More interesting is this:
Although the variability of effects about zero was small, we also found evidence of an allegiance effect. Specifically, our analyses indicated that as allegiance to compared treatments became more unbalanced, the expected difference between treatments increased, in favor of the treatment forwhich there was researcher allegiance. Allegiance accounted for most of the variability in treatment differences in alcohol measures.
Not shocking to anyone who reads journals often. Researcher’s findings strangely almost always support their hypothesis. It’s worth noting that this study is no exception.
The authors suggest that the most important predictor of outcomes is the alliance between the helper and the client. I’m inclined to accept this, particularly with less severe substance use disorders and short term outcomes. They call for “research that focuses on developing models of the complex interactions that occur between patients”, and I’d agree.
What I believe is missing from their formulation is the importance of factors like intensity and duration. I suspect that the most successful approaches will be those that respond to addiction as a chronic illness and recruit patients into managing their recovery over their life course.