Send email for updates
Imel Z.E., Wampold B.E.
Psychology of Addictive Behaviors: 2008, 22(4), p. 533–543.
Unable to obtain a copy by clicking title? Try asking the author for a reprint by adapting this prepared e-mail or by writing to Dr Imel at email@example.com.
After combining results from studies comparing talking therapies for alcohol problems, this ingenious analysis finds any structured approach grounded in an explicit model as good as any other. We have, it's argued, been looking in the wrong direction for therapy's active ingredients.
Summary To estimate the relative efficacy of alcohol use disorder treatments, the authors used meta-analysis A recognised set of procedures used to summarise quantitative results from several studies of the same or similar interventions. Usually undertaken to allow an intervention's effectiveness to be assessed with greater confidence than on the basis of the studies taken individually. In this case used to assess differences in the effectiveness of different interventions. techniques to aggregate the outcomes of studies which directly compared two bona fide psychological treatments. They faced the problem of how combine a study where for example treatment A was better than B, with another where A was worse than C. To overcome this they randomly assigned a positive or negative sign to an effect size A standard way of expressing the magnitude of a difference (eg, between outcomes in control and experimental groups) applicable to any quantitative data. In this context, a standardised measure of the extent of the difference in drinking outcomes between the treatments compared in each of the studies included in the meta-analysis. expressing the magnitude of the difference between any two treatments. Then they estimated how far the distribution of effect sizes conformed to the shape expected if in reality there were no differences. A highly variable distribution would indicate that there really were differences in the effectiveness of the treatments which were not due simply to sampling error. A smoother, more homogenous distribution would suggest that real differences in effectiveness were at best minor, and that studies which found differences did so largely due to chance variations. For alcohol measures as a whole, and for measures of abstinence in particular, the analysis did indeed find that effects were homogenously distributed, indicating that different treatment comparisons yielded a common effect size which was not significantly different from zero. Further analyses indicated that the researcher's allegiance to the treatment accounted for a significant portion of what variability there was in differences between treatments. The authors argue that the results are consistent with an emphasis on therapeutic processes common across different treatments (such as the relationship between therapist and client) and common mechanisms of change, rather than specific techniques supposedly stimulating specific mechanisms keyed to a specific complaint.
commentary The analysis brings alcohol therapy squarely within the ambit of a fundamental debate across psychotherapy – whether the drive to devise more effective therapeutic programmes is fundamentally misguided because it is not the specific programme which matters, but 'common factors' which cut across these programmes, such as entering a therapeutic setting within which the patient expects to be helped to get better, the credibility of the therapy to both patient and therapist, its ability to (for that patient) make ordered sense of the patient's 'disorder', in doing so to structure a route out of that disorder which generates optimism, its ability to provide a platform for engaging the client in their recovery, and the therapist's ability to create a supportive environment which facilitates these processes. Perhaps the greatest common factor lies in the patients and clients. Typically they have reached the point where they desperately want to get better, have realised they need help to do so, and decided to follow a culturally sanctioned route to gaining that help – formal treatment.
In his influential book, one of the authors of the current analysis exhaustively analysed these issues in respect of psychotherapy in general, concluding that the evidence overwhelmingly supported a common factor model of how therapy works rather than the medical model of a specific treatment for a distinct complaint. The current analysis goes part way to extending that verdict to alcohol therapy in particular. By including only comparisons of 'bona fide' therapies among (presumably) treatment-seeking samples, the analysis effectively evened out many of the presumed common factors. If these truly were the common core responsible for the effectiveness of seemingly distinct therapies, then comparisons between the outcomes of these therapies should find differences no greater and no more often than would be expected by chance. Differences there would be, but they would be smoothly clustered around the zero difference mark, just as random variation in wind and bounce would leave identical apples falling from the same tree smoothly clustered around a point immediately below. On the other hand, if the differences between the therapies were important active ingredients, outcome differences too would be substantial. Rather than clustering around the zero mark, they would often be found in more extreme territory. The results of the analysis were far closer to what would be expected if the common factor model held. They became closer still when one further candidate common factor was accounted for – how committed the researchers (and perhaps also therefore the therapists and through them the patients) were to one of the therapies they were testing compared to the other(s).
The conclusions are similar to those reached by some of the researchers responsible for two of the largest ever alcohol treatment trials, the US Project MATCH study and the British UKATT trial. After pitting deliberately distinct psychosocial therapies against each other, both studies concluded that the outcomes differed little overall, and that there were few indications that certain types of patient benefited more from one therapy than another. Faced with these findings, MATCH researchers argued for a common factor model, and drew the important implication that "If most treatments are similar in their effectiveness, the real value of having an array of treatments available is to promote healthy competition for the wide variety of people who would benefit from any treatment, but who would be more attracted to one because of reputation, convenience, or personal preference". A leading UKATT researcher has argued for attention to be redirected from 'brand name' therapies to "change processes that are common to treatments with different names and theoretical rationales".
Intriguing as it is, the featured analysis restricted itself to impacts on outcomes directly related to drinking as assessed immediately after therapy ended. Sometimes distinctions between the effectiveness of therapies have emerged only in the longer term Persisting effects are particularly noticeable in respect of cognitive-behavioural therapies. On the other hand, it is not unusual for outcomes to converge over the longer term, as happened in the study which registered the greatest outcome difference in the featured analysis (Miller W.R. Behavioral treatment of problem drinkers: a comparative outcome study of three controlled drinking therapies. Journal of Consulting and Clinical Psychology: 1978, 46, p. 74–86). Had the later outcomes from this study been used instead, the analysis would have even more overwhelmingly come down on the side of there being no real differences between therapies. and sometimes a therapy's main advantage over the alternatives lies beyond drinking itself. Such as couples or family based therapies which particularly help improve life for partners and children. See for example http://findings.org.uk/PHP/dl.php?file=BCT.nug. It is also important to remember that the analysis included only recognised or explicitly structured and theoretically based approaches. Its findings do not mean that an irrational or obviously irrelevant approach would also do as well as any other. Nor can it exclude the possibility that some therapies really are preferable to others; it simply established that generally this has not been the case.
Drug and Alcohol Findings has published a series of articles dedicated to exploring of the impact of some of the common factors which might be important in the treatment of substance use problems.
Thanks for their comments on this entry in draft to Jim Orford of the University of Birmingham. Commentators bear no responsibility for the text including the interpretations and any remaining errors.
Last revised 04 January 2009
Comment/query to editor
Give us your feedback on the site (one-minute survey)
Open Effectiveness Bank home page
Add your name to the mailing list to be alerted to new studies and other site updates