Length of stay and program component effects in the treatment of alcoholism: A comparison of two techniques for process analyses.

The effects of length of treatment and specific treatment components (therapy sessions, Alcoholics Anonymous meetings, and films and lectures on alcoholism) of three residential alcoholism programs were examined. Two statistical techniques, partial correlation and treatment-effect correlation, were compared for their estimates of treatment effects after controlling for patient background characteristics and functioning at intake. Longer periods of treatment were associated with better outcome for residents of a halfway house but not for patients at a milieu-oriented program or a Salvation Army center. Some evidence suggested that the three program components tended to have moderately beneficial effects on outcome, although the results varied in some cases, depending on whether partial or treatment-effect correlation was used in the analysis. The substantive and methodological implications of the findings are considered. Evaluations of inpatient alcoholism treatment programs have proliferated in recent years. As in most areas of evaluative research, however, the focus has usually been summative (Scriven, 1967), or on the overall impact of treatment, whereas the process of treatment, the effects of the therapeutic components that comprise programs (Suchman, 1967), has virtually been ignored. Consequently, ineffective or even counterproductive program components may be retained or expanded, and effective elements may not be assigned the resources they merit. A related issue that has received more attention in the empirical literature concerns the association between length of stay (LOS) in inpatient alcoholism programs and treatment outcome. Patients, practitioners, and third-party payment sources want to know whether length of hospital