Alcoholics Anonymous, the 80-year-old self-help program, has always had critics, who fault it for being too religious and unscientific. Journalist Gabrielle Glaser revives both these charges in her April Atlantic article, "The False Gospel of Alcoholics Anonymous." She claims that "researchers have debunked central tenets of A.A. doctrine and found dozens of other treatments more effective."

The addiction-treatment industry is a racket, which cries out for critical investigation. But Glaser's article is embarrassingly shallow and one-sided. She cherry-picks data and anecdotes to make A.A. look bad and alternatives look good. Her article has already provoked some blowback, including this critique in New York Magazine, but I'd like to add my own complaints.

Here's an example of how Glaser misrepresents sources: She quotes a 2006 report by the Cochrane Collaboration on A.A. and other twelve-step programs (so-called because they are based on A.A.'s recommendations for maintaining sobriety). Cochrane Collaboration is a terrific source of independent analyses of health-related issues, but for the most part its work does not support Glaser's thesis.

She quotes Cochrane's conclusion that "no experimental studies unequivocally demonstrated the effectiveness of AA or [12-step] approaches for reducing alcohol dependence or problems." She neglects to mention that the 2006 report also examined studies comparing twelve-step programs to other treatment methods. The result? "Severity of addiction and drinking consequence did not seem to be differentially influenced by [twelve-step programs] versus comparison treatment interventions," Cochrane states, "and no conclusive differences in treatment drop out rates were reported."

Glaser faults the zero-tolerance tenet of A.A. and touts programs that seek reduction rather than elimination of drinking. But a 2012 Cochrane evaluation found no rigorous studies of so-called "managed alcohol programs." The report states: "The lack of evidence does not allow for a conclusion regarding the efficacy of [managed alcohol programs] on their own, or as compared to brief intervention, moderate drinking, no intervention or 12-step variants."

Glaser is also keen on pharmaceutical treatments, particularly those involving naltrexone, which blocks opioid receptors and is more commonly used to counter opioid addiction. Glaser reports that after taking naltrexone herself for ten days, she "no longer looked forward to a glass of wine with dinner," and she lost two pounds to boot.

Glaser cites other personal testimonials and studies that supposedly demonstrate naltrexone's effectiveness. Actually, the evidence is, at best, mixed. A 2010 Cochrane analysis of 50 studies involving 7,793 subjects concluded that "more patients who took naltrexone were able to reduce the amount and frequency of drinking than those who took an identical appearing, but inert substance." The effect was hardly overwhelming. "On average," the Cochrane report noted, "one out of nine patients was helped by naltrexone." Even that modest effect was not supported by a double-blind study reported in the New England Journal of Medicine in 2001. Those authors found no difference between naltrexone and placebos for treating "chronic, severe alcohol dependence."

I know many people who have tried to overcome addictions with Alcoholics Anonymous and other twelve-step programs, meditation, psychotherapies, medications and various combinations of the above. Some people have mastered their compulsions, others haven't. I wish I could recommend a particular treatment, but I suspect that all are subject to the notorious "dodo bird verdict." Psychologist Saul Rosenzweig coined the phrase in the 1930s to describe the possibility that all psychotherapies are roughly as effective, or ineffective, as each other.

The phrase refers to an episode in Lewis Carroll’s fable Alice’s Adventures in Wonderland in which a dodo persuades Alice and other characters to race around an island. The dodo eventually proclaims, "Everyone has won, and all must have prizes!" The dodo bird verdict has been repeatedly borne out by modern research, including this 2013 comparison of cognitive-behavioral therapy to other psychotherapies. Divergences from the dodo verdict often display the allegiance effect, researchers' tendency to find evidence for the therapy they favor.

The dodo verdict applies to most Cochrane studies of treatments for addiction, including this 2014 analysis, which offers a weirdly double-negative conclusion: "There is low-quality evidence to suggest that there is no difference in effectiveness between different types of interventions to reduce alcohol consumption in concurrent problem alcohol and illicit drug users."

Evidence that treatments are equivalent in effectiveness flies in the face of Glaser's claim that "researchers have debunked central tenets of A.A. doctrine and found dozens of other treatments more effective."

And that brings me to my final complaint. Glaser unfairly conflates A.A. with rehabilitation centers, such as Hazelden, that employ twelve-step programs and charge patients for their services. Glaser estimates that Americans spend $35 billion annually on alcohol and substance-abuse treatments, including twelve-step programs. She is justly outraged that "some bare-bones facilities charge as much as $40,000" a month for treatment. She expresses the hope that the Affordable Care Act, which expands coverage of substance-abuse treatments, will lead to better evaluations of and hence advances in such treatments. I share this hope.

But as psychologists Hal Arkowitz and Scott Lilienfeld noted in Scientific American in 2011, given "the wide availability of meetings and the lack of expense, A.A. is worth considering for many problem drinkers." [Italics added.] If all treatments are as effective as each other, cost should be the determining factor. Judged by that criterion, no treatment beats A.A., because A.A. is free.

*The above headline replaced and is vastly superior to the original: "Alcoholics Anonymous Ain't Perfect, But At Least It's Free."

A.A. logo: