How should the effectiveness of treatment for alcohol problems be evaluated? Available to Purchase

In times of information overload, researchers have found ways to synthesise a large amount of data from numerous studies bearing on the effectiveness of treatment for alcohol problems. Two of these ways are discussed: formal meta‐analysis and the box‐score ‘mesa grande’. It is concluded that meta‐analysis cannot answer questions regarding which treatments give the best results in the alcohol field. The mesa grande has certain limitations, which are described, but is useful for the formation of treatment policies when a clear summary of the research evidence on treatment effectiveness is needed. By contrast, large multi‐centre randomised controlled trials with enough statistical power to detect small effects of treatment should normally be preferred when a decision has to be made as to which of two or more specified treatments should be implemented in practice. Unfortunately, two multi‐centre trials, one in the USA and one in the UK, have given rise to the ‘dodo bird’ verdict of equivalent effectiveness of four treatment modalities. The findings of the UK Alcohol Treatment Trial cannot be used to advise treatment providers and practitioners which one of two treatments, MET or SBNT, should be preferred in practice. In the absence of relevant research findings, four possible ways of making this decision are outlined, including the suggestion that MET should serve as the first step in a stepped‐care model of treatment provision.