would permit a well-controlled comparison of clients admitted to TC treatment versus an equivalent group (e.g., persons seeking treatment but denied admission, individuals admitted to other treatment modalities or arbitrarily excluded from TC treatment as a result of program closure) are not feasible or appropriate; when attempted, such experimental protocols have failed (see Bale et al., 1980). Currently, the strongest conclusions on the effectiveness of TCs are based on nonrandomized or nonexperimental but rigorously conducted studies of clients seeking admission to therapeutic communities. It is therefore worthwhile to look more closely at the nature, strengths, and weaknesses of such evidence.

The Character of Nonexperimental Evaluations

In nonrandomized or nonexperimental studies of treatment effects, conclusions generally depend on two kinds of comparisons. One is the contrast of observed TC outcomes with the record of similarly troubled individuals from the pretreatment era (e.g., those seen at Lexington or other prisons or hospitals). The problem with such comparisons is that one cannot be certain that the people of one historical period are totally similar to those of another. Likewise, there may or may not be similarities between a group seeking TC (or any other specific) treatment and a group seeking detoxification, or between a self-selected group from the community and a group culled from the drug-dependent population by the criminal justice system. Those seeking admission to TCs might (although they just as easily might not) represent a different kind of drug population or a very specialized slice of the population, or at least a different enough slice to honestly confound any comparisons of this sort. Because the same data are not collected on the different groups being compared, one cannot really reduce this uncertainty very much.

The second comparisons are internal ones, between those who enter TC treatment and those who apply for it but break off the process before entry, and between clients staying for longer versus shorter periods of treatment (receiving, in effect, larger and smaller "doses" of TC). In this case, the groups at least are being compared within the same time and data collection frame. Still, there may be selection effects that threaten the validity of the comparison, that is, its capacity to determine treatment effects. Those who stayed may have been different to begin with from those who left earlier. For example, they may have been intrinsically more or less likely to do well, treatment or no treatment (because of lesser or greater initial criminality, shorter or longer drug histories, better or worse family support). These differences may bias the comparison one way or another—either in favor of or against treatment effectiveness.

To guard against such biases, researchers rely on baseline measurements and statistical adjustments to control for preadmission client characteristics that might account for differential retention or outcome.13 These procedures increase one's assurance that the results are not confounded by selection effects; however, because some pretreatment characteristics that might conceivably affect retention and outcome may not have been measured well or even measured at all, they do not offer as much assurance as a successfully implemented, randomized clinical trial with minimal attrition.

The lack of randomized trials involving TCs is in some ways not surprising; most medical and criminal procedures became widely used without the benefit of such trials. The early success stories from the therapeutic communities Synanon and Daytop Village, in contrast to most treatment modalities' gloomy prior experience with heroin addiction, were positive and convincing enough that many clinicians and policymakers backed the establishment of TCs in the late 1960s and early 1970s. The scientific community paid them relatively little attention (a notable exception was Yablonsky, 1965), and many researchers viewed randomized trials as impossible to perform because heroin cases are so prone to noncompliance. 14

protocols is not specific to experiments involving TCs. Noncompliance has compromised attempts to compare alternative pharmacologically based modalities, as vividly demonstrated in several large-scale studies, including the attempted comparison of the effectiveness of methadone maintenance versus maintenance with the narcotic antagonist naltrexone (National Research Council, 1978) or methadone versus the longer acting methadone congener LAAM (Savage et al., 1976; Ling et al., 1978).

The Bale Study

The one notable attempt to undertake an experimental evaluation of the effectiveness of TCs compared with groups who were not treated or who were treated in other ways was conducted in California by Bale and coworkers (1980). This study, which examined methadone maintenance as well as TCs, did not work well as a random-assignment trial; in addition, the subject population was skewed from national norms. Nevertheless, its results are unique, important, and deserving of detailed attention for they underwrite much of the confidence that can be attached to results from studies that had no untreated control groups.

The subjects were 585 heroin-addicted male veterans who sought and gained entry to the Veterans Administration (VA) Medical Center in Palo Alto, California, for a 5-day opiate detoxification program during an 18-month intake period in the mid-1970s—who also met the study's requirements.15 When asked, about one-fifth of the subjects denied any interest in transferring to a VA drug treatment program after detoxification (some later changed their minds). The balance (plus the changers) were randomly assigned to either of two methadone maintenance clinics or one of three residential programs, each a different kind of 6-month TC.

The clinical staff invested significant time in trying to enlist every subject in his assigned program, and the overall rate of transfers from detox to VA programs doubled as a result. Nevertheless, the random-assignment design was thoroughly compromised (Tables 5-2a and 5-2b). Less than half of the randomly assigned subjects entered and spent as long as a week in any of the VA treatment programs, and only half of those entered the specific programs they had been assigned to (the others waited out at least a 30-day exclusion period to enter their own preferred program). Altogether, 42 percent of the total study cohort did not enter any kind of treatment during the follow-up year, about 28 percent entered one of the VA TCs, 12 percent entered a VA methadone clinic, and 19 percent entered a non-VA program.

TABLE 5-2A. Subject Compliance (percentage) with Assignment to a Therapeutic Program.


Subject Compliance (percentage) with Assignment to a Therapeutic Program.

TABLE 5-2B. Subject Compliance (number and percentage) with Assignment, Combining Therapeutic Communities (TCs).


Subject Compliance (number and percentage) with Assignment, Combining Therapeutic Communities (TCs).

The lack of compliance affected the study so profoundly that research analysts (who were independent of the clinical staff) were obliged to switch from the simplicity of randomizing assumptions to the use of multivariate statistical procedures to control for initial differences in age, ethnicity, prior treatment, drug use patterns, and criminal history among treatment and nontreatment groups.

At the one-year follow-up, those who had been successfully recontacted (the follow-up contact rate was 93 percent) were divided among the nontreatment (41 percent), non-VA (21 percent), short-term TC (14 percent), long-term TC (14 percent), and methadone (11 percent) options. 16 Controlling for pretreatment characteristics, the no-treatment, non-VA treatment, and short-term TC groups were statistically indistinguishable from each other at the follow-up. Compared with these groups, however, the long-term TC and methadone client groups (comprising one-fourth of the total sample originally contacted during the detox program) were clearly different. The long-term TC and methadone clients were:

  • two-thirds as likely to have used heroin in the past month (41 percent versus 64 percent);
  • three-fifths as likely to have been convicted during the year (22 percent versus 37 percent);
  • one-third as likely to be incarcerated at year's end (7 percent versus 19 percent); and
  • one-and-a-half times as likely to be at work or in school at year's end (59 percent versus 40 percent).

The long-term TC group ranked somewhat better than the total methadone group on each measure, but the differences were not large enough to be statistically distinguishable in a sample of this size.

Other Significant Follow-up Studies

Beyond the efforts of Bale and colleagues, there is a significant controlled observational literature on therapeutic communities. The bulk of these studies have focused on clients admitted to particular programs such as Phoenix House and Daytop Village in New York; in addition, the DARP (Simpson et al., 1979) and the TOPS (Hubbard et al., 1989) separately examined clients who were admitted to about 10 TCs across the country (not the same programs and 10 years apart).

The most extensive outcome evaluations from a single program come from Phoenix House in New York. De Leon and coworkers (1982) studied a sample of 230 graduates and dropouts and found that before admission the two groups were very similar with respect to criminal activity and drug use but that dropouts had somewhat greater employment. After treatment, the status of both groups was much better than before, but graduates had dramatically superior posttreatment outcomes compared with dropouts (Table 5-3).17

TABLE 5-3. Follow-up Results of Treatment at Phoenix House (New York City) Measured on Crime, Drug Use, and Employment Indices (percentage).


Follow-up Results of Treatment at Phoenix House (New York City) Measured on Crime, Drug Use, and Employment Indices (percentage).

The Drug Abuse Reporting Program provided further important controlled observational findings about the effectiveness of therapeutic communities (Sells, 1974a,b). The mean and median lengths of stay in the traditional TCs involved in the DARP were close to 7 months, which was well below the average 16-month treatment plan. At 12 months after admission, 71 percent of those admitted had left the TC voluntarily or by expulsion, although only 5 percent had completed their treatment plan by then; the ultimate graduation rate was 23 percent (Simpson et al., 1979).

Most of the DARP's outcome measures at one year after discharge (daily opiates, daily nonopiates, arrests, incarceration) were significantly better for TIC clients compared with the outcomes of detoxification-only and intake-only cases (Simpson et al., 1979). As in the Bale study, the multivariate-adjusted outcomes for TCs and methadone maintenance clients (matched for time since admission) on daily opiate use, nonopiate use, employment, and a composite index were quite similar. The length of stay in treatment was a positive, robust, significant predictor of posttreatment outcomes (drugs, jobs, and crime). Among clients staying more than 90 days in treatment, there was a positive and linear relationship between outcome and retention. The outcomes among clients staying less than 90 days were indistinguishable from detox-only and intake-only cases, and there was no discernible relation between outcome and short lengths of stay.

The final results of the TOPS, which were derived using multivariate logistical regression to control for pretreatment demographics, drug use, and criminality, yielded the familiar positive relationship between length of stay and outcome but with no clear threshold (Hubbard et al., 1989; see Figure 5-3). One year or more in a TC was significantly related to reduced heroin use, lower crime involvement, and increased employment at a 12-month follow-up. The odds of having problems with heroin or crime were about two-fifths as great for the long-term residential clients as for early dropouts, and their odds of having a job were nearly 1.7 times higher. Cocaine use followed a similar pattern, but the effect was not statistically significant. Alcohol problems were not related to treatment retention.

FIGURE 5-3. Outcomes and retention in therapeutic communities based on data from the Treatment Outcome Prospective Study and shown as odds ratios derived from multivariate analyses.


Outcomes and retention in therapeutic communities based on data from the Treatment Outcome Prospective Study and shown as odds ratios derived from multivariate analyses. The odds that members of the intake-only group will report a successful outcome at (more...)

In summary, multisite evaluation of the DARP (Simpson et al., 1979; Simpson, 1981) and the TOPS (Hubbard et al., 1989) both produced strong results supporting those of Bale and coworkers and the one or two useful single-program studies. Even in the absence of clinical trials, it is difficult to credit any explanation of these results other than the following: TCs can strongly affect the behavior of many of the drug-dependent individuals who enter them, and retention in treatment after some minimum number of months—how many seems to vary with the program—is positively and significantly related to improved outcomes as measured by illicit drug consumption, other criminal activity, and economically productive behavior.

Why Do the Results of Therapeutic Communities Vary?

No one really knows why there is such variation in TC performance and client responses (although strong views are often expressed about the matter) because there has been virtually no systematic research about the determinants of client success and failure in TCs. It is highly plausible that the results of TC treatment depend on its primary elements: the client's motivations, the quality and quantity of staffing, and the psychosocial organization and therapeutic design of the program. The committee heard anecdotally that TC staffing has been problematic during the 1980s as a result of constant budget pressures (staff numbers or salaries can be cut or held down more readily than room-and-board expenses) and rising competition with private-tier outpatient and chemical dependency treatment providers for credentialed, experienced staff. Yet there are no studies that specifically investigate how TC staffing relates to the effectiveness of treatment.

There are clearly wide variations in outcome indicators across programs. Client-Oriented Data Acquisition Process (CODAP) reports published from 1976 through 1981 make it possible to examine variations across cities in client status at discharge. The crude city differences are not adjusted to account for differences in the characteristics of clients treated in the various cities, nor can they be broken down to the program level. Nevertheless, the 1976–1981 CODAP reports demonstrate graphically that effectiveness varied significantly from area to area and undoubtedly even more so from program to program.

The year 1980 was one of relative program stability: the treatment system had been in place from five to six years and had not yet been disrupted by the massive system changes that resulted from the institution of block grants with their devolution of management responsibility to the states. Yet very large variations were seen in the treatment ''completion" rates reported for that year by residential programs (Figure 5-4), most of which were TCs. From the figure it is apparent that TCs in some cities diverged widely from the national average. Although the average residential completion rate across the nation was 10 percent, a sizable number of communities had averages well below and above this rate: 23 cities had rates between 5 and 15 percent, 9 cities were below 5 percent, 13 were between 15 and 24 percent, and 9 were above 25 percent. These variations have not been analyzed for possible attribution to differences in client characteristics, treatment process, quality of staff, or random processes. There is also currently no usable evidence of national scope showing whether client discharge statuses still exhibit such differences across geographic areas, or why.

FIGURE 5-4. Variations in "completion rates of opiate clients in residential programs in U.


Variations in "completion rates of opiate clients in residential programs in U.S. cities, 1980. Source: National Institute on Drug Abuse (1981)

Costs and Benefits of Therapeutic Community Treatment

Most evaluations of TCs indicate that they are cost-effective or cost-beneficial, or both. There have been fewer rigorous evaluations of costs and benefits than of cost-effectiveness, however. A simulation by Maidlow and Berman (1972) showed that a TC produces $213,000 of economic benefits to society per client at a cost of $14,700 (a benefit/cost ratio of 14.5 to 1). These authors concluded that TCs were highly cost-beneficial compared with prisons (which were much more expensive and had high recidivism rates). Rufener and coworkers (1977a) focused only on benefits after treatment (ignoring benefits during treatment) and tried to sort out the benefits that accrue variously to the client, the government account, and society as a whole. They estimated that the combined benefit/cost ratio of TCs after the treatment period was 1.9 to 1. As discussed earlier, however, both of these sets of benefit/cost ratios are biased upward because the assumptions used to produce them were overly optimistic compared with what is now known about treatment retention and effectiveness.

A more realistic cost-effectiveness study using the DARP data base (Rufener et al., 1977a) found that TCs generally produced greater differentials than methadone or outpatient nonmethadone treatment in terms of legitimate income and employment status after treatment versus before. But methadone was decidedly more cost-effective — measured as the cost per added day of desirable outcome — because it was cheaper. Of course, these comparisons work only to the degree that those entering one treatment would as readily have entered the other.

A cost/benefit study of the Gaudenzia House TC (Griffin, 1983) compared the expense of operations over a five-year period with the benefits from reduced criminal activity and increased social productivity. The analysis distinguished the benefits to be derived from treatment "successes" and "failures," finding positive ratios of benefits to costs for both groups (9 to 1 for "successes'' and 3.4 to 1 for "failures"). Benefits accrued even for "failures" because while in residence for treatment they were unable to commit as many street crimes (analogous to the incapacitation effect of incarceration) as they would have if not in residence.

Most recently, Hardwood and colleagues (1988) analyzed the TOPS data base, examining the reduced crime-related impacts on society that result from drug treatment. A particularly important finding was that TC treatment, as with methadone treatment (see the section above entitled "Costs and Benefits of Methadone Treatment"), virtually pays for itself during the time it is delivered, owing to the reduced criminal activity of clients in treatment relative to either the pre- or posttreatment periods. Further benefits accrue after leaving treatment. The final benefit/cost ratio for TCs was 3.8 to 1 using the primary measure (the costs of crime) and 2.1 to 1 using a more conservative employment-oriented measure.


TCs are for the most part designed to treat individuals who are badly impaired by drug problems and other deficits, and client decisions about whether to seek TC treatment reflect an awareness of that design. Even those who do seek treatment often drop out of TCs in short order, in contrast to the much higher retention rates of those who enter methadone maintenance. There is, nevertheless, a sizable population who not only find TC treatment initially attractive but also will remain in this modality for a substantial fraction (up to the whole course) of planned treatment. This segment is distinct from the typical methadone maintenance population: it is appreciably younger, more heavily white, and more likely to use multiple drugs.

The committee considers the evidence about the following to be fairly persuasive (although not ironclad): those clients who stay in TCs for at least a third or half of the planned course of treatment, a threshold that seems to vary greatly from program to program — that is, those who stay in treatment for at least 2 to 12 months, varying from program to program for reasons that are not yet clear — are much closer to achieving the treatment's goals at follow-up than those who drop out earlier. The outcomes of the earlier dropouts basically cannot be distinguished from those of individuals who did not enter any treatment modality.

These improvements over nontreatment, which are estimated to be in the neighborhood of one- to two-thirds reductions in the rates of primary drug consumption and other criminal activity and half-again increases in the rates of employment or schooling, vary with the amount of time spent in treatment. TC graduates have outcomes that are even better than these rates, but they are a small percentage (usually 15 to 25 percent) of total TC admissions. What is most important here is that graduates are not the only ones who benefit themselves and society as a result of spending time in a TC. Even for those individuals who "split" early, even for those who show no later effects, the TC may be a good social investment considering that a day in a TC is a day away from street crime.

Outpatient Nonmethadone Treatment

What is Outpatient Nonmethadone Treatment?

Outpatient nonmethadone (OPNM) programs range in duration from the one time assessments and referrals of drop-in and "rap" centers to virtual outpatient therapeutic communities with daily psychotherapy and counseling intended to continue for a year or longer (Kleber and Slobetz, 1979). In between are the vast majority of programs, which see clients once or possibly twice weekly and deliver services based on theoretical approaches from psychiatry, counseling psychology, social work, therapeutic communities or the 12-step Anonymous creed. Some programs contract extensively with Treatment Alternatives to Street Crime agencies or probation departments (see Chapter 4), monitoring the shared clients' compliance with probation conditions — particularly through administration of drug tests — and offering no other therapeutic services.

Some OPNM programs utilize psychoactive medications prescribed by psychiatrists or other physicians on staff. These agents may be medications used initially in detoxification to ameliorate withdrawal symptoms, maintenance antagonists that prevent intoxication (e.g., naltrexone), medications to control drug cravings after withdrawal (especially innovative cocaine pharmacotherapies), or drugs that address psychiatric disorders (depression, mood disorders, schizophrenia, etc.). Programs with the requisite resources may deliver or link their clients to formal education, vocational training, health care (such as AIDS testing or treatment), housing assistance (especially for homeless clients), support for battered spouses and children, and other social services.

The diversity of OPNM treatment defies easy summary and is matched by the heterogeneity of its client populations. These populations generally are not abusing opiates, usually are not involved in the criminal justice system (at least, were not so during the DARP and TOPS periods), and include significant proportions of abusing rather than dependent individuals — differing in all these respects from typical methadone and TC clients.

How Well Does Outpatient Nonmethadone Treatment Work?

The major conclusion that can be offered about the effectiveness of outpatient treatment is a familiar one: clients who remain in treatment longer have better outcomes at follow-up than shorter-term clients. These conclusions are based entirely on multivariate results of the two major multisite evaluations, the DARP and the TOPs. In the Drug Abuse Reporting Program study of clients entering treatment from 1969 to 1972 (Sells, 1974a,b; Simpson, 1981), OPNM clients exhibited statistically significant follow-up improvements relative to pretreatment in terms of employment and consumption of opiates and nonopiates, but not in terms of arrest rates, which were much lower before treatment than they were in TIC or methadone maintenance clients. The DARP comparison groups, those in detox programs and those who only made contact with treatment during intake, reported no significant preto posttreatment changes except in opiate consumption (Simpson et al., 1979).

Analyses of retention (Simpson, 1981) produced results identical to those for TC clients: clients staying in treatment less than 90 days showed no improvement relative to the detox and intake-only clients, whereas those staying longer had improved outcomes on a composite score that incorporated drug, criminality, and productivity scales. For the 90-days-plus group, outcome scores were strongly and significantly correlated with total length of stay.

The larger and more recent Treatment Outcome Prospective Study (Hubbard et al., 1989) collected data on 1,600 OPNM clients admitted to 10 programs. Clients again reported better performance during and after treatment than before admission, and multivariate analyses strongly related posttreatment outcomes to length of stay, using multivariate logistical regression to adjust for client drug use histories and sociodemographic characteristics at admission (Figure 5-5). The analysis suggested that the critical retention threshold may be six months, but only 17 percent of TOPS outpatient clients were retained this long. OPNM dropout rates were quite high—significantly higher than for methadone or TCs. At four weeks the programs retained only 59 percent of clients; 18 percent eventually completed the course of treatment.

FIGURE 5-5. Outcomes and retention in outpatient nonmethadone programs based on data from the Treatment Outcome Prospective Study and shown as odds ratios derived from multivariate analyses.


Outcomes and retention in outpatient nonmethadone programs based on data from the Treatment Outcome Prospective Study and shown as odds ratios derived from multivariate analyses. The odds that members of the intake-only group will report a successful (more...)

Why Do the Results of Outpatient Nonmethadone Treatment Vary?

There is no answer to this question for OPNM programs. Although there is evidence of variation in program retention rates, there is very little information about what the "active ingredients" in this treatment modality are that might lead to these variations. One can only speculate that the same factors that emerge from methadone and TC research, in particular, staff quality and program design, may be equally important here.

Benefits and Costs of Outpatient Nonmethadone Treatment

Both of the major multisite studies, the DARP and the TOPS, have been analyzed with respect to the costs and benefits of OPNM treatment. Rufener and colleagues (1977a) compared the cost-effectiveness of the major treatment modalities for the DARP subsample of opiate clients. For this population, OPNM generally had poorer cost-effectiveness than methadone and TCs, but no attempt was made to address whether OPNM was more cost-effective than no treatment or whether longer treatment was more cost-effective than brief episodes.

Harwood and coworkers (1988), using the methods described above for methadone and TCs, estimated a benefit/cost ratio of 1.3 to 1 for OPNM in the TOPS data base. Compared with a similar detox sample, increased treatment retention in OPNM programs had a modest but measurable impact on the amount of theft while in treatment, even though the OPNM treatment population was less criminally active than the populations in methadone treatment and TCs. An alternative measure, improvement in the amount of legitimate employment, produced a benefit/cost ratio of 4.3 to 1, indicating that the benefits of OPNM are more pronounced in terms of the secondary goal of employment rather than as reductions in already low-levels of criminal activity. Unlike the results of TC treatment, crime-related benefits of OPNM after discharge were not discernible.

Chemical Dependency Treatment

What Is Chemical Dependency Treatment?

Chemical dependency (CD) treatment (also called the Minnesota model, 28-day, 12-step, or Hazelden-type treatment) is the predominant therapeutic approach taken by the privately financed inpatient and residential programs identified in Chapter 6 as the "private tier" of providers. Virtually all of these programs were originally oriented toward alcohol problems but have increasingly served clients with illicit drug problems. The CD theory of the disorder and the modality's treatment approach have expanded from a focus on alcoholism that depended on the Alcoholics Anonymous principles (the 12 steps) to one more broadly addressing dependence on any chemical substance. Cook (1988a,b) has provided a concise historical review of the development of the Minnesota model. He notes the similarities between the underlying theories that shape CD and TC treatment but observes that they developed almost completely independently of each other.

Almost exclusively, the goal of CD treatment is abstinence from alcohol and drugs. The client is viewed as a victim of a disease process but also as the person with the primary responsibility for making behavioral changes that will promote abstinence, which will in turn eliminate problems resulting from alcohol or drugs.

In its most sophisticated formulation, the CD approach views drug problems as having multiple causes. There is a physiological phenomenon at work, but psychological component often requires some pharmacological intervention as an integral aspect of treatment. The treatment's psychological dimension highlights the impact of emotional, motivational, and learning problems on dependence. Sociocultural models explore the relation of drinking and drug problems to socialization processes and environments. CD treatment practices represent a blending of the Alcoholics Anonymous model of recovery, certain insights and prescriptions of somatic medicine, and psychiatric and behavioral science principles.

Chemical dependency treatment is usually an intensive, highly structured three- to six-week inpatient regimen. Clients begin with an in-depth psychiatric and psychosocial evaluation and then follow a general education-oriented program track of daily lectures plus two to three meetings per week in small task-oriented groups. Group education teaches clients about the disease concept of dependence, focusing on the harmful medical and psychosocial effects of illicit drugs and excessive alcohol consumption. There is also an individual prescriptive track for each client, meetings about twice a week with a "focal counselor," and appointments with other professionals if medical, psychiatric, or family services are needed. Recently, there has been increasing emphasis on family (or "codependent") therapy and the concept that others may be acting as "enablers" of drug and alcohol consumption.

Clients actively engage in developing and implementing a recovery plan, which is patterned on the "step work" (working through the 12 steps that lead to recovery) of Alcoholics Anonymous. Self-help is a large part of therapy; clients work with each other and are generally required to attend Alcoholics/Cocaine/Narcotics Anonymous (AA/CA/NA) meetings.

Aftercare is considered quite important in CD treatment, but there are relatively few program resources devoted to it. It can last from three months to as long as two years and range in intensity from a simple monthly telephone follow-up to intensive weekly group therapy and individual counseling as needed. Clients are urged to continue an intensive schedule of AA/CA/NA attendance through the follow-up period, with continued contacts thereafter at a lower rate.

CD treatment has some elements in common with the TC approach: abstinence as a goal, striving for behavioral changes to achieve abstinence, the client taking primary responsibility for his or her problems, and recovery in the context of mutual support, including that of counselors. But there are noteworthy differences between the two modalities. The inpatient or residential phase of the CD treatment plan is short relative to TC treatment, and the extended follow-up or aftercare phase is seldom if ever a strong and integrated program element. Because the hospital-based services of CD treatment do not require clients to perform housekeeping duties, there is more time for psychotherapy and educational work; in the TC process, however, housekeeping and other program maintenance responsibilities are considered an integral component of therapeutic learning. CD program staff, like TC staff, are a mixture of stable, recovering (from alcohol or drug dependence) individuals and professional clinicians from traditional health care, mental health, and social service disciplines. However, CD staff tend to be more heavily credentialed.

CD treatment is full of educational work, including writing, reading, and lectures; there is little of the daily job routines or ladder of work responsibilities that are intrinsic to TC treatment and by which client progress is symbolized. (In CD programs, progress is made by ascending spiritual steps.) Most TCs depend heavily on advanced clients to direct the progress of new clients. Prior to admission, CD clients are usually enacting some stable social roles, whereas TC clients almost always have massive functional and social deficits. CD programs, with their residential treatment duration, are more attractive to clients with greater initial functional and social resources: indeed, the prototypical CD client used to be fortyish, middle class, employed, white, and dependent on alcohol. Today, although the clientele is more diversified (programs are now seeing more clients with combined cocaine/alcohol problems, as well as a segment of adolescents with both psychiatric and drug diagnoses), these origins continue to shape the CD approach.

How Well Does Chemical Dependency Treatment Work?

Although CD programs have come to play a major role in the drug treatment world, the research data on this type of treatment for illicit drug problems are weaker than for the other modalities. There are no relevant experimental or quasi-experimental studies; there were no CD programs in the DARP or TOPS samples. Only one of the available observational studies of CD programs employs an untreated comparison group (Rawson et al., 1986),18 and none have collected data on admissions with short lengths of stay. There is also practically no use of multivariate statistics.

The extent of reasonably certain knowledge about CD treatment is that clients who present drug problems at admission have poorer outcomes at the posttreatment follow-up than alcohol clients (with no illicit drug consumption) in the same programs. This finding is consistent across studies by the CareUnit system, the Chemical Abuse/Addiction Treatment Outcome Registry (CATOR) follow-up service, and the Hazelden center in Minnesota.

The CareUnit study (Comprehensive Care Corporation, 1988) sampled 1,002 adult clients who stayed at least five days in 1 of 50 different CareUnit programs in 1987. (CareUnits treated 46,000 adults and adolescents in more than 200 locations.) About 53 percent of the sample had used multiple substances before admission, and 29 percent reported polydrug consumption on a daily basis. Clinical program staff interviewed 723 clients from the sample at least one year after discharge. Sixty-one percent were classified as recovering at follow-up (fewer than four instances of use since discharge). Abstinence was poorer for preadmission consumers of illicit drugs (54 percent for those who had used cocaine and 48 percent for those using marijuana) and polydrugs (56 percent) than for consumers primarily of alcohol (63 percent). The strongest indicator of outcome was attendance at self-help groups after discharge: only 48 percent of nonattenders were recovering, compared with 79 percent of those attending the groups more than 29 times.

The CATOR study is a multisite comparison of independent programs. Hoffmann and Harrison (1988) found that at least 38 percent of clients in 22 adult inpatient programs in the Midwest had an admission history that included illicit drugs. However, the study excluded clients with fewer than 10 days in treatment, followed virtually no one who did not complete treatment, and reached only 37 percent of completers at the two-year follow-up interview. Few results were detailed specifically for individuals with drug problems, but the authors note that "[p]revious CATOR analyses have consistently found that polydrug users have the poorest prognosis for abstinence, followed by regular marijuana users… The relationship of use pattern to recovery is confounded to some extent by age since the polydrug and marijuana groups contain a larger proportion of younger patients" (p. 31).19

Why Do the Results of Chemical Dependency Treatment Vary?

There are no useful studies that distinguish the reasons why some clients in CD programs recover and others do not. As with other treatments, client motivation and program staff quality are suspected factors. But there is no readily available information on variations in drug client outcomes across CD programs or any attempts to relate such differences to systematic variations among clients or in the therapeutic approach.

Benefits and Costs of Chemical Dependency Treatment

There are no studies available on the costs and benefits or cost-effectiveness of this modality. There is some discussion of cost data, however, in Chapters 6 and 8.


Detoxification, unlike the previous modalities, is not a treatment for drug-seeking behavior. Rather, it is a family of procedures for alleviating the short-term symptoms of withdrawal from drug dependence (NIDA, 1981, 1983b; Kleber, 1987).20 The major procedure is observation (because withdrawal is self-limiting and ordinarily not life-threatening, although it can be uncomfortable). There is some standard clinical indications for administering pharmacological agents during detoxification: to ameliorate severe withdrawal symptoms, to induce relaxation, to prevent seizures in the case of sedative-hypnotic drugs,21 or to counteract severe depression.

Detoxification of different drugs involves different durations and medications. Various pharmacological agents are used for withdrawal from opiate addiction, which has been extensively studied and reported for more than 60 years. The most common detox drug is methadone, but benzodiazepines, clonidine, and some other agents also are frequently used to control withdrawal symptoms. Opiate detoxification has often been done rather slowly—over several weeks or even several months—particularly in cases in which there is a long, continuous history of dependence. Today, however, new, more rapid forms of detoxification using combinations of drugs such as clonidine and buprenormine are being tested and used in residential and outpatient settings.

Detoxification of cocaine, particularly crack-cocaine dependence, has been especially difficult, but some promising approaches are now emerging. Cocaine dependence typically involves a series of binges that last from 12 to 36 hours each. These binges are usually followed by several days without cocaine use but with gradually mounting withdrawal symptoms that include mood alterations, diminished capacities for experiencing pleasure (anhedonia), and craving for cocaine. These symptoms may not abate for four to eight weeks, thus yielding another binge cycle in very short order. The critical task in detoxification is to disrupt the imminent return of the cocaine cycle.

There has been some success in the management of cocaine withdrawal symptoms and craving in ambulatory clinical trials using desipramine hydrochloride (Gawin et al., 1989a), amantadine (Tennant and Sagherian, 1987), bromocriptine (Dackis et al., 1987), flupenthixol decanoate (Gawin et al., 1989b), and buprenorphine (Mello et al., 1989), among other drugs, in conjunction with once-a-week outpatient counseling. These treatments reduced short-term rates of relapse two– to threefold for a majority of those treated. Unfortunately, most of these agents do not begin to have their major clinical effect for one to two weeks, during which outpatient dropout often occurs (in the programs in which trials have been conducted, dropout rates range from 30 to 70 percent).

Comfort, the avoidance of seizures (the most common cause of fatalities), screening and treatment of infections and other medical problems, and the achievement of a condition in which withdrawal distress is not evident are and should be the primary goals of detoxification. In these terms, current detoxification procedures for most drugs are virtually always effective if they are completed, permitting a transition to abstinence with only attenuated symptoms of withdrawal. The key to completing detoxification successfully is compliance with the detox protocol: taking medication in prescribed amounts and schedules and avoiding intervening use of the drug on which the client is dependent and any other nonprescribed drugs.

Inpatient or residential detoxification appears logically to offer better opportunities for clinicians to ensure compliance with detoxification prescriptions. There is little evidence, however, by which to judge whether this supposition is, indeed, true. Inpatient, residential, and outpatient drug detoxification have not been adequately compared to permit confident conclusions on which has the best compliance record or who belongs in which setting. On technical grounds, detoxification of most illicit drugs in most cases can occur as safely and effectively on an ambulatory basis as in a bedded setting. Hospital treatment in particular calls for justification on relevant medical grounds, such as history of seizures, concurrent conditions needing hospital care, or special cases of risk such as neonates of dependent mothers. On the basis of cost, an ambulatory detox is therefore preferable for most individuals when the medical criteria dictating inpatient detoxification are not present. These issues will be discussed further in Chapters 7 and 8.

It is crucial to underscore the fact that the goals of detoxification are quite limited. This restricted scope is mainly a product of extensive experience with the lack of longer term effects of detoxification, especially of heroin dependence. Consistently, without subsequent treatment, researchers have found no effects from detoxification that are discernibly superior to those achieved by untreated withdrawal in terms of reducing subsequent drug-taking behavior and especially relapse to dependence. No appreciable success in increasing rates of recovery from heroin dependence after detoxification alone has been demonstrated for different pharmacological agents or for various detoxification protocols (e.g., rapid versus slow tapering of dose). Review articles reaching this decisive conclusion include those by Resnick (1983), Newman (1983), Cole and colleagues (1981), Moffet and coworkers (1973), and Sheffet and colleagues (1976). There is much less of a literature on cocaine detoxification, but clinicians who are experienced in treating opiate dependence do not believe that short-term detoxification alone will prove any more effective with cocaine.

On the other hand, a detoxification episode offers clinicians a major opportunity to recruit clients into treatment, as the Bale team's study (1980) demonstrated (see also NIDA, 1981; Kleber, 1987). Success at recruitment may well be a more critical outcome for detoxification programs than the conventional primary goals of comfort and suppressing withdrawal symptoms. There appear to be significant variations across U.S. cities in successfully enlisting detoxification patients into treatment. Discharge data from the 1980 CODAP report indicated that only 14 percent of opiate detoxification clients were transferred or referred to further treatment, although there was substantial variation around his average: out of 62 reporting areas, 12 had transfer/referral rates lower than 5 percent, and 14 had rates greater than 25 percent. There are no studies to indicate whether such variations relate to systematic differences in clients, the treatment process, or staff performance, or to chance.

Correctional Treatment Programs

The overall record of research on prison-based drug treatment programs is moderate in scope, and the findings mostly correspond to the largely negative results observed in the treatment of criminals during incarceration in hopes of reducing their recidivism (Vaillant, 1988; Besteman, 1990; Chaiken, 1989). Yet Falkin and colleagues (1990) sound a more optimistic note:

Given the current array of treatment programs (many offering only occasional counseling, drug education or other limited services), the finding of evaluation research that many programs are ineffective is not surprising. To adjudge that drug treatment is unable to control recidivism because many programs do not is to miss the crucial point that some programs have been quite successful. With the proper program elements in place, treatment programs could achieve a significantly greater reduction in recidivism than by continuing a policy of imprisonment without adequate treatment.

Their list of the elements necessary for a successful prison drug treatment program22 is succinct:

  • a competent and committed staff;
  • adequate administrative and material support by correctional authorities;
  • separation from the general prison population;
  • incorporation of self-help principles and ex-offender aid;
  • comprehensive, intensive therapy aimed at the entire lifestyle of a client and not just the substance abuse aspects; and
  • an absolute essential—continuity of care into the parole period.

Three controlled evaluations of prison-based programs that incorporate these criteria are available and are discussed in the sections below. The first used three control groups: a group of program applicants (this was a voluntary program) who did hot receive treatment for lack of timely openings—essentially a random selection process—and participants in two other kinds of treatment in the same prison system. The second study (also a voluntary program) used as controls an early-dropout group and an untreated group from the same prison system. The third study, sampling a very large prison/parole program with more than 1,000 admissions per year (partially voluntary), used a sophisticated case-control matching procedure involving the early dropouts from the program. All three studies collected data on the entire group entering treatment for periods of 2 to 11 years after release from confinement. Overall, the results indicate that sizable positive effects can be obtained from treatment, although the results are not unequivocal.

Stay'n Out and Cornerstone

The most recent and currently most influential study (see the discussion of Project REFORM in Falkin et al., 1990; also Frohling, 1989) is of Stay'n Out, a New York program that operates a four-unit, 146-bed prison program for male inmates and a separate 40-bed program for female inmates. The program is based on the social organization of a major therapeutic community, Phoenix House, and adapted to the prison setting; it works closely with community-based TCs to extend treatment contact after release. Stay'n Out clients from 1977–1984 (N = 682) were compared with similar groups of drug-abusing and dependent prisoners. The comparison groups received either regular drug abuse counseling (N = 576) or milieu therapy, which is a staff-intensive congregate-residential counseling approach or quasi-TC (N = 364); there was also an untreated control group who applied and were waiting for Stay'n Out admission but who were not treated because there were not enough openings during their window of eligibility, the 6 to 12 months before their first parole hearing (N = 197). The groups were followed through 1986 (i.e., from 2 to 9 years after release from prison).

As indicated in Table 5-4, the TC group was arrested significantly less often than the other groups, with differences of 8 to 14 percentage points (which represent 22 to 35 percent reductions in rearrest rates) for men and 6 to 12 percentage points (25 to 40 percent reductions) for women. Because for every arrest, such criminally inclined individuals have generally committed hundreds of crimes (Ball et al., 1981; Johnson et al., 1985; Speckart and Anglin, 1986), these differences in rearrest rates are a valuable result. The authors indicate, however, that intergroup differences at follow-up in rates of reincarceration, rapidity of rearrest, and parole revocation were statistically or substantively negligible, except that significantly more Stay'n Out-treated women than untreated women successfully completed their parole term.

TABLE 5-4. Results of Evaluation of Stay'n Out Prison Treatment Program (New York) Compared to Groups Receiving Other Treatment Modalities or No Treatment.


Results of Evaluation of Stay'n Out Prison Treatment Program (New York) Compared to Groups Receiving Other Treatment Modalities or No Treatment.

A similar controlled observational study of the Cornerstone program has been reported by Field (1984, 1989; see Table 5-5). Cornerstone is a modified TC program (a mixture of milieu therapy and TC principles) located in Oregon State Hospital in Salem. It is designed for state prisoners in the last year prior to eligibility for parole; after release, the parolees move to a halfway house that includes some therapeutic contacts. Study results indicate that prisoners in the program were convicted significantly less often than comparable parolees in the three years following release.23

TABLE 5-5. Results of a Three-Year Follow-up of the Cornerstone Treatment Program (Oregon) Comparing Program Graduates, Program Dropouts, and Untreated Parolees.


Results of a Three-Year Follow-up of the Cornerstone Treatment Program (Oregon) Comparing Program Graduates, Program Dropouts, and Untreated Parolees.

Graduates of Cornerstone did much better at follow-up than early dropouts from the program. In the Stay'n Out study, and in several other well-regarded, well studied voluntary correctional programs (see Falkin et al., 1990), length of stay in treatment correlated strongly with positive follow-up measures, the same result seen in community-based programs. The fact that early dropouts from prison programs are even more likely to recidivate, by every measure, than are untreated controls suggests that prison-based TCs may be more efficient than community-based programs at sorting out and excluding (or encouraging self-exclusion of) the poorest responders.

The California Civil Addict Program

A different type of correctional treatment program combines treatment in a penal institution with specialized parole supervision, including access to a variety of community-based treatment opportunities. The most comprehensive and well-studied example of this kind of program was CAP, the California Civil Addict Program, which began in 1961 (McGlothlin et al., 1977; Anglin and McGlothlin, 1984; Anglin, 1988). Two similar civil commitment programs, one federal and one operated by the state of New York, fell far short of their design goals, ended fairly quickly, and were roundly regarded as failures (Besteman, 1978, 1990; Inciardi, 1988). Even the CAP effort operated as designed only until 1969, after which much of its original character was lost, principally because the strict therapeutic rationale was overturned by the general fiscal leanness and operational leniency that overtook the California penal system during then-Governor Ronald Reagan's second term. In addition, community-based treatment programs funded largely by the federal government became available after 1970, creating attractive treatment alternatives for criminal justice agencies and clients. As discussed earlier, this expansion of treatment coincidentally presented a research opportunity to compare the results of the correctional treatment program and methadone maintenance.

CAP permitted adjudication of heroin-dependent individuals through a civil commitment procedure rather than regular criminal sentencing. 24 The first (repeatable) stop for CAP clients, once they had been committed, was a term in the California Rehabilitation Center at Corona, a medium- security prison with a large staff of psychotherapists. This period began a seven-year term of supervision, three-fourths of which, on average, was spent on parole in the community rather than in the center. (The seven-year commitment term could be terminated after three consecutive drug-free years in the community.) The community supervision component involved specially trained parole officers, smaller caseloads of only 30 parolees, and weekly drug testing (Anglin, 1988).

The conditions conducive to a case-control study were inadvertently created during the initial years of the program. The original commitment law was complex enough that legal-procedural errors were made in committing at least half of the early CAP clients. Sooner or later, most of these 1961–1963 commitments were challenged by writs of habeas corpus, and the individuals were released by court order from CAP incarceration and supervision and returned to the regular track of criminal adjudication, with credit for the time served during CAP. The overall writ-released group differed from those who continued in CAP in that many writ-releases had less serious offenses for which the CAP commitment of seven years was a longer term than the sentence (including parole or probation) they would probably otherwise have served. On the other hand, virtually all of the continuing CAP group would probably have had longer sentences without the CAP diversion. The researchers therefore used matching procedures to select from within the writ-released group a comparison sample that was as similar as possible to the continuing group on 15 criteria, including criminal and drug histories and demographic characteristics.

During their years under CAP, individuals retrospectively reported that they reduced their heroin use (Figure 5-6) as well as total criminality while unincarcerated to levels that were half or less than half the amount reported by the comparison group. These reductions became apparent immediately after their release into the community, and they were sustained. The difference between CAP parolees and the comparison group on these dimensions narrowed over the next several years as more members of the writ-released group (some of them recommitted to CAP for new offenses) reduced their heroin use and other criminal behavior. By the time the continuing CAP group's parole ended, the control group was at a more nearly similar level, especially considering pretreatment (baseline) differences. The subsequent recovery paths of the two groups remained parallel.

FIGURE 5-6. Effects of the California Civil Addict Program on daily narcotics use.


Effects of the California Civil Addict Program on daily narcotics use. The percentage of nonincarcerated time during which subjects reported using narcotics daily is show for 8 pre- and 13 postadmission years. The vertical line at A denotes admission (more...)

In summary, the residential and community supervision components of CAP were evidently effective in accelerating the recovery of a significant fraction—at least half—of the treated group.

A different result of the CAP study was to examine the effects of methadone maintenance treatment during CAP supervision (Anglin et al, 1984). In 1971, as discussed earlier in the chapter, methadone programs were opened in a number of California's cities, and some members of both the CAP and the writ-released comparison group who had continued active heroin use elected to enter methadone programs. (Parole officers neither insisted on this option nor opposed it.) In both samples, entry to methadone had powerful effects on individuals who, by and large, had not otherwise begun recovery—effects as great or greater than those of CAP parole itself (Figure 5-7). There were no significant differences between the CAP and the comparison group in how methadone affected their heroin-seeking or other criminal behavior.

FIGURE 5-7. The effect of methadone maintenance on daily narcotics use in the California Civil Addict Program and control groups.


The effect of methadone maintenance on daily narcotics use in the California Civil Addict Program and control groups. The percentage of nonincarcerated time during which subjects reported using narcotics daily is shown for 8 pre- and 6 postadmission years. (more...)

Boot Camps

A final type of prison-based treatment that has received much attention recently is the "boot camp" or "shock incarceration" (SI) concept for young offenders. This treatment constitutes a three- to six-month sentence for young offenders who are remanded to a facility employing rigorous physical exercise and a small-group organizational structure similar to Outward Bound or military training camps. A number of states, beginning with Georgia in 1983, have opened such facilities, largely as a way to reduce prison costs and improve resource management. Shock incarceration segregates young offenders, who would otherwise be mixed with the general penitentiary population (in this case, SI reduces penitentiary overcrowding) or with the general probation population (in this case, using the SI option increases the need for correctional facilities).

There are several studies under way to improve understanding of how these programs work. Boot camps vary in nature. Some are entirely militaristic environments with few if any therapeutic staff or procedures; others incorporate many drug treatment elements that the more successful prison treatment efforts display but lack still other requirements—particularly continuity of care when the individual returns to the community. Parent (1989:4,5), in a report to the National Institute of Justice, summarizes current knowledge:

Preliminary case tracking data raises questions about SI's capacity to reduce recidivism. The Oklahoma Department of Corrections used survival analysis to compare return rates of SI graduates with similar non-violent offenders sentenced to the DOC. After 29 months almost half the SI graduates, but only 28 percent of the other group, had returned to prison.

In a three year follow-up, the Georgia DOC found that 38.5 percent of their SI graduates returned to prison. For Georgia SI graduates who were in their teens when admitted to SI, 46.8 percent returned to prison within three years of release. In an earlier study, Georgia researchers found little difference in one-year return to prison rates for SI graduates, and similar offenders sentenced to prison and to a youthful offender institution. It should be emphasized that neither of these studies involved carefully constructed comparison groups.

Until evaluation results become available, policymakers should view claims of incredible success with skepticism, and should be cautious about proceeding with SI development on the basis of high hopes, preliminary data, or press clippings.

Conclusions about Prison Treatment

Prisoners with drug problems are "hard cases," but in terms of avoidable social damage, success in accelerating the recovery of even a modest proportion of them yields substantial social benefits. The limited research information on correctional treatment indicates that some programs have delivered this benefit, but many have not. The research does not clearly demonstrate why only a few prison programs have curbed recidivism, but clinical judgments about the key differences between effective and noneffective programs are consistent with the available evidence and bear repeating here.

First, clients need not be dragooned into treatment in order to enlist substantial participation in correctional programs. In the three programs reviewed here in detail, and in most well-regarded programs, entry has largely been a matter of negotiation or multilateral consent, requiring the fulfillment of certain obligations by the prospective client, program staff, and custodial authorities. The principal requirement for effective correctional treatment programs is responsiveness: the program must respond to individual client behaviors as surely as the individual must respond to clinical protocols and queries. The treatment programs have had authority to exclude clients. Mutual consent and performance are a recurrent theme, evident in the formulation of entry contracts and treatment plans, the incorporation of self-help principles and systems of earned program privileges, and roles for program graduates.

Successful correctional treatment requires clinically skillful staff who are strongly committed to their work. To maintain staff skills and commitment in the face of difficult cases is impossible without adequate material and administrative support from correctional and other authorities. Another vital element is follow-up research to let staff know what effects their efforts are having.

Treatment is not an alternative to penalties for committing violent and acquisitive crimes such as robbery, burglary, and larceny, for which offenders with drug problems are so frequently apprehended. Treatment decisions (including admission and termination) need to be made on therapeutic grounds in terms of program goals and rules; decisions in the interest of justice and custodial security must also be made by the appropriate authorities on their own merits. But decisions in either sphere must be consistent with explicit rules, and agencies must be prepared to follow through on them. Contingencies such as revocation and return to custody in the event of noncompliance with release conditions must be believable and consistently enforced.

Summary and Conclusions about Treatment Effectiveness

The committee is both satisfied and disappointed with the conclusions that can be drawn about the effectiveness of the major drug treatment modalities. It is satisfied that some modalities have been studied with sufficient skill and methodological integrity that conclusions can, indeed, be drawn (even though there is still much to be desired in the way of useful knowledge). It is disappointed that the same cannot be said of other modalities and that the overall state of knowledge about treatment effectiveness has not grown more rapidly in the past 5 to 10 years. Most of what is known is based on data collected between 1969 and 1981.

Table 5-6 is a succinct statement of this disappointment. Of the four major modalities, methadone maintenance has received the most extensive study, using all of the main type of treatment evaluation research techniques. Therapeutic communities have received the next most extensive assessment; outpatient nonmethadone treatments have been evaluated at a somewhat lower level. Chemical dependency treatment has the least extensive useful body of knowledge concerning its effectiveness.25 Yet according to the committee's analysis of a 1987 national survey of drug treatment providers (detailed in Chapter 6), the order of expenditures for these modalities is exactly the reverse of the order of knowledge about their effectiveness.

TABLE 5-6. Comparison of Types and Strength of Evidence on Effectiveness, Numbers of Admissions, and Revenues for the Major Drug Treatment Modalities.


Comparison of Types and Strength of Evidence on Effectiveness, Numbers of Admissions, and Revenues for the Major Drug Treatment Modalities.

In the final section of this chapter, the committee offers its ideas on how to go about repairing the sources of its disappointment. These ideas are presented as a series of specific research recommendation. First, however, the committee summarizes below its findings about the respective modalities.

Methadone Maintenance

Methadone maintenance is a treatment for extended opiate dependence (which is usually heroin). A sufficient daily oral dose of methadone hydrochloride, which is a relatively long-acting narcotic analgesic, yields a stable metabolic level of the drug. Consumption once daily of a stable, clinically adjusted dose is not behaviorally or subjectively intoxicating and does not impair functioning or generate appreciable morbid side effects. Once a newly admitted client reaches a stable, noneuphoric "blockade" state, free of the psychophysiological cues that precipitate opiate craving, he or she is amenable to counseling, environmental changes, and other social services that can help shift his or her orientation and lifestyle away from drug-seeking and related crime toward more socially acceptable behaviors.

Methadone maintenance has been the most rigorously studied modality and has yielded the most incontrovertibly positive results. However, it is also the most controversial treatment, largely on the grounds that methadone clients have "merely" switched their dependence to a legal narcotic and that some clients (the proportion varies from program to program) continue to take heroin and other drugs intermittently and to commit crimes, including the sale of take-home methadone. In the committee's judgment, these controversies and reservations are neither trivial nor compelling. The great majority of methadone clients had been consuming high levels of illicit drugs and committing other crimes (including drug selling) on a daily basis prior to admission. The issues are to what extent undesirable behaviors are reduced and positive behaviors increased as a result of methadone maintenance (in comparison to no treatment or to alternative measures), and whether enough is known about such treatment to improve poorly performing programs.

Research on methadone has demonstrated the following:

  • There is a strong evidence from clinical trials and similar study designs that, on average, heroin-dependent (or other opiate-dependent) individuals have much better outcomes in terms of illicit drug consumption and other criminal behavior when they are maintained on methadone than when they are not treated at all, when they are simply detoxified and released, or when methadone is tapered down and terminated arbitrarily.
  • Methadone clinics have significantly higher retention rates among opiate-dependent populations than do other treatment modalities for similar clients.26
  • When assessed following discharge from methadone treatment, clients who stayed in treatment longer have better outcomes than clients who left earlier.
  • Methadone dosages need to be clinically monitored and individually optimized. Clients do much better, however, when they are stabilized on higher rather than lower doses within the typical ranges that are currently prescribed (30-100 mg/day). Program characteristics such as inadequate methadone dosage levels and differences between counselors (which are not yet fully defined) are significantly related to differences in client performance while in treatment.
  • Methadone treatment, when implemented at the resource levels observed in the late 1970s, provides individual and social benefits over a term of at least several years that are substantially higher than the cost of delivering this treatment, which is no $3,000 per year and which should be at least $4,000 per year to be comparable to earlier programs. The daily continue drug taking at a lower level.

Therapeutic Communities

Therapeutic communities are residential programs with expected stays of generally 9 to 15 months, phasing into independent residence with continuing contact for a variable period. TC programs are highly structured blends of resocialization, milieu therapy, behavioral modification practices, progression through a hierarchy of occupational training and responsibility within the TC, community reentry, and a variety of social services.

Therapeutic community clients are more diverse in their drug use patterns than methadone clients because the modality is not specific to any particular class of drugs. From the 1960s to the early 1980s, a majority of TC clients were primarily dependent on heroin. In the late 1980s, cocaine dependence began to predominate in many programs. Therapeutic communities are designed for individuals with major impairments and social deficits, including histories of serious criminal behavior. The results of research on the effects of TC treatment are as follows:

  • TC clients end virtually all illicit drug taking and other criminal behavior while in residence and perform better (in terms of reduced drug taking and other criminal activity and increased social productivity) after discharge than before admission. They also have better outcomes at follow-up than individuals who simply underwent detoxification or who contacted but did not enter a TC program. The length of stay is the strongest predictor clients remaining in treatment. The bias if all admissions, not only the ones remaining in treatment, are compared across modalities. of outcomes at follow-up, with graduates the best outcomes at that point.
  • Attrition from TCs is typically high—above the rates for methadone maintenance but below the rates for outpatient nonmethadone treatment. Typically, about 15 percent of admissions will graduate after a continuous stay; the figure is higher (20 to 25 percent) once later readmissions are considered.
  • The minimum retention necessary to yield improvement in long-term outcomes seems to be several months, which covers one-third to one-half of a typical program's admissions. Improvements continue to be manifested for full-time treatment of up to one year in length.
  • The benefits of TC treatment are substantial and they virtually repay the costs on a day-by-day basis, although the per diem costs are higher than for methadone maintenance: generally, about $13,000 per year-probably $20,000 for a model program-yielding somewhat lower benefit/cost ratios than for methadone but ones that still favor the use of this treatment.

Outpatient Nonmethadone Programs

Outpatient nonmethadone programs display a great deal of heterogeneity in their treatment processes, philosophies, and staffing. Their clients generally are not opiate-dependent but otherwise vary across all types of drugs. Usually, OPNM clients have much less serious criminal histories than methadone or TC clients and include more nondependent individuals. Outpatient nonmethadone programs generally provide one or two visits per week for individual or group psychotherapy/counseling, with an expected course averaging about six months.

Despite the heterogeneity of programs and their clients, the limited number of outcome evaluations of OPNM programs have generated conclusions qualitatively similar to those from studies of TCs:

  • Outpatient nonmethadone clients during and following treatment exhibit better behavior than before treatment. Those clients who are actually admitted to programs have better outcomes than clients who contact but do not enter programs (and clients who only undergo detoxification). Outcome at follow-up is positively related to length of stay in treatment, and completers have better outcomes than dropouts.
  • Retention in outpatient nonmethadone programs is poorer than for methadone maintenance and therapeutic communities.
  • The benefits of OPNM treatment are fewer than for methadone or TCs, but the cost of the treatment, at about $1,350 for six months (about $1,800 for a model program), is low. As a result, the yields are favorable for those who stay longer than three months, and the aggregate program ratios are mildly favorable.

Chemical Dependency Programs

Chemical dependency programs generally are residential or inpatient, with a three- to six-week duration, followed by up to two years of attendance at self-help groups or a weekly outpatient therapy group. CD programs are based on an Alcoholics Anonymous (12-step) model of personal change, a belief that dependence is a permanent but controllable disability, and goals of total abstinence and lifestyle alteration. The proportion of the CD population who are drug involved is similar to the outpatient nonmethadone population in that the primary drugs are cocaine and marijuana. The modal CD client, however, is an older, socially well-supported, alcohol-dependent individual.

CD programs are often located in hospitals, but the core therapeutic elements of this modality do not require the presence of acute care hospital services. There is little evidence on whether hospital-based CD programs are more or less effective for drug problems than CD programs that are not sited in hospitals, or whether they are more or less effective than no treatment at all. Chemical dependency programs treat mainly primary alcoholism and have not been adequately evaluated for treatment of drug problems. A few follow-up studies of individuals who have completed CD treatment indicate that primary drug clients have poorer outcomes than primary alcohol clients. There are no cost/benefit analyses for chemical dependency treatment.


Detoxification is therapeutically supervised withdrawal to abstinence over a short-term—that is, up to several months but usually five to seven days, often employing pharmacological agents to reduce client discomfort or the likelihood of complications. Detoxification is seldom effective in itself as a modality for bringing about recovery from dependence, although it can be used as a gateway to other treatment modalities.

Clinicians generally advocate that detoxification not be considered a modality of treatment in the same sense as methadone, TCs, outpatient counseling, and CD units because of its narrow, short-term focus and poor outcomes in terms of relapse to drug dependence.

Detoxification episodes are often hospital-based and may begin with emergency treatment of an overdose. Much drug detoxification (an estimated 100,000 admissions annually) is now taking place in hospital beds. It is doubtful whether hospitalization (especially beyond the first day or two) is necessary in most cases, except for the special problems of addicted neonates, severe sedative-hypnotic dependence, or concurrent medical or severe psychiatric problems. For clients with a documented history of complications or flight from detoxification, residential detoxification may be indicated. Detoxification may, in the committee's judgment, be undertaken successfully in most cases on a nonhospital residential, partial day care, or ambulatory basis.

Correctional Treatment

Treatment of drug-involved prisoners is fairly common, but at least two-thirds of prison treatment programs are equivalent to outpatient nonmethadone treatment—that is, periodic individual or group therapy sessions. This level of intervention is probably not intensive enough to do much for this group. The other prison treatment programs are similar to stays in a therapeutic community, including separation from the general prison population for the expected 6- to 12-month duration of the program.

Most of the prison drug treatment programs that have been studied, including specialized ''boot camp" or "shock incarceration" facilities, have not been shown to reduce the typically very high postrelease rates of recidivism to drug-seeking and other criminal behavior that occur among untreated prisoners. Nevertheless, a small number of well-designed controlled studies, involving prison TCs and residential programs that have strong linkages to community-based supervision and/or treatment programs, indicate that prison-initiated treatment can reduce the treated group's rate of rearrest by one-fourth to one-half; clear correlations are observed between positive outcome rates and length of time in treatment, just as in studies of entirely community-based modalities. The results have some anomalies and there have been difficulties in sustaining the integrity of prison-based treatment programs, but the results argue that these program should be carefully encouraged.

* * * * *

If a single phrase could succeed in capturing most of the findings in this chapter, it would be an expression that—much like the current treatment modalities—dates from the 1960s: different strokes for different folks. No single treatment "works" for a majority of the people who seek treatment. Each of the treatment modalities for which there is a baseline of adequate studies can fairly be said to work for many of the people who seek that treatment; and enough of them do find the right treatment, and stay with it long enough, to make the current aggregate of treatment programs worthwhile.

Selection of the most appropriate treatment modality by clients or others (e.g., judges, probation officers, employee assistance counselors, family members) is constrained by poor information about programs, location/ transportation issues, waiting lists at some portals and aggressive recruitment at others, and cost questions. In most locations, there is no comprehensive intake (assessment and referral) unit or agency to advise or assign applicants. (This triage feature, which was relatively common in the multimodality programs and municipal treatment agencies of the 1960s or 1970s, was often abandoned in the cost-cutting of the early 1980s). Most of all, the search for the right program is bedeviled by variations in program quality. The signs of poor program performance (particularly of poor response to the prospective client's specific set of problems) are not readily apparent, and the general lack of reliable information about program outcomes does not offer incentives for programs to change for the better.

There is a great deal of room for improvement, and there are indications in the research literature on how to bring that about. Much of Chapters 7 and 8 is informed by the committee's reading of those indications. Before moving to the final third of the report, however, the committee considers it vital to lay out a plan for restocking and expanding the limited store of knowledge it has had to draw on so that if another group is charged with studying the treatment system 5 or 10 years from now, they will not have to be as disappointed as this body was about the knowledge gains in the intervening years. The last section of this chapter therefore presents a brief but systematic template of recommendations for a national program of treatment research.

Recommendations For Research On Treatment Services And Methods

Rebuilding the Research Base

Federal support for drug research, including research on treatment methods and services (alternatively, clinical and services research), surged during the early 1970s, declined steadily in real terms for the next decade, and began to surge again as a result of the Anti-Drug Abuse Acts of 1986 and 1988 and recent initiatives for AIDS-related research (Figure 5-8). Unfortunately, but quite predictably, the base of capable researchers declined during the decade-long period of stagnation, as scientists moved on to other fields and very few new ones entered the drug research area. The number of centers of excellence in treatment-oriented research—active programs generating sound new results on current data—declined substantially; where there were formerly close to two dozen, located in all parts of the country, there are now just a handful in a few major metropolitan centers.

FIGURE 5-8. Annual research obligations of the National Institute on Drug Abuse (in both nominal [current] and real [1989-equivalent] dollars) for fiscal years 1974–1990.


Annual research obligations of the National Institute on Drug Abuse (in both nominal [current] and real [1989-equivalent] dollars) for fiscal years 1974–1990. Source: National Institute on Drug Abuse, unpublished data, 1989.

The national research infrastructure must be rebuilt and the number of local centers of excellence in research on treatment methods and services increased to reverse the shortage of experienced investigators. Current funding increases are sufficient to rebuild the needed base of treatment research excellence but only if the current level is sustained for at least four or five years and expenditures are patterned during that time to ensure attention to the perennial questions that face clinicians and policymakers responsible for the system. It is critical that this base be maintained through a program of steady incremental funding changes and not be dismantled once again, a course that would leave the nation unprepared to respond quickly to whatever new epidemic of drug use might arise in the future—and the lesson of history is that some new wave will arise.

To evaluate and improve the adequacy and effectiveness of treatment plans and expenditures, the national services research program in particular needs rebuilding. The prospects for maintaining and improving treatment quality as well as continuing to develop more effective treatment methods depend to a great extent on treatment services research. The National Institute on Drug Abuse (NIDA), the agency most responsible for maintaining treatment research, is, of course, not autonomous. Its budget and priorities are proposed by the President and disposed by Congress. Moreover, providers of drug treatment services are very much at fault for permitting and in some cases tacitly encouraging the paucity of treatment research over the past decade. Programs have been characterized too much by a fear of failure and too little by the courage of their convictions. The results of earlier treatment enterprises tell an enlightening and reasonably heartening tale, and there is little possibility of improving current therapeutic practices further without careful study of outcomes, not only in research units, with their limited patient protocols and cadre of university-based researchers, but also in all other treatment programs.

Most importantly, the advances in knowledge that came out of clinical and services research in the 1970s have not been followed up, and as a result analysts today are not better prepared to answer questions about the effectiveness, costs, and benefits of current treatment than they were a decade ago. Data systems and analytic capabilities that were designed to answer policy questions have not been well maintained. It would be a travesty of prudent governance if once again the federal government and the states were to proceed to build, or rebuild, a major instrument of national drug control policy without assuring themselves and the taxpayers that there would be timely, necessary research and evaluation to understand that instrument's performance and facilitate its improvement.

One more note needs sounding in this context. A critical longer term role is played by basic epidemiological, behavioral, biological, and neurochemical research to address such issues as the role of genetic predispositions in addiction, the factors that contribute to the plasticity of addictive behavior, the effects of social factors, and methods for reducing drug craving. The goal of such work should be to integrate the biological and behavioral sides of the drug problem. This integration will remain difficult so long as a continuing imbalance persists between substantial investments in high-quality biomedical research and meager ones in high-quality biobehavioral and psychosocial research.

Major Research Questions

The core questions that need to be addressed for the various modalities of public treatment are the following: What client and program factors influence treatment-seeking behavior; treatment retention and efficacy, and relapse after treatment? How can these factors be better managed? Treatment-seeking factors include community outreach, health promotion and disease prevention efforts such as experimental needle-exchange programs, family and employer interventions, and program intake and triage procedures. Retention and efficacy factors include optimal treatment durations and schedules, pretreatment motivations, counselor or therapist behavior, incentives and conditions of employment, clinic procedures, criminal justice contingencies, and ancillary services. Posttreatment factors include relapse prevention interventions, abstinence monitoring, and environmental reinforcement.

The questions need to be attacked in a variety of ways. Despite the difficulties of maintaining the integrity of controlled experiments in treatment programs, these studies provide the most incontrovertible evidence about comparative treatment effects, and efforts to conduct them should be strongly encouraged. A more detailed understanding of treatment processes through ethnographic and case study methods is also badly needed. This work is the basis for the design and interpretation of survey instruments. Studies should be initiated within as well as across each major treatment modality to answer the following question: What are the relations of treatment performance (that is, differential outcomes, taking initial client characteristics into account), the content and organization of treatment (specific site arrangements, service offerings, therapeutic approaches, staffing practices), and the costs of treatment?

Services Research

Health services research is a critical element in building treatment systems. An important foundation for services research as well as program accountability is the development, maintenance, and analysis of a system of data acquisition on treatment programs, client performance, and costs. Results from studies that use these kinds of data will permit better and more cost-effective decisions about facility characteristics, staff salary and training levels, services coordination methods, intensity of services, reasonable charges, and other components. Systems of this sort were established in the 1970s but were effectively disassembled as a matter of federal policy in the 1980s. Treatment data acquisition system must be rebuilt and effective managed and utilized if the improvement of treatment knowledge and practice is to be evaluated and facilitated in the 1990s. Data on treatment effectiveness and costs should become the cornerstone of decisions about treatment coverage by public and private programs.

NIDA, in conjunction with its sister agency, the Office of Treatment Improvement, needs to give more adequate, focused attention to the drug treatment delivery system as a whole. Stronger services research programs at NIDA are a critical complement to the research and service responsibilities of the Alcohol, Drug Abuse, and Mental Health Administration (ADAMHA). Fulfilling this responsibility requires close linkages to practice and thus some responsibility to and for service delivery. Existing legislative authority directing these linkages should be implemented fully.

The responsibilities for research coordination, however, do not stop at the boundaries of ADAMHA. Collaborative and coordinative arrangements with the National Institute of Justice, the Bureau of Justice Statistics, the National Institute of Corrections, and other relevant agencies in the Department of Justice and other federal departments should be extended beyond current levels. More extensive relationships would encourage critical technical improvements, such as the inclusion in epidemiological and treatment surveys of "linkage" items to facilitate syntheses with data from criminal justice populations. For example, treatment applicants should be asked how many emergency room admissions and arrests they have undergone during the year prior to treatment, which would not only serve to build baseline data for outcomes research but also provide calibrations with respect to the Drug Abuse Warning Network and Drug Use Forecasting data systems.

Some of the most compelling results of treatment research have come from large longitudinal studies involving thousands of clients: the DARP (Drug Abuse Reporting Program) study of a 1969–1971 national admission cohort, which included a 12-year follow-up, and TOPS (the Treatment Outcome Prospective Study), which involved a 10,000-person national sample of 1979–1981 admissions to 41 drug treatment programs in 10 cities. There is reason to believe that some findings about the treatment modalities—such as the importance of time in treatment—will prove robust in the face of changing drug markets, but others may not.

Another such national treatment sample study (DATOS, or the Drug Abuse Treatment Outcome Study) is beginning in 1990, and some smaller scale studies, such as the Drug Services Research Survey, are in process. Intervals of 10 years between entry cohorts to major studies as important as these are far too long. New study panels composed of 3-year entry cohorts (an efficient period of admission to a multiwave design) should begin at no greater than 5-year intervals.

The responsibility to study treatment services in the field generally is not met by demonstration grant programs. Demonstrations have historically functioned as a stop-gap measure to provide a new kind of service for which there seemed to be a need but no certain knowledge about how to fill it—knowledge lacking at least in part because adequate research systems were not already in place to generate it. Demonstrations are not a reasonable substitute for a strong program of treatment services research. Demonstration grants should be made only when objectives are carefully specified and independently designed and performed collaborative evaluations are funded. Collaborative clinical trials are the basis for developing standardized protocols in other forms of treatment and should be implemented as models for demonstration programs. Such a plan would allow effective programs or program components to be adequately described, replicated and, if found useful, incorporated into certification standards.

A services research issue worth noting here is the difficulties that drug treatment programs experience in securing zoning approval for clinical facilities, a problem usually summarized as "not in my back yard" (NIMBY). This problem, of course, is not confined to siting community drug treatment programs but confronts public utilities and services of many kinds. There is currently a NIDA-sponsored market research project (Technical Assistance & Training Corp., 1989) to create technical assistance materials to overcome this "barrier" to treatment. Research support for more definitive studies of program site effects—for example, on local real estate values and criminal victimization rates—would provide a better foundation for this work.

Chemical Dependency

Chemical dependency programs are the least well studied of the drug treatment modalities. The aggressive marketing deployed by many such programs has created suspicions about them in many quarters that cannot be allayed without investments in objective treatment research and evaluation. The optimal site of delivery and length of programming, including the duration of intensive treatment and aftercare periods, and the specific therapeutic elements necessary for an effective program should be investigated more closely.

Only a few chemical dependency treatment providers have played positive roles in providing data and research opportunities for effectiveness studies. Many more need to do so to answer these questions: What is the effectiveness of chemical dependency treatment for drug-impaired clients of varying characteristics? Are there variations in program effectiveness—and if so, why? What are the actual costs and benefits of the most effective components of chemical dependency treatment?

Cocaine Treatment

The major efforts to date to investigate cocaine treatment efficacy occurred prior to the epidemiological reemergence of cocaine in the 1980s. There is reason to believe that some findings about treatment modalities—such as the importance of time in treatment—will prove robust in the face of changing drug markets, but others may not. The infrastructure of treatment research centers decayed during the stagnation of drug research funding, and as this capability is rebuilt, it should specifically address the following questions about cocaine treatment: What are the most effective treatment elements for cocaine dependence and abuse? To what degree can current modalities be effective for crack-cocaine? What new or existing pharmacological and nonpharmacological treatment elements can improve the clinical picture?

Women, Children, and Adolescents

The majority of individuals in treatment are adult males who are 20 to 40 years old, and their responses dominate treatment research statistics. The major findings of research to date on the effectiveness of different modalities and elements of treatment seem to apply roughly as well to adolescents and women with young children as they do to the more prevalent demographic groups (Hubbard et al., 1989). Yet the potential significance of child-bearing and child-rearing women and adolescent clients in terms of the future benefits of present treatment (or the future costs of present nontreatment) is great. Research plans in all areas need to devote special attention to differentiated knowledge about the two populations of adolescents and women with young children (including pregnant women).

It seems clear from earlier studies that women in treatment who are pregnant or have young children are especially likely to bring particular needs to the treatment system (Beschner et al., 1981; Reed et al., 1982). For example, drug-abusing or dependent women on average have poorer self-esteem than men and suffer from greater anxiety, depression, and detachment; as a result, therapists who rely too heavily on confrontative techniques may worsen such problems rather than help reduce them. Because of their child care responsibilities, long-term residential treatment in TCs may be ruled out for many women unless there are special provisions for child care. In many states, long-term TC treatment becomes doubly problematic because extended residential treatment may jeopardize family eligibility for Aid to Families with Dependent Children (welfare) or threaten the mother s custody of the children.

The federal block grant for alcohol, drug abuse, and mental health services mandates that 10 percent of the grant be set aside to provide special services for women. According to the Institute of Medicine analysis of the 1987 National Drug and Alcoholism Treatment Utilization Survey, about one-third of the more than 80,000 women in drug treatment were in programs that had at least some special services for women, although there is no further specification of the nature or extent of these services. Both clinical and services research are needed to gain an understanding of the nature and efficacy of current practices and the potential of innovations.

The state of knowledge about adolescent treatment is, if anything, even less satisfactory. The number of useful studies of adolescents is small, and most work in this area is based too heavily on studies of treatment in much earlier periods (e.g., Friedman and colleagues [1986] analyze data on adolescents in the Drug Abuse Reporting Program of the early 1970s). There are major obstacles to research on adolescents, including conceptual issues, such as discordant terminology for adolescent treatment service components, and logistical constraints, such as unmanageable requirements for obtaining parental consent.

The committee recommends that a special study initiative be undertaken by the National Institute on Drug Abuse, in conjunction with other relevant agencies of the Public Health Service, on the treatment of drug abuse and dependence among adolescents and women who are pregnant or rearing young children. he initiative should review and summarize all available sources of evidence and insight from research and clinical experience, provide as much guidance as possible for current treatment efforts, and develop a comprehensive research agenda. The agenda in turn should be pursued by research agencies of the federal government and other sources of research support and carried forward by the community of clinicians and scientists.



Because methadone maintenance programs are virtually always conducted on an outpatient basis but are set apart by the specific reference to methadone, all other outpatient programs are conventionally lumped together as outpatient nonmethadone or outpatient drug free. In light of the frequent use of other psychotropic medications during outpatient treatment, the committee views the terms "nonmethadone" as more accurate than "drug free." The lumping together of all patient nonmethadone treatment is testimony to the prominence and distinctive nature of methadone maintenance and the fact that the population it serves is sufficiently homogeneous and different from the populations served by other outpatient programs. It should also be noted that methadone may be used in modalities other than maintenance, which technically refers to a planned treatment duration of 180 days or longer. (Shorter periods—usually 3 weeks to 2 months—are considered methadone detoxification.) Planned methadone-to-abstinence tapers of longer than 180 days are also incorporated into some programs plans.


Although CD programs incorporate numerous therapeutic components in addition to Alcoholics Anonymous-type meetings, the 12 steps of the Anonymous creed are so fundamental to the CD modality that the latter has been referred to as the "professionalization of Alcoholics Anonymous." There is no scientific literature on the Oxford House approach, which combines residential proximity with the fellowship principles.


There are three main types of narcotic analgesics: those derived from opium, such as morphine, heroin (diacetylmorphine), and codeine, and the two major synthetics, meperidine (best known as Demerol) and methadone. There are numerous congeners of each major narcotic type that have varying degrees of activity. The natural and synthetic compounds have dissimilar chemical bases but share certain critical structural properties that result in their penetrating and affecting the "endogenous opioid" neurotransmitter system in similar ways. There are significant differences, however, in how the major narcotic types are absorbed and metabolized outside the brain; these difference affect the duration and rate of their central nervous system effects.


There continue to be widespread negative beliefs among the general public and some policymakers about methadone (see, for example, the results of focus group discussions reported by the Technical Assistance & Training Corporation [1989]). The drug is suspected, for example, of being unsafe even in clinically controlled usage; it is said to "rot" the bones (or the brain, or the liver) and to create lassitude or stupefaction among individuals who take it for any length of time or at any dose except a minimal one. It is also said that indefinite maintenance is "just substituting one addiction for another," so the most important clinical goal should be to "get off methadone" as soon as possible. It is thought that most of the people enrolled in methadone maintenance programs sell some or all of their daily methadone dose and use the proceeds to buy heroin and other drugs, Putting all of these beliefs together, methadone can be portrayed as an assault on the well-being of communities in which methadone maintenance clinics are located, rather than a therapeutic response to local drug problems.

This set of beliefs about methadone is based partly on shards of experience (often reported by journalists), partly on philosophical or ideological premises that may be impervious to evidence, and partly on frank skepticism about the existence of a therapeutic rationale or base of evidence underpinning methadone maintenance treatment. This section should at least be useful in addressing the last of these sources of belief.


The argument has been made that even illegally marketed methadone represents a significant public health improvement over street heroin. Although this result is theoretically plausible, an opposite result is equally plausible, and their is little evidence to support either theory. Therefore, in policy terms, street methadone sales are a negative effect.


There was extensive research from the late 1960s to the late 1970s on a longer acting methadone congener, levo-alpha-acetylmethadyl (LAAM), that requires less frequent doses—every two or three days instead of daily. LAAM has been studied in a series of phased clinical trials but has not yet been approved for nonexperimental use, although its safety and freedom from toxic side effects appear comparable to those of methadone (Savage et al., 1976; Ling et al., 1978; Blaine et al., 1981). Overall, during the trials, methadone was more successful than LAAM in retaining clients in treatment (by 20 percentage points), largely because more LAAM recipients felt that the medication was not ''holding," that is, not keeping opiate withdrawal symptoms from beginning to emerge between doses, a result that Goldstein and Judson (1974), after a double-blind study, judged to be more psychological than physiological in origin. LAAM recipients who stayed in treatment used less heroin and performed better on other clinical measures than methadone clients, particularly those on lower methadone doses. Some clinicians reported a substantially improved therapeutic climate in LAAM clinics owing to the more relaxed three-days-per-week visiting schedule (Goldstein, 1976). There are probably clients who would do better on LAAM than on methadone, and vice versa, with results for both likely to improve with better dose optimization and counseling about differences between the two drugs. A revival of interest in LAAM and an attempt to restore the initiative toward approval by the Food and Drug Administration for nonexperimental use are under way.


Observational studies of the original Dole-Nyswander program cohorts, which probably engaged the most highly motivated clients and had relatively high-quality staff and resources, yielded good data over time confirming the long-term efficacy of methadone maintenance (Dole et al., 1968; Gearing, 1970, 1974; Dole and Joseph, 1978). Studies in these and later programs also indicated the close relation of retention in treatment and good outcomes; attrition after a short period in treatment was associated with higher rates of relapse (Simpson et al., 1979; Hubbard et al., 1989).


One other significant experimental study was reported by Newman and Whitehill (1978) from Hong Kong. This study demonstrated both the attractiveness or retentive power of methadone as such and the difficulties of conducting randomized clinical trials with drug-dependent populations when they are able to act on their own strong preferences about treatment assignment. (Another illustration of that difficulty in the United States was reported by Bale and coworkers [1980].) Newman and Whitehill studied 100 male heroin addicts who were seeking methadone maintenance. The men were hospitalized for one week and stabilized on 60 mg/day of methadone. They were then randomly assigned to ambulatory methadone maintenance or to slow detoxification. The maintenance group started out at 60 mg/day and ended by averaging 97 mg/day. The detoxification group was taken down 1 mg/day over 60 days, after which they were given placebos. The medication was given on a double-blind basis: neither patients nor clinicians had certain knowledge of which group they were in.

About 60 percent of maintenance patients were retained in treatment for the entire 2.5-year trial period, a rate commensurate with retention studies in the United States. In contrast, the patients who were detoxified dropped out of treatment rapidly. By the time they reached the placebo state, only 20 percent remained in treatment; nearly all had dropped out by the end of a year. Dropouts from the control group were subsequently recruited into methadone maintenance and had the same retention rates from that point on as the original maintenance group. Most of the control group sensed that they were being detoxified rather than maintained, and many quit the study to reenroll in methadone maintenance.


Only four counselors participated in the McLellan study, however.


Rufener and coworkers (1977a) examined the cost-effectiveness of three major treatment modalities (methadone maintenance, TCs, and outpatient nonmethadone) based on an analysis of the DARP data base (Sells, 1974a,b). Methadone maintenance was decidedly the most cost-effective treatment in terms of lowest cost per opiate-free days, non-opiate-free days, days not spent in criminal activity, and legitimately employed days. Goldschmidt (1976) similarly compared methadone maintenance and therapeutic communities, but his effort identified the benefits of both the in-treatment and posttreatment periods. He concluded that methadone and therapeutic communities produced similar ''effectiveness units" (percentage of addicts meeting success criteria). The cost advantage of methadone, however, made its cost-effectiveness about twice that of therapeutic communities.


This estimate derives from experiments such as that of Bale and colleagues (1980), which is described in the following section, and from national surveys of the treatment system, described in Chapter 6, combined with estimates of the prevalence of heroin dependence.


TC clients were 57 percent white, 34 percent black, and 9 percent Hispanic. Methadone clients were 16 percent white, 58 percent black, and 26 percent Hispanic (Sells, 1974a).


The causal model here attributes the client's status at a later point in time to three kinds of factors: predisposing conditions, which are controlled for by the baseline measurement procedures (e.g., why the client sought treatment, how much recovery the client wants to achieve); exterior factors during treatment, which are assumed to affect clients more or less at random; that is, they are not correlated with being admitted to treatment (changes in the price of drugs, for example, or police attitudes toward an individual, or the likelihood of being caught in a job layoff); and the units of treatment received, the element whose effects the researcher really wants to measure. There are three corresponding sources of error: unmeasured predisposing conditions, exterior factors that are correlated with being in treatment, and variations in the consistency of treatment units.


The problem of heroin-dependent individuals' noncompliance with experimental and control


There were 710 total drug detox admissions; exclusions from the study sample were for pending felony charges (51), major psychiatric problems (41), falsified eligibility for VA treatment (13), and miscellaneous reasons (19). The study population differed from the opiate-dependent DARP sample in several important particulars: they were all honorably discharged veterans (100 percent versus 25 percent in the DARP), all male (100 percent versus 77 percent), and mostly high school graduates (71 percent versus 39 percent) and ex-convicts (80 percent versus 60 percent) who used other drugs in addition to heroin (72 percent versus 52 percent). They were also less often black (41 percent versus 54 percent) or Hispanic (13 percent versus 22 percent) and more often employed (76 percent versus 57 percent) in white-collar jobs (36 percent versus 11 percent).


Retention in treatment was not high for the 6-month residential programs. About 13 percent of the clients stayed less than 1 week (these were considered "no treatment"), 57 percent dropped out within 7 weeks, and 85 percent left treatment before 6 months. In contrast, about 65 percent of clients entering methadone maintenance were continuously in treatment for the follow-up year.

The TC group was therefore divided at the median length of stay (for all admissions who had remained longer than a week), which was 50 days. The short-term group stayed in treatment about 3 weeks on average; the long-term group stayed about 20 weeks on average. The methadone group on average stayed in treatment about 40 weeks.


One smaller study that is notable for its careful execution followed a random sample of graduates and dropouts from a Connecticut TC (Romond et al., 1975) with an 18-to 24-month treatment plan. The authors found few pretreatment differences between the graduate and dropout groups except that women were much less likely than men to graduate. All 20 graduates in the sample were successfully contacted; 10 of 31 dropouts in the sample were not located, and 1 refused an interview, yielding 20 successful contacts. Graduates had spent on average 21 months in treatment, compared with 5.7 months for dropouts (range: 10 days to 16 months). Interview data were corroborated through formal and informal community networks.

Graduates had consistently better outcomes. Only 1 of 20 graduates relapsed to dependence for some part of the follow-up period, another 5 sometimes used nonopiate drugs, and 14 remained drug free throughout the interval; altogether, graduates spent 0.5 percent of the follow-up period dependent. Of the 20 dropouts interviewed, 14 relapsed to dependence for some of the follow-up period, 2 more used nonopiates occasionally, 1 was incarcerated for the entire period, and 3 had used no drugs; 35 percent of the dropouts posttreatment time was spent as drug dependant. Ninety-four percent of the graduates' posttreatment time had been in school or employed, and at the time of the interview, none were institutionalized and 2 had some criminal justice involvement (probation, parole, pending case). Forty percent of the dropouts' posttreatment time had been in school or employed; at the time of the interview, 4 were in other drug programs, 1 was in a psychiatric hospital, 4 were in jail, and 7 others had pending court case or were on probation or parole.


This study reported on 83 individuals who responded to advertisements offering referral to cocaine treatment and who then self-selected a CD program, an outpatient program, or no treatment following an education/information session. The study found no significant differences between the CD and no-treatment groups eight months later.


Studies of Hazelden drug clients (Laundergan, 1982; Gilmore, 1985) are too limited methodologically to merit detailing, which is unfortunate, given the prominence of this program. The findings are consistent with CATOR and CareUnit results in indicating lower abstinence rates at follow-up for drug clients than for alcohol-only clients.


Diagnoses of abuse ordinarily do not call for detoxification procedures, although in occasional cases of abuse there is reason for 2 to 24 hours of medical observation to monitor clearing of severe intoxication or acute overdose (possibly needing emergency intervention if vital signs are poor). These treatments of single episodes of excessive dosing may be thought of as logical counterparts to detoxification from dependence, but they are not detoxification.


Detoxification of barbiturates is particularly liable to involve seizures and is more likely than other drug withdrawal to need management in a supervised environment—a hospital or other residential facility with appropriate medical staff and equipment.


These elements could also apply to community programs, which is not surprising because many of the same clients are seen in both program settings.


The net figure in Table 5-5 for the whole Cornerstone group—that is, the 54 percent who were convicted of new crimes after release—is the yield within the combined group of dropouts (less than one month in the program) and graduates. In private communications with the Cornerstone staff, the committee was told that most dropouts from the program leave within the first few weeks; therefore, ignoring the small numbers who dropped out between the first month and graduation, for whom follow-up data have not been published, would not appreciably modify the above result. Those in the parolee comparison group, according to Field (1984:54), ''do not have the chronic substance abuse histories nor the chronic criminal histories of Cornerstone graduates [and therefore] would be expected to do better at avoiding criminal recidivism than Cornerstone graduate—except, of course, for the treatment results."


As at Lexington, voluntary as well as criminal commitments to the facility were permitted. Most CAP clients, however—about 70 percent before 1970 and 93 percent afterward—had been convicted of felonies, largely for non-drug offenses such as burglary and robbery. In addition, before 1970, about 15 percent had been referred by police officers on the non-criminal basis of "believed addicted."


Correctional treatment has not been included on this chart because it is not a distinct modality. Knowledge about prison-based programs is approximately at the same level as that for community-based TCs. Detoxification also is not included because it is not considered a treatment for drug-seeking behavior in the same way as are the major modalities.


It should be noted that higher retention can "load the dice" when during treatment among different modalities. Because dropouts generally show worse behavior and have somewhat poorer prognoses to begin with, a program that retains more of its initial clients, even if equal in its effect on each client, will have a lower average effectiveness on

Copyright © 1990 by the National Academy of Sciences.
Bookshelf ID: NBK235506


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (3.0M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...