ENGLISH :
This paper identifies and illustrates the challenges of conducting cost-effectiveness analysis (CEA) of addiction treatments given the multiple important outcomes of substance abuse treatment (SAT). Potential problems arise because CEA is intended primarily for single outcome programs, yet addiction treatment results in a variety of outcomes such as reduced drug use and crime and increased employment. Methodological principles, empirical examples, and practical advice are offered on how to conduct an economic evaluation given multiple outcomes. An empirical example is provided to illustrate some of the conflicts in cost-effectiveness (CE) ratios that may arise across the range of outcomes. The data are from the Philadelphia Target Cities quasi-experimental field study of standard versus "enhanced" (e.g. case management and added social services) drug treatment. Outcomes are derived from of the Addiction Severity Index (ASI), while cost data were collected and analyzed using the Drug Abuse Treatment Cost Analysis Program (DATCAP). While the results are illustrative only, they indicate that cost-effectiveness ratios for each of several different outcomes can produce conflicting implications. These findings suggest that multiple outcomes should be considered in any economic analysis of addiction treatments because focusing on a single outcome may lead to inadequate and possibly incorrect policy inferences. However, incorporating multiple outcomes into a CEA of addiction treatment is difficult. Cost-benefit analysis (CBA) may be a preferable and more appropriate approach in some cases. (Review's abstract.)