The trials of Alcoholics Anonymous
Version of Record online: 18 APR 2006
Volume 101, Issue 5, pages 617–618, May 2006
How to Cite
HUMPHREYS, K. (2006), The trials of Alcoholics Anonymous. Addiction, 101: 617–618. doi: 10.1111/j.1360-0443.2006.01447.x
- Issue online: 18 APR 2006
- Version of Record online: 18 APR 2006
Timko et al. (2006) have advanced the field in two critical respects. Substantively, their randomized trial showed that intensive referral to 12-Step self-help groups leads to higher group involvement and better substance use outcomes, a particularly striking result given the already high rate of 12-Step meeting attendance in the control condition and the severe, long-standing addictions of the study participants. Methodologically, the study is in the vanguard of a new generation of experimental evaluations of Alcoholics Anonymous (AA) and other 12-Step self-help groups.
Two decades ago, countless members of the field expected the meeting of experimental methods and AA to resemble a traffic accident. Only one intrepid team attempted a controlled study of AA in the 1970s and 1980s. Their reward was an array of methodological problems and attendant ambiguous results that never made it into the peer-reviewed literature (Brandsma et al. 1980). Some researchers—and I must admit to having been one of them (Humphreys & Rappaport 1994)—were extremely skeptical that the level of professional control and standardization required for a randomized trial could be fully compatible with an ecologically valid evaluation of informal, voluntary, peer-led self-help groups. Others expected (some with fear, some with glee) internecine war within the addiction field if randomized trials suggested that AA's millions of members and admirers were deluded in their belief that the organization was effective. Before the results of Project MATCH appeared, for example, a colleague commented to me without irony ‘I don’t see why people are complaining that there isn’t an untreated control condition—don’t they know that a third of the participants are being sent to AA?’.
Walsh et al. (1991) then bulldozed conventional wisdom by showing not only that AA could be studied in a randomized trial, but that such a project could be methodologically sophisticated enough to grace the pages of the New England Journal of Medicine. Their finding that AA, in combination with in-patient treatment, produced better substance abuse outcomes than did advice to attend AA alone has become less relevant as in-patient alcoholism treatment has virtually disappeared around the world (Bao et al. 2001), but the study none the less laid the groundwork for the efforts that followed later in the 1990s. Subsequent US National Institutes of Health-supported comparisons of ambulatory interventions for alcohol (Project MATCH Research Group 1997), cocaine (Weiss et al. 2000) and nicotine (Martin et al. 1997) patients, as well as for significant others of addicted individuals (e.g. Miller, Meyer & Tonigan 1999), all had a treatment condition embracing 12-Step principles and facilitating self-help group involvement. Together these studies constitute the first wave of high-quality trials of AA and other 12-Step groups.
This first generation of trials produced made major methodological advances, including improving the measurement and conceptualization of self-help group involvement, and developing readily usable, manualized, group involvement facilitation strategies. Confounding the skeptics of self-help groups, these trials also showed that individuals randomized to 12-Step conditions improved significantly over time on a range of measures. Yet these remarkable studies did not test AA's effectiveness directly because they did not hold professional treatment services constant while randomizing their samples to different levels of self-help group involvement. Walsh's design, for example, included an in-patient + AA and an AA-only condition, which allows an inference about the absolute effectiveness of in-patient treatment more readily than it does a parallel inference about AA. In Project MATCH, patients randomized to 12-Step facilitation counseling indeed went to many more self-help group meetings than did patients randomized to motivational enhancement therapy, but they also received three times as many psychotherapy sessions. The designs of the first wave of ‘AA trials’, in short, used randomization primarily to draw valid inferences about professionally provided treatments.
We are now seeing the initial fruits of a second generation of trials in which randomization is intended to draw inferences about AA and its sister fellowships per se. These studies were made possible in no small part by the predecessors’ demonstration that self-help group involvement is influenced powerfully by clinician behavior and is therefore partially subject to randomization.
McCrady, Epstein & Hirsch (1999), for example, conducted a trial where two patient groups received alcohol behavioral couples therapy, only one of which also received an intensive referral to AA and Al-Anon. The 6-month follow-up results were most intriguing: randomization to 12-Step self-help groups clearly enhanced the likelihood of abstinence (41.7% versus 31.8% in alcohol behavioral couples therapy only) yet just as clearly hindered achievement of non-problem drinking (4.2% versus 18.2% in alcohol behavioral couples therapy). Debating the meaning of this result would be an enjoyable exercise that would take more space than is available here; what better proof that the second wave of AA trials is producing stimulating findings?
Like Kingree & Thompson's (2000) trial of the additive benefits of randomization to Al-Anon/Adult Children of Alcoholics groups during residential substance abuse treatment, Timko et al.'s (2006) trial yields a consistently positive conclusion about the benefits of self-help group involvement. Perhaps the current generation of trials will ultimately show that facilitating 12-Step group involvement is less consistently beneficial after a cognitive–behavioral intervention (e.g. alcohol behavioral couples therapy) than it is after a 12-Step-based treatment (Humphreys et al. 1999), implying a need for clinical methods of referring patients to cognitive–behaviorally oriented self-help groups.
The second wave of trials has already taught us a good deal, and there is more to come. Jason et al. (in press) will soon release the results of the first-ever randomized comparison of a residential self-help organization (Oxford House) to usual aftercare for drug and alcohol in-patients. Also under way is the first experimental evaluation of group-provided 12-Step facilitation (Kaskutas & Oberste 2003) and a randomized trial of three strategies for facilitating AA attendance during out-patient treatment (Walitzer 2003). Strong views about AA one way or the other will always survive, no matter what evidence accumulates, but the studies of the past 15 years have established beyond any reasonable doubt that high-quality AA trials are possible, and that such studies usually reinforce rather than undermine the excellent reputation the fellowship enjoys around the world.
Preparation of this article, which is a statement of personal opinion, was supported by a Research Career Scientist Award from the US Department of Veterans Affairs.
Declaration of interest
I served as an unpaid consultant to the Jason et al., Kaskutas & Oberste and Timko et al. studies cited above, which involved critiquing drafts of the original grant proposals but did not involve any aspect of data collection, analysis or interpretation.
- 2001) Changes in alcohol-related inpatient care—an international trend comparison. Journal of Addictive Diseases, 20, 97–104. , & (
- 1980) Outpatient Treatment of Alcoholism: a Review and Comparative Study. Baltimore, MD: University Park Press. , & (
- 1999) Alcoholism: Clinical and Experimental Research, 23, 558–563. , , & (
- 1994) Researching self-help/mutual aid groups and organizations: many roads, one journey. Applied and Preventive Psychology, 3, 217–231. & (
- Communal housing settings enhance substance abuse recovery. American Journal of Public Health. , , & (in press)
- 2003) Making Alcoholics Anonymous Easier: MAAEZ. Berkeley, CA: Alcohol Research Group. & (
- 2000) Mutual help groups, perceived status benefits, and well-being: a test with adult children of alcoholics with personal substance abuse problems. American Journal of Community Psychology, 28, 325–342. & (
- 1997) Prospective evaluation of three smoking interventions in 205 recovering alcoholics: one-year results of Project SCRAP-Tobacco. Journal of Consulting and Clinical Psychology, 65, 190–194. , , , , , et al. (
- 1999) Maintaining change after conjoint behavioral alcohol treatment for men: outcomes at 6 months. Addiction, 94, 1381–1396. , & (
- 1999) Engaging the unmotivated in treatment for alcohol problems: a comparison of three strategies for intervention through family members, Journal of Consulting and Clinical Psychology, 67, 688–697. , & (
- Project MATCH Research Group (1997) Matching alcoholism treatments to client heterogeneity: Project MATCH posttreatment drinking outcomes. Journal of Studies on Alcohol, 58, 7–29.
- 2006) Intensive referral to 12-Step self-help groups and six-month substance use disorder outcomes. Addiction, 101, 678–688. , & (
- 2003) Facilitating involvement in AA outpatient treatment. Available at: http://crisp.cit.nih.gov[accessed 20 December 2005]. (
- 1991) A randomized trial of treatment options for alcohol-abusing workers. New England Journal of Medicine, 325, 775–782. , , , , , et al. (
- 2000) Self-help group attendance and participation among cocaine dependent patients. Drug and Alcohol Dependence, 60, 169–177. , , , , , et al. (