5080.0: Wednesday, November 15, 2000 - 9:30 AM

Abstract #2842

The Productivity of Outpatient Treatment for Substance Abuse

Mingshan Lu, Department of Economics and Department of Community Health Sciences, University of Calgary, University of Calgary, 2500 University Dr NW, Calgary, AB T2N 1N4, Canada, (403) 220-5488, lu@ucalgary.ca and Thomas G. McGuire, Department of Economics, Boston University, 270 Bay State Road, Boston, MA 02215.

This paper studies the effectiveness of outpatient care for substance abuse (mainly alcohol abuse) taking place in private non-profit agencies during more than 10,000 episodes of treatment occurring over four years in Maine. Effectiveness is measured as the reduction in the rate of drug use between admission and discharge. In a non-experimental setting without a formal "control group," we take steps to obtain unbiased estimates of the effect of treatment in an ordered logit model framework. Following recent papers measuring treatment effects, we employ instruments for treatment to deal with the potential bias introduced by a correlation of quantity of treatment with unmeasured factors affecting change in rates of drug use. We found consistent evidence that more treatment for drug abuse improves outcomes for clients with moderate or heavy drug use problems. Evidence for a "regression-to-the-mean" appeared in some specifications. Furthermore, we found evidence for diminishing marginal treatment productivity. After controlling for the selection bias, marginal effects of treatment appear to be positive for most clients. The results in this paper are supportive of the position that the selection bias associated with treatment in a non-experimental setting cannot be ignored and that IV methods are helpful. In the IV model regressions, higher treatment effects are found for the majority user groups. This result remains robust to different choices of IVs as well as different model specifications. In our data set, the unobservable factors caused a downward bias in traditional model estimation.

Learning Objectives: At the conclusion of the session, the participant (learner) in this session will be able to: 1. Recognize the importance of using information from real-world settings (non-experimental setting) to study the effects of treatment, as well as the related problem of selection bias that needs to be controlled for; 2. Describe and apply the statistical methodology of Instrumental Variable approach (IV) to control the selection bias and estimate treatment effectiveness; 3. Discuss the strength and weakness of applying IV approach in health treatment outcome evaluation

Keywords: Treatment Outcomes, Evaluation

Presenting author's disclosure statement:
Organization/institution whose products or services will be discussed: Publicly funded non-for-profit substance abuse agencies in the state of Maine
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.

The 128th Annual Meeting of APHA