Participant selection biases in a randomized clinical trial of alcoholism treatment settings and intensities

Eligible participants and decliners in a randomized study of inpatient, intense outpatient, and standard outpatient treatments for alcoholics were compared and contrasted on a series of demographic, social stability, psychological, legal, drug use, problem severity, and treatment history variables. Among 302 individuals meeting eligibility requirements, those agreeing to participate, compared with decliners, were more likely to be unemployed, be residentially less stable, have legal problems, use other drugs, have a more severe alcohol problem, have a recent treatment history, and were less likely to have problems with violence. Participants also were more likely to be male and non-white, although gender and racial effects were not significant when other variables were controlled for. The implications of these findings for generalizing the results of inpatient-outpatient studies are discussed, and the need for routine reporting of decliner characteristics in research reports is stressed.