APHA
Back to Annual Meeting Page
 
American Public Health Association
133rd Annual Meeting & Exposition
December 10-14, 2005
Philadelphia, PA
APHA 2005
 
3321.0: Monday, December 12, 2005 - Board 9

Abstract #117637

Role of selection effects in evaluating substance abuse treatment outcomes in three California counties

Jason Bond, PhD1, Martha Beattie, PhD1, and Thomas K. Greenfield, PhD2. (1) Alcohol Research Group, 2000 Hearst Avenue, Berkeley, CA 94709, 510-642-7965, jbond@arg.org, (2) Public Health Institute, Alcohol Research Group, 2000 Hearst Ave, Berkeley, CA 94709

A study of public sector substance abuse treatment systems compared MidState, a California county that reorganized its treatment system using managed care principles, to two other California counties taking different approaches, NorthState and SouthState. It was hypothesized that MidState would have better outcomes due to its emphasis on quality of care. Adult clients (N=714) were interviewed entering treatment from 6/2000 to 6/2001 and again twelve months later. Interviews assessed functioning in the seven Addiction Severity Index (ASI) domains – alcohol, drug, psychiatric, legal, employment, medical, and family/social. The 714 clients interviewed represented 70% of the n=1020 initially recruited with actual recruitment rates varying by site (85%, 87%, 66% in SouthState, NorthState, and MidState). These participation rates raise sample self-selection issues and indicate the need for a study of variables that may influence results from analyses. Unadjusted analyses of baseline and follow-up ASI differences revealed no differences in six of seven ASI domains; for psychiatric functioning, SouthState had better outcomes than MidState. Using Heckman selection models to adjust for the selection effect of background variables and prior substance use information (available for all N=1020 recruited whether interviewed or not) from county databases, differences were also found the Medical ASI outcome (p=.05) with Midstate producing better outcomes than SouthState. More years of substance use emerged as the best predictor of a higher likelihood of selection out of the sample. Attention to differences in respondents' substance use characteristics, especially for those refusing to participate in studies, may be critical to evaluating program performance.

Learning Objectives:

Keywords: Alcohol Problems, Evaluation

Presenting author's disclosure statement:

I wish to disclose that I have NO financial interests or other relationship with the manufactures of commercial products, suppliers of commercial services or commercial supporters.

Alcohol Consumption Distributions: Interventions and Policies Poster Session

The 133rd Annual Meeting & Exposition (December 10-14, 2005) of APHA