The comparability of research participants with "real-world" patients influences the generalizability of treatment outcome studies. Here this issue is examined in the context of alcoholism treatment research.

The characteristics of adult participants in alcoholism treatment research were determined from a recently published meta-analysis of all 404 multiple-group outcome studies (357 published and 47 unpublished) conducted from 1970 to 1998 (1). The characteristics of 701,962 U.S. adults who were seeking treatment for a primary alcohol problem were determined from the 1999 Treatment Episode Data Set (TEDS) (www.samhsa.gov/oas/samhda.htm).

As can be seen from F1, the differences in demographic characteristics between research participants and typical patients were sizable. Research participants had a much higher level of social stability, as indicated by being employed and being married, which raises concerns that alcoholism treatment outcome research overestimates the effectiveness of treatment among typical patients. Racial and gender differences were less pronounced but are still worrisome given scientists' widely shared commitment to an equitable distribution of the burden and benefits of human research.

A number of noncompeting explanations could be offered for these differences between research participants and typical patients. Studies' eligibility criteria may disproportionately exclude patients who are from racial minorities, who are unemployed, and who are socially unstable (1). Furthermore, persons with certain characteristics—for example, those who are college educated—may be more likely to agree to participate in scientific research. The differences may also be partly due to the fact that research is sometimes conducted in treatment settings, such as some private hospitals, that do not provide patient data to the TEDS.

Understanding the mechanisms through which these differences arise and minimizing them when possible should increase the ability of alcohol treatment outcome research to inform clinical practice.

This work was funded by the National Institute on Alcohol Abuse and Alcoholism and the Veterans Affairs Mental Health Strategic Healthcare Group.

The author is affiliated with the Veterans Affairs and Stanford University Medical Centers, 795 Willow Road (152-MPD), Menlo Park, California 94025 (e-mail, knh@stanford.edu). Harold Alan Pincus, M.D., and Terri L. Tanielian, M.A., are editors of this column.