My interpretation of overall conclusions: Individuals who complete an AA program are not significantly different from those on other treatments. The studies that attempt to randomize treatments show worse results for AA. The two studies that mention adherence rates show AA has more drop-outs.
Summary of ~12 studies: http://en.wikipedia.org/wiki/Effectiveness_of_Alcoholics_Anonymous
My interpretation of overall conclusions: Individuals who complete an AA program are not significantly different from those on other treatments. The studies that attempt to randomize treatments show worse results for AA. The two studies that mention adherence rates show AA has more drop-outs.
Best meta-study I could find: http://www.ncbi.nlm.nih.gov/pubmed/2648497
Conclusion: AA is not significantly different other treatments.