Dan Waldorf & Patrick Biernacki

Dan Waldorf, M.A., has worked in the field of addiction for over ten years. His initial work was with a study of addict careers in New York City at Columbia University which culminated in the book, Careers in Dope. Since that time he has been involved with numerous evaluations of drug treatment programs, a historical study of the Shreveport, Louisiana morphine maintenance clinic. and an ethnographic study of a group of cocaine users in San Francisco. Most recently, he has been working as a consultant to a P.C.P. ethnographic study and is Co-Principal Investigator of the Recovery Project, an exploratory study of natural recovery of heroin addicts.

Patrick Biernacki, Ph.D. co-author of the Natural Recovery article, teaches Sociology at San Francisco State University. He has conducted a research study that analyzed the life careers of 1400 addicts who had been civilly committed to the California Rehabilitation Center. He has also directed research projects in the areas of alternative education, criminal justice and consumer affairs. At the present time he and Dan Waldorf are conducting a study of the processes involved in the “natural” termination of heroin addiction careers.

This paper reviews the literature in an area which has received little attention of drug researchers spontaneous remission. The paper reviews all the research studies that have looked at the phenomena of the “natural” recovery from heroin addiction natural in the sense that some addicts manage to stop using heroin and not become re-addicted without the help of treatment intervention. Some areas for future research are also suggested.


Conventional wisdom among clinicians and researchers in the field of drug abuse and addiction is that heroin addicts seldom, if ever, overcome addiction without treatment. Occasionally researchers have speculated that there may be something akin to spontaneous remission among addicts, but until recently it was thought that the numbers and percentages of such recoveries were very small (5-15%) and insignificant. New evidence suggests that the rate of natural recovery may be much higher than expected. Furthermore, new studies suggest that addicts who do not go to treatment recover at approximately the same rates as those who do go to treatment.

The Incidence Literature

The first evidence to suggest natural recovery came from Charles Winick’s famous “maturing out” study published in 1962. Winick traced the official records of addicts in files of the Federal Bureau of Narcotics and found that age was associated with such traces. As addicts approached ages 35-40 years they tended to drop out of the files which suggested to Winick that some life cycle processes were involved. He postulated that addicts gave up their addiction just as some adolescents matured out of juvenile delinquency.

There are, however, problems with Winick’s study; he did not know exactly what happened to persons who were no longer in the file and assumed that they had given up their addiction. A 1973 report of George Vaillant’s longitudinal study of 100 New York addicts (originally admitted to Lexington Hospital and followed for 20 years) questions Winick’s assumption of recovery. Vaillant found that ” . . . more than half of the actively addicted men of [his] study [were] able to go for five years or more without being reported to the Federal Bureau of Narcotics and Dangerous Drugs,” and that “Over 25% of active addicts went for five years without being reported to the New York Narcotics Register.” (Vaillant, 1973) These data suggest that Winick’s assumption of recovery may not be completely justified. The next study to suggest natural recovery was conducted in 1964 and 1965 and published in 1966. Robert Scharse working in the East Los Angeles Halfway House asked known addicts in the program to identify and locate friends who had used heroin with them but had since given it up. Scharse identified 71 ex-users by this means and interviewed 40 of them in a dual interview situation (both the addict and the ex-user). He found that at least 9 of the 40 interviewed reported that they had experienced physical dependence from heroin and had recovered without going to treatment. (Scharse, 1966)

Social survey data amplified the exploratory studies of Scharse and Winick in 1967. Lee Robins working out of Washington University in St. Louis published the results of a social survey of a sample of black males born during 1930-1934 in St. Louis and who attended schools in that city. This was the first study of drug use of a non-treatment sample (called “normal” by Robins) and she found that 10% (22) of the 235 men interviewed had been addicted to heroin while 4% (9) had been to Lexington and Ft. Worth Hospitals for drug treatment (at that time there were few other treatments available). Of the 22 persons reporting heroin addiction only 16% (4) reported heroin ule during the previous year (1964-1965); 2 of the 4 had been to treatments (or 22% of the 13) and the remaining (15% of the 13) had not been treated. Put another way, 78% of the treated and 85% of the untreated addicts reported no heroin use for the previous year. (Robins, 1967) These findings were so unusual and so much at odds with the accepted knowledge of addiction at the time that many persons were cautiously skeptical.

This skepticism subsided somewhat in 1973 when Robins published her milestone study of returned Vietnam veterans. Startled by reports of widespread heroin use in Vietnam during the war, the federal agencies (more specifically the Special Action Office for Drug Abuse Prevention) commissioned a study of returned veterans in 1972 to learn more about their drug use in Vietnam and also since returning. A sample of 898 men who had returned from Vietnam during September 1971 were interviewed in 1972 from 8-12 months after their return. Of the 898 it was found that nearly one in two had used narcotics in Vietnam (45%) and one in five (20%) had been addicted to heroin. After returning only 10% reported using narcotics between the time of their return and the interview and only 1% had been re-addicted. At the time of the interview only 2% (8% of those addicted in Vietnam) reported to have been currently using narcotics and 1% were detected to have used opiates through urine analyses.

Research findings concerning the differences between treated and untreated addicts were not as expected. Veterans who did not get treatment for their addiction did just as well upon return as those treated. More specifically the study found that 37% of the treated and 49% of the untreated veterans who were dependent and detected (DEROS urine samples) narcotics users (186) were drug positive at the interview. 48% of the treated and 24% of the untreated veterans who were dependent but not detected narcotic users (76) were drug positive and; 13% of the treated and 16% of the untreated persons who claimed never to have been dependent in Vietnam (12) were drug positive. These findings were even more startling than those of the first Robins study. The idea that addiction was nearly always a long lasting phenomena and the old myth 11 once an addict, always an addict” had to be abandoned in the face of these findings. Very clearly, the natural history of addiction had to be re-thought to account for these new data.

Discussing the findings in 1975 she and her co-authors said:… it does seem clear that the opiates are not so addictive that use is necessarily followed by addiction nor that once addicted, an individual is necessarily addicted permanently. At least in certain circumstances, individuals can use narcotics and even become addicted to them but yet be able to avoid use in other social circumstances. (Robins, Helzer and Davis, 1975)*

Still another large scale survey also lends support to the findings of the two Robins’ studies. In 1976 John O’Donnell and researchers from the Universities of Kentucky and California (at Berkeley) published preliminary results from a survey of 2,510 males taken in 1974 and 1975. (O’Donnell, Voss, Clayton, Slatin and Room, 1976). From a sample of all the males in the United States born between 1944-1954 and known to draft boards, O’Donnell and his associates found that 6% of the sample had used heroin and 2% were considered heavy users (using 100 times or more). Only 20 of the heroin users reported going to treatment; this number constituted 13% of all the heroin users. Half the heavy heroin users had been to treatment. Comparing those who had been to treatment with those who had not the authors found major differences in current heroin use (use during 1974 and 1975):

… 65 percent of the men who had been treated for heroin use were currently using it, in contrast with 27 percent of the men who had never been treated for use of heroin. These data suggest that users who enter treatment comprise those least likely to succeed in terminating the use of heroin. (O’Donnell, Voss, Clayton, Slatin and Room, 1976) Unfortunately, O’Donnell did not organize his data on the current use of treated and untreated users into addicts or non-addicts, heavy or light users; as a result, we can not tell from his presentation the extent of the prior heroin use of the non-treated sub-sample. It could be that the majority of the non-treated users were light or experimental users rather than heavy users or addicts. We expect that the authors will clarify this confusion in future analyses.

Still other sources of data that suggest natural recovery are two large scale treatment evaluation studies that employed control groups. These studies are the Macro Systems, Inc., evaluation of the New York City Addiction Services Agency (A.S.A.) programs and the Burt Associates study of the effectiveness of the National Treatment Association (N.T.A.) of Washington, D.C. Prior to these two studies, evaluations of drug treatment did not to the best of our knowledge incorporate control groups against which the treatment groups could be compared. Consequently, there has been no base to compare the relative outcomes of treatment groups and little information about the remission of drug users who do not partake in long-term treatment programs. The first evaluation conducted by Macro Systems, Inc., a profit research firm, followed up a sample of 462 persons who had been to a variety of A.S.A. treatment programs (during the last 6 months of 1971). One-third of the sample (156) were persons who had stayed in treatment less than 10 days (and had not undergone any subsequent treatment); this group was designated as the control group. (Macro System, 1975) Three years after entry in treatment the evaluation findings showed that narcotics use by the control group was no greater than it was for those who had been to treatment. Using an index of narcotics use as a basis of comparison they found that the controls had a score of 0.29 while those in treatment from 10-90 days had 0.21 and those persons who had been in longer than 90 days had a score of 0.20. The differences between the three scores were not statistically significant. The authors summarized: These findings have an iconoclastic tenor insofar as they challenge widely held orthodoxies and substantive implications upon the future course and direction of drug treatment efforts. The findings, however, are not consistent with theories related to the natural history of addiction, the healing effect of time, and the inner psychological motivation of drug users. (Macro Systems, Inc., 1975)

These assertions must be tempered, however, in light of some methodological problems with the study. Macro Systems had a low interview completion rate-they initially claimed to have had completed interviews with 74% of the sample but a subsequent report indicated a much lower completion rate of only 61%. (Burt Associates, 1977) Ile researchers had considerable difficulty in locating and interviewing Puerto Ricans in the sample, particularly those living in the South Bronx, and as a result Puerto Ricans were underrepresented in the interviewed group and this may have biased the findings.

Burt Associates in their evaluation of the National Treatment Association programs used a similar design in that they also employed a control group. They successfully located and interviewed 81% of an initial sample of 360 persons who had previously been to treatment one to three years earlier. One-third of those interviewed were persons who had stayed in treatment five days or less and were used as the comparison or control group. One in five (29%) of the total sample were considered “fully recovered” at the follow-up interview, while 37% were considered ‘partially recovered”. Full recovery was defined by the study as persons who two months before the

1. used no illicit drugs (except marijuana),

2. had not been arrested or incarcerated and

3. who were employed, in school, or job training or a

Partial recovery was more complexly defined but usually included one negative response to the arrest and employment criteria or some daily illicit drug use. When the treatment sample was compared with the comparison sample, no significant differences were found between the two. The comparison sample defined as the non-treatment group did just as well in terms of the definitions of recovery as did the treated group. Furthermore, time in treatment had no particular association with outcome; people who stayed in treatment one day did just as well as those who stayed a year, two years, or five years. (Burt Associates, 1977)

The next evidence to suggest natural recovery comes from an imaginative study of a heroin epidemic that occurred in two high schools in a San Francisco Bay Area suburb. Using known heroin users as informants the authors of the study, David and Kathleen Graeven, asked them to identify persons they had used heroin with from high school yearbooks. The study identified 294 heroin users and interviewed 120 of them. Of those interviewed 51% (76) were found to have been addicted (used heroin daily and experienced dependence) and of those addicted 53 (71%) had been to some drug treatment. (Graeven and Graeven, n.d.)

Comparing the treated with the untreated addicts the authors found that: By 1975 the percent of untreated addicts not using heroin (52%) was quite similar to those who had only one or two treatment experiences (50%), however, those with more treatment experiences were more likely to use heroin. [ 28% of those with three or more treatment experiences were not using heroin at the interview in 1975.] Furthermore, the authors concluded that:

… the untreated addict is more likely to stop using heroin and less likely to have high use of heroin than the treated addict.

This in part can be explained by the smaller habits of untreated addicts, but it also is due in part to the fact that untreated addicts had better family lives, more self-esteem and less involvement [with the criminal justice system]. Given the size of the sample, the conclusions of the authors are, perhaps, premature, but the data does suggest that there are similar frequencies of recovery for both untreated and treated groups.

The most recent study to suggest natural recovery comes from a longitudinal health study conducted in Central Harlem (in New York City) by Ann Brunswick at Columbia University School of Public Health. In 1975-1976, during a second wave of interviews, 535 of the original sample of 668 black youth (18-23 years of age) were interviewed; this was 80% of the original sample. The follow-up interviews revealed that 16% of the sample had used heroin at least once and 13% had used it daily or a “few times a week” at some time in their life (usually during the five year interval between the first and second interview).

When respondents were asked about their most recent use of heroin, only 25% of the 69 persons (who used heroin more than twice) reported using heroin in the last year. Organizing the data into those who had received treatment and those who had not it was discovered that 25% of the treated group reported to have used heroin in the previous year while only 16% of the untreated had. (Brunswick, 1978) For purposes of this review, the best incidence data to support natural recovery is the second Robins study, of returned Vietnam veterans. It is best because it utilized urine samples rather than merely relying on self reports, which is the principal source of data used in all the other studies reviewed here except Winick’s which used arrest reports. Robins second study is also special in another way, the users in the sample were predominantly smokers of heroin and opium and were subject to dramatic environmental changes (returned from wartime Vietnam). The study population may be so unique that it would be unwarranted to generalize findings from it to typical addicts in the United States who most commonly use heroin through intravenous injection.

All of the studies reviewed suggest natural recovery but do not by any means settle the issue with detailed and conclusive data. Only three of the studies explored the length of heroin abstinence (the Brunswick study set a one-year criteria, O’Donnell et al had a nine month to one year period, and the Burt Associates study used two months) while most reported heroin use only at the time of the interview. It is possible that a good number of the persons in the other studies were experiencing some short-term or periodic hiatus from heroin use and might be merely in temporary remission. It is well known that addicts use heroin cyclically (Waldorf, 1970; O’Donnell, 1965); seldom is heroin used uninterruptedly. Occasionally individuals initiate an abstinence from use themselves only to eventually drift back to regular use.

Other important concerns with the study of natural recovery are the processes related to the recovery and the characteristics of those who do overcome their heroin addiction. Only two of the above studies dealt with process because most were not designed to explore it. In most instances the researchers did not anticipate findings of this type and thus were not prepared to explore the why and wherefore of the abstinence. Moreover, in many instances the size of the sub-samples of treated and untreated groups were too small to allow even simple statistical correlations to be done to say nothing of more complex kinds of analyses. Only two studies suggested reasons for natural recovery and both had very small numbers of untreated addicts. The Columbia University study which included 36 treated and 32 untreated heroin users found that the treated used heroin for longer periods and the untreated had greater education attainment. The second study conducted by the Graevens had similar small numbers of treated (53) and untreated (23). Their data suggest that the untreated heroin addict had higher self-esteem and better family relations than the treated addicts. Both suggest that personal and family resources contribute to natural recovery.

Conclusion and Recommendations For Future Research

The review of the studies presented here documents the fact that significant numbers of heroin addicts naturally recover from their addiction without treatment intervention. Some of the evidence even suggests that the untreated addict may have as good or possibly even a better chance at breaking the bond of addiction than the treated heroin user. However, there is a virtual absence of substantive information concerning the dynamics and processes-social, psychological and environmental-that may be at work to bring about recovery.

Reasons for the sparcity of knowledge in this area are without doubt numerous and complex but certainly a major one is the widely-held belief of both the lay and professional communities of “once an addict, always an addict”. Such a rigid, unyielding and deterministic viewpoint toward heroin addiction precludes serious consideration of the possibility that some addicts might “naturally” find their own path to recovery. As a result, most research efforts at this time have concentrated on documenting the incidence and spread of addiction, on developing methods to control heroin use and its distribution and on creating prevention and treatment programs.

In short, society has over the past decade responded to heroin addiction in much the same way it did to alcohol use and alcoholism. It has implicitly accepted the phenomena as a intrangent social problem that should be eradicated or at least controlled. The perspective that shaped the current social policy and programatic responses to the heroin problem inadvertently turned research efforts away from developing a more complete substantive understanding of the natural course of the career of heroin addiction and its possible natural discontinuance.

In order to gain a more definitive understanding of the career of heroin addiction and its various possible outcomes, we recommend future research be undertaken to:

1. establish the length of recovery so as to be sure that the non-addictive state is not short term or temporary;

2. explore the possibility of controlled use after addiction-it may be that ex-addicts can resort to a more controlled pattern of drug use rather than strict abstinence as recent findings on recovered addicts (Harding et al., 1978) and recovered alcoholics has shown (Tuckfeld, 1976; Armour et al., 1976);

3. discover the extent that individuals substitute other drugs for heroin-most particularly alcohol and barbiturates as they are much more dangerous than opiates;

4. determine the characteristics and resources of persons who recover naturally and compare them with their treated counterparts and the larger populations, and lastly

5. learn what are the actual processes of recovery for both treated and untreated addicts-what initiated the attempt to recover, how the individuals cope and what kind of interpersonal, familiaI and community support are utilized.

Researchers interested in the area should be sure to consult the literature on alcoholism recovery as there has been much more work done in that field than in the field of illicit drugs. In recent years there has been a growing concern in the field of alcoholism about the incidence of recovery without the help of formal treatment. Data from social surveys of the general population and control groups of treatment evaluations has suggested natural recovery rates of 4170-63%. (Knupfer, 1972; Armour et. al., 1976; Rutledge, 1973; Smart, 1975) Much of this variation may be accounted for in the different ways researchers have defined recovery. Recovery for some has meant total abstention while for others there can be “normal” drinking. In an effort to apply the various criteria and definitions for recovery to data from a San Francisco survey it was found that recovery without treatment could vary from a low of 11% to a high of 71% depending upon how recovery was defined. (Roizen, Cahalan and Shank, 1976)

These findings and those of a recent RAND evaluation of NIAA sponsored treatment centers (Armour et. al., 1976) kindled interest and debate around the issue and NIAA contracted with a research firm to conduct an in-depth study of the processes of untreated alcoholic remission. (Tuckfeld, 1976) This study which is available from the National Technical Information Service, is, at present, the only study that deals with the dynamics of natural recovery. There are, however, several studies of the processes of treated recovery for opiates (Bess, et. al., 1972; Brill, 1972; Bull, 1972; Waldorf, 1973) of which Brill’s book is the most detailed.


*While this paper was at the printers we received from Dr. Robins a copy of her report of a three year follow-up of 617 Vietnam veterans. That report confirms her earlier data as only 12% of all men addicted in Vietnam became readdicted. Of those treated 47% were re-addicted in the second period
and of those not treated. 17%.

1. The research for this paper was supported by a National Institute on Drug Abuse grant #H81 DA01 988-01.

2. The authors of this paper are currently in the process of exploring the areas suggested here in a study supported by a two year NIDA research grant. The study’s goal is to explore and analyze the experiences of two groups of ex-addicts–one that has received some form of treatment and a second that has never undergone any treatment for their addiction. The criteria for inclusion in the study are in general physical addiction for a total of at least one year and following that to not have been re-addicted for at least two years. The research goal is to interview 100 persons in each group.


Armour, David J. et. al 1976 Alcoholism and Treatment Santa Monica: Rand Corporation, Report #1739-NIAA

Bess, Barbara, et. al. 1972 Factors in successful narcotics renunciation. American Journal of Psychiatry. 28 (7).

Brill, Leon 1972 The De-Addiction Process. Springfield, Ill.: Charles Thomas. 

Brunswick, Ann F. 1978 Black Youth and Drug Use Behavior. Mimeographed. New York: Columbia University School of Public Health.

Bull, James 1972 Coming Alive: The Dynamics of Personal Recovery. Mimeographed. Ph.D. Dissertation. Santa Barbara, Ca.: University of California.

Burt Associates. 1977 Drug Treatment in New York City and Washington, D.C.: Follow-up Studies. N.I.D.A. Monograph.

Graeven, David B. and Kathleen A. Graeven Treated and untreated addicts: factors associated with participation in treatment and cessation of heroin use. Mimeographed, n.d.

Harding, Wayne M., Zinberg, Norman E., Stelmack, Shirely M., and Barry, Michael 1979 “Formerly Addicted-Now-Controlled Opiate Users” Mimeographed, 1978 to be published in The International Journal of the Addictions 14(7), 1979.

Knupfer, Genevieve 1972 Ex-problem drinkers. In Life History Research in Psychopathology. Edited by Merrill Kopp, Lee N. Robins and Max Pollack. Minneapolis: University of Minneapolis

Macro Systems, Inc. 1975 Three year follow-up study of clients enrolled in treatment programs in New York City. Phase III Final Report.

O’Donnell, John A. 1968 The relapse rate in narcotic addiction: a critique of follow-up studies. In N.A.C.C. Reprints 2 (1). Originally published in 1965.

O’Donnell, John A., Voss, Harwin L., Clayton, Richard R., Slatin, Gerald T., and Room, Robin G. W. 1976 Young Men and Drugs-A Nationwide Survey. NIDA Research Monograph 5.

Robins, Lee N. and Murphy, George T. 1967 Drug use in a normal population of young negro men. American Journal of Public Health. 57 (9), 1967.

Robins, Lee N. 1973 The Vietnam Drug User Returns. Washington D.C.: U.S. Government Printing Office

Robins, Lee N., Helzer, John E. and Davis, Darlene H. 1975 Narcotic use in Southeast Asia and afterward. Archives of General Psychiatry. 23.

Rutledge, Carolyn, et. al. 1973 A socio-epiderniological study of alcoholism in East Baton Rouge parish. Baton Rouge, Louisiana: Alcohol and Drug Abuse Section, Division of Health.

Scharse, Robert 1966 Cessation patterns among neophyte heroin users. The International Journal of the Addictions. 1

Smart, Reginald 1975- Spontaneous recovery in alcoholics: a review and analysis of available 1976 research. Drug and Alcohol Dependence. 1 (4)

Tuckfeld, Barry S., Simuel, Judy B., Schmitt, Mary L., Ries, Janet L., Kay, Debra L. and Waterhouse, Gloria J. 1976 Changes in patterns of alcohol use without the aid of formal treatment: an exploratory study of former drinkers. Research Triangle Park, North Carolina: Research Triangle Institute. Contract #ADM 281-75-0023.

Vaillant, George E. 1973 A 20-year follow-up of New York narcotic addicts. Archives of General Psychiatry. 29

Waldorf, Dan 1970 Life without heroin: some social adjustments during long-term periods of voluntary abstention. Social Problems 18 (Fall).

Waldorf, Dan 1973 Rock bottom. Chapter 9 in Careers in Dope by Dan Waldorf. Englewood Cliffs: Prentice-Hall, Inc.

Winick, Charles 1962 Maturing out of narcotic addiction. U.S. Bulletin on Narcotics. 14.