12-Step vs Non-12-Step: What the Evidence Shows

01 What the two approaches actually are

Note the last row carefully. It describes how much each has been studied — which is not the same as how well each works.

02 The finding both sides tend to misquote

In 2020 a Cochrane review examined Alcoholics Anonymous and clinically delivered 12-step facilitation across 27 studies and 10,565 participants. It is the most rigorous look this question has had, and it is quoted selectively by almost everyone.

What it found in favour of 12-step: manualized 12-step facilitation improved rates of continuous abstinence at 12 months compared with other clinical treatments such as CBT — risk ratio 1.21 (95% CI 1.03 to 1.42), rated high-certainty evidence, and the effect stayed consistent at 24 and 36 months. The review also concluded it probably produces substantial healthcare cost savings.

What it found that is quoted far less often: on percentage of days abstinent, 12-step facilitation performed about the same as the other treatments — a mean difference of 3.03 with a confidence interval running from −4.36 to 10.43. That interval crosses zero, which means no reliable difference was demonstrated.

Put together, the review says something quite specific: 12-step approaches are better at producing complete, unbroken abstinence. They are not clearly better at reducing drinking overall. If the goal is total abstinence, that is a meaningful advantage. If the goal is drinking less, the evidence does not favour either approach.

One honest caveat about the evidence quality: risk of selection bias was rated high in 11 of the 27 studies, and the strong abstinence comparison rests on 2 studies covering 1,936 people. High-certainty applies to that specific comparison, not to everything in the field.

03 Why "less evidence" does not mean "worse"

The most common mistake in this comparison is treating the size of a research literature as a verdict on the thing it studies.

A systematic review of SMART Recovery set out to do for it what has been done for the 12-step approach. It found 12 studies in total, of which only 3 evaluated effectiveness. Alcohol was the primary focus; assessment of other substance use was infrequent; functional outcomes were rarely reported; no economic analysis had been done at all. The authors' own summary is that a rigorous exploration of SMART Recovery has yet to be undertaken.

That is a statement about research activity, not about whether it helps people. SMART's methods are drawn from cognitive behavioural therapy, which is itself well evidenced. What has not been established is how well the mutual-aid package performs as a whole compared with alternatives.

So the fair summary is: 12-step is better evidenced; non-12-step is less examined. Anyone claiming the research shows one is superior overall is reading more into it than is there.

04 The matching question, which has a surprising answer

The intuitive idea is that different people suit different philosophies — the spiritually inclined to AA, the analytically minded to SMART. A very large trial was built specifically to test that logic.

Project MATCH randomly assigned people with alcohol dependence to one of three manualised treatments — cognitive behavioural therapy, motivational enhancement, or 12-step facilitation — across two parallel trials, following up more than 90% of participants. Every group improved substantially. But there was "little difference in outcomes by type of treatment," and the matching effects the study was designed to find largely did not appear.

One exception held up: outpatients low in psychiatric severity had more abstinent days after 12-step facilitation than after CBT. Beyond that, the authors concluded that providers need not take client characteristics into account when assigning people between the three approaches, despite their very different philosophies.

05 What this means for a family

Three things follow, and none of them is "make them go to AA."

• If the goal is complete abstinence, 12-step has the better track record — and it is free and available almost everywhere, which is not a small consideration.

• If a group is not a fit, switching is reasonable. The alternative is less-studied, not established as worse, and the evidence gives little support for the idea that a particular personality needs a particular philosophy.

• Attendance is the part that is consistently associated with outcomes. A group someone actually returns to beats a better-evidenced group they quietly stop attending.

An important limit on everything above: this research is overwhelmingly about alcohol. The evidence for Narcotics Anonymous and for drug use disorders is considerably thinner, and for opioid use disorder in particular, mutual-aid attendance is not a substitute for medication — see what MAT is and detox vs rehab.

Meetings of both kinds, and treatment programmes generally, can be found independently through FindTreatment.gov.