An introduction to the evaluation of alcoholism outcome studies.

The results of outcome studies in the field of alcoholism have been challenged as not providing clear and definitive demonstrations of the effectiveness of alcoholism treatment. This has not been due to studies showing negative results, or to a lack of studies reporting positive results. The difficulty has been that the studies failed to satisfy the standards of experimentally minded methodologists. Many methodological problems with respect to patient, treatment, and criterion variables were described which will require a great deal of basic research for their solution. Others may not be soluble. Drawing a representative sample of alcoholics may not be feasible. Assigning patients at random to treatments and no treatment who are uninterested, who desire particular treatments, or who are in need of specific treatments is impractical and socially unacceptable. Matching or randomly assigning therapists to provide consistent, measured, and equal amounts of specific therapies is similarly unrealistic. Determining what constitutes a good therapeutic outcome may be more of a philosophic than a scientific issue. Applying experimental standards to human outcome studies is inappropriate. The state of the art with respect to evaluating the effectiveness of alcoholism treatment is not ready for experimental investigations or cost-benefits analyses. Despite all these problems, a review of 384 studies indicates that two-thirds of the patients improve following alcoholism treatment. Research studies are increasingly sophisticated. A great deal has been and is continuing to be learned about patient characteristics, treatment methods, and matching patients to treatments.