Some references:
- https://www.thecut.com/2015/03/why-alcoholics-anonymous-work...
- https://www.wbur.org/commonhealth/2020/03/11/medical-science...
"Mostly it [the new review] looks like studies of either sending people to AA or frequent counseling, and that yields similar results to other kinds of alcohol use disorder psychotherapies. That's what their main finding is," Saitz said. "And I, unfortunately, I don't think that tells us so much about AA effectiveness."
As an aside the average AA member attends 2.5 meetings a week and membership is 89% white[1], I had no idea. Where do non-white people get help in America?
[1] https://www.aa.org/assets/en_US/p-48_membershipsurvey.pdf
Cochrane reviews are considered the gold standard for meta-reviews. If the data, which does show AA [1] helps some alcoholics become abstinent, wasn’t scientifically sound, it wouldn’t had become part of the review.
Saitz either thinks a 7% improvement is a “similar result”, or he wasn’t familiar with the study when interviewed.
Obviously, people are free to think the science doesn’t “conclusively support” a given theory. Using this line of thinking, some people think radiocarbon dating doesn’t “conclusively support” the idea of the earth being older than 6,000 years -- but that’s hardly an evidence-based scientific line of reasoning.
[1] Since it’s not practical to use randomization to compare AA per se to non-AA recovery because of “contaminated control” issues, modern studies compare twelve-step facilitation (TSF) treatment with non-TSF treatment.