Alcoholics Anonymous® 12-Step Programs
The featured article, written by a freelance journalist, was highly critical of faith-based Alcoholics Anonymous (AA) 12-step programs and questioned why evidence-based medications are used so infrequently for treating alcohol use disorders (Glaser, 2015). The author and individuals interviewed did not believe that 12-step programs are effective or should play a central role in the treatment of alcohol use disorders.
Glaser (2015) stated:
A meticulous analysis of treatments, published more than a decade ago in The Handbook of Alcoholism Treatment Approaches but still considered one of the most comprehensive comparisons, ranks AA 38th out of 48 methods. At the top of the list are brief interventions by a medical professional; motivational enhancement, a form of counseling that aims to help people see the need to change; and acamprosate, a drug that eases cravings. (An oft-cited 1996 study found 12-step facilitation—a form of individual therapy that aims to get the patient to attend AA meetings—as effective as cognitive behavioral therapy and motivational interviewing. But that study, called Project Match, was widely criticized for scientific failings, including the lack of a control group.) (p. 54)
A scientific debate about Project Match can be read in dueling commentaries (Cutler & Fishbain, 2005; Miller, 2005). The Handbook that Glaser (2015) refers to is by Hester and Miller (2002), where the “meticulous analysis of treatments” (p. 54) is contained in Chapter 2 (“What works? A summary of alcohol treatment outcome research”). Miller (2005) was the lead author of this chapter, a principal investigator on Project Match, and a co-investigator on the COMBINE study (Anton et al., 2006). Miller (2005) may have appreciated Glaser’s (2015) glowing praise of his chapter, but not necessarily her scientific take on Project Match.
While emphasizing the “meticulous analysis of treatments” from the Handbook and devaluing findings from Project Match, Glaser (2015) neglected to mention findings from the federally funded COMBINE study (Anton et al., 2006), which was randomized and blinded. In this large study of 1,383 recently alcohol-abstinent patients, eight groups received medical management with 16 weeks of naltrexone (Revia®), acamprosate (Campral®), or both, and/or both placebos, with or without a combined behavioral intervention (CBI). A ninth group received CBI only (no pills). The CBI integrated aspects of cognitive-behavioral therapy, 12-step facilitation, motivational interviewing, and support system involvement external to the study. A motivational interviewing style was used throughout. Patients receiving medical management with naltrexone, CBI, or both fared better on drinking outcomes, whereas acamprosate showed no evidence of efficacy, with or without CBI. No combination produced better efficacy than naltrexone or CBI alone in the presence of medical management. Placebo pills and meeting with a health care professional had a positive effect greater than that of CBI during treatment.
Glaser (2015) also stated (while referencing Ferri, Amato, and Davoli, 2006):
In 2006, the Cochrane Collaboration, a health-care research group, reviewed studies going back to the 1960s and found that “no experimental studies unequivocally demonstrated the effectiveness of AA or [12-step] approaches for reducing alcohol dependence or problems.” (p. 56)
Rather than relying on Glaser’s selective quote, it is instructive to read carefully the eight studies included by Ferri et al. (2006) in their analysis, as well as their discussion and conclusions. Control interventions in these studies included no treatment, a variety of psychological interventions, or 12-step program variants. In my opinion, the conclusions to be drawn from their analyses are more nuanced than definitive. Ferri et al.’s (2006) plain language summary included:
As well as AA, there are also alternative interventions based on 12-step type programmes, some self-help and some professionally led. AA and other 12-step approaches are typically based on the assumption that substance dependence is a spiritual and a medical disease. The available experimental studies did not demonstrate the effectiveness of AA or other 12-step approaches in reducing alcohol use and achieving abstinence compared with other treatments, but there were some limitations with these studies. Furthermore, many different interventions were often compared in the same study and too many hypotheses were tested at the same time to identify factors which determine treatment success. (p. 2)
Ferri et al. (2006) further stated, “In general, the available research seems to be concentrated on prognostic factors associated with assumedly successful treatments rather than on the effectiveness of treatments in themselves” (p. 8). The italics emphasize what I believe is a key point of these studies.
My interpretation of this study analysis is not that 12-step programs are ineffective, but that they have neither been adequately studied nor shown to be more effective than comparison interventions. It is not clear whether these studies were designed as noninferiority studies (i.e., demonstrating relative equivalence among therapies) or with adequate power to demonstrate superiority. Indeed, Ferri et al. (2006) added, “The main methodological problem with the included studies is the statistical power which was never mentioned and possibly not taken into account when designing the studies” (p. 6).
This methodological issue often is encountered when comparing different medication therapies or psychotherapies for other disorders, such as depression. Drugs are first approved based on safety and efficacy established in placebo-controlled trials. Subsequent studies comparing treatments previously shown to be effective in placebo-controlled studies typically are conducted using noninferiority designs (Greene, Morland, Durkalski, & Frueh, 2008). Based on noninferiority study designs, it can be concluded that one treatment is not worse than another; it cannot be proven that one is superior.
Establishing the efficacy of psychosocial therapies is much more difficult than for drug therapies. Creating psychosocial control groups and blinding them is technically and conceptually challenging. By contrast, pharmacological placebos are easier to design and blind. According to Ferri et al. (2006), blinding was never mentioned in any report or publication of the studies analyzed.
The problem with comparing the efficacy of psychosocial therapies for alcohol use disorder is likely to be similar to that seen when they are investigated for the treatment of depression. In a recent review of psychotherapies for depression, Cuijpers (2015) explained:
Since the first trials were conducted in the 1970s, every few years ‘new’ psychotherapies are discovered. These new therapies claim that they are more effective than the ‘old’ or ‘traditional’ therapies. This has resulted in dozens of different types of psychotherapies. In the field of depression, however, there are no indications whatsoever that such new therapies are more effective than longer existing therapies. Meta-analyses of studies directly comparing different psychotherapies consistently show that there are no or only small differences between therapies. If a new therapy claims to be more effective than existing therapies, it should be assumed that the difference between the new and the existing therapies is small. In order to find this, large sample sizes are needed. For example, the largest difference between two therapies for depression has been found to be small (effect size d = 0.20). If a trial is designed to show that a new therapy is more effective than existing ones, and the expected differential effect size is d = 0.20, this trial needs to have 491 patients in each condition…. No such trial has ever been conducted…. From this perspective, it does not seem very useful to develop new psychotherapies for depression. There are already many different kinds of psychotherapies, there is no evidence that one type is considerably more effective than others, and it is very difficult to show such differences in trials with sufficient statistical power. (pp. 25-26)
If this is the situation with depression, then a characterization of 12-step programs as not being evidence-based may be specious when the real problem simply is a paucity of high-quality, methodologically sound studies. I do not necessarily disagree with critiques of 12-step programs, but the truth about their effectiveness probably lies somewhere between detractors and proponents. In addition to criticizing 12-step programs, Glaser (2015) admitted that there is no incentive to rigorously study their effectiveness.
Medication for the Treatment of Alcohol Use Disorder
The brief guide prepared for SAMHSA (2015) was based on the findings of a scientific consensus panel that reviewed current evidence of the effectiveness of available medications for the treatment of alcohol use disorders. However, in the guide, the only medications described are those that have been approved by the U.S. Food and Drug Administration (FDA) for this indication: acamprosate, disulfiram (Antabuse®), oral naltrexone (Revia®), and extended-release injectable naltrexone (Vivitrol®). It is disappointing that off-label use of other FDA-approved medications that have been investigated is not covered, but the guide provides readable and useful clinical information. Glaser (2015) mentioned these four medications as well as topiramate (Topamax®), baclofen, varenicline (Chantix®), and nalmefene (Revex®), which all have evidence for efficacy based on one or more randomized placebo-controlled studies (Howland, 2012; Keating, 2013; Litten et al., 2013; Martinotti et al., 2014). Given her penchant for promoting the use of evidence-based medications, it is surprising that Glaser (2015) failed to mention gabapentin (Neurontin®), whose efficacy for alcohol use disorders has been demonstrated in at least four placebo-controlled trials (Howland, 2013).
A problem only mentioned in passing by Glaser (2015) is that efficacious evidence-based medications are not always effective when used in clinical practice, which is also true of psychotherapies. Availability of services and cost are barriers for many patients. Treatment adherence with medication or therapy is variable and likely to be worse for substance use disorders than other medical or psychiatric conditions. Medication efficacy typically is established based on relatively short-term studies, but information about long-term adherence, efficacy, and safety is often lacking. Another problem not mentioned by Glaser (2015) is that medications have side effects and toxicities (some more than others).
According to Glaser (2015) and SAMHSA (2015), evidence-based medications are underused for the treatment of alcohol use disorders. Why might this be true? Glaser (2015, p. 53) noted, “Nowhere in the field of medicine is treatment less grounded in modern science,” criticizing the culturally ingrained dominance of faith-based, 12-step thinking that she said draws nothing from modern science. She also decried the profit motives of the rehabilitation industry. I would agree that substance abuse should be within the realm of science and medicine. Although, as previously mentioned, 12-step programs simply have not and probably cannot be investigated in the same way as medication. Substance abuse and 12-step programs are also not the only examples in the United States where faith competes with, or even trumps, science.
Whereas 12-step programs may be criticized for their lack of scientific evidence, pharmaceutical companies often are vilified for overselling benefits and underemphasizing safety of drug therapies. This may be a particular issue for substance abuse treatments. The prescription opioid drug problem illustrates how a well-intentioned idea (e.g., more treatment options for better pain control) can run amok (Paulozzi, 2012). The medical use of a buprenorphine/naloxone combination (Suboxone®) for treating opioid dependence is an example of how treatment itself can be abused (and sold on the street). I discussed this type of concern on the use and abuse potential of gabapentin (Howland, 2013, 2014a,b). In addition, profit motives are not unique to the rehabilitation industry. Pharmaceutical companies are readily criticized for profiteering, but are only one member of the much larger American medical-industrial complex (Relman, 1980).