A physician named Robert Holbrook Smith had his last alcoholic drink 79 years ago this week.   He and Bill Wilson, the friend with whom he had sobered up, thus anointed June 10 as the official founding date of the organization they created together: Alcoholics Anonymous.  AA is the most commonly-sought source of help for problem drinking in the United States and many millions of people have participated in it and the other 12-step organizations it inspired around the world.

But does AA really work?

For most of the 12-step fellowship’s existence, professionals in the addiction field held widely varying opinions of its value.  Some praised AA as an extremely valuable resource for people seeking recovery, but others viewed it as unsophisticated folk medicine and even a bit cultish.  Other tensions emerged from turf issues: Medical professionals can be dismissive of – at times even hostile to – those they consider well-intentioned amateurs.  Just as some obstetricians resent midwives, some addiction treatment professionals looked down on the non-professional AA members in their midst.

To be sure, a large number of studies conducted from the 1950s through the 1980s showed that AA participation was correlated with decreases in problem drinking as well as improvements in mental health and quality of life.   But because these studies were of individuals who had chosen on their own to attend AA, many suspected that the putatively positive findings were actually due to “selection bias.”  In this case, selection bias would be present if only the most motivated and organized problem drinkers attended AA.  This would undermine the case for AA’s effectiveness because such individuals might well have recovered without the organization’s help.

Meanwhile, an increasing number of professionally-designed psychological therapies for alcohol dependence were proving beneficial in randomized clinical trials.   Such studies are considered the gold standard of proof in medicine because researchers rather than patients determine who receives what sort of help, thereby eliminating the problem of selection bias that had bedeviled AA research.

A watershed in scientist’s views of the value of AA occurred in the 1990s with Project MATCH, the largest study of alcohol dependence treatment ever undertaken.   Two well-validated professionally-developed psychotherapies were evaluated head to head against “twelve-step facilitation counselling.”  This counselling approach adapted AA ideas and goals into a 3-month long psychotherapist-delivered outpatient treatment protocol and also strongly encouraged involvement in community-based AA groups.

AA skeptics were confident that by putting AA up against the best professional psychotherapies in a highly rigorous study, Project MATCH would prove beyond doubt that the 12-steps were mumbo jumbo.  The skeptics were humbled: Twelve-step facilitation was as effective as the best psychotherapies professionals had developed.

A subsequent randomized clinical trial eliminated the twelve-step counselling component and simply evaluated the effect of a brief, structured introduction to AA (as well as Narcotics Anonymous, if appropriate).  Those connected by researchers to 12-step groups had substantially lower rates of using alcohol and other drugs over time.  This proved that the groups themselves have a positive impact, even when they are not coupled with extended professionally-provided twelve-step facilitation counselling.

Studies such as these dramatically reduced the ranks of AA critics among scientists.  AA’s value is still questioned in a few quarters, but as Harvard Professor of Psychiatry John Kelly notes, this is becoming less true as the years go by: “The stronger scientific evidence supporting the effectiveness of AA is relatively new. It takes time for evidence to disseminate into clinical practice as well as into broader society.”

Keith Humphreys is a Professor of Psychiatry at Stanford University.  Follow him on Twitter @KeithNHumphreys