programs since 1984, although no patient-level data are collected. In contrast, the major strength of the DATOS research is the extensiveness of patient-level data obtained from more than 10,000 adults in 99 substance abuse treatment programs between 1991 and 1993. Researchers using these data are attempting to evaluate how changes in program resources and the financing of treatment, organizational structure and program design, and treatment practices and treatment duration relate to patient outcomes, controlling for patient and environmental characteristics.
Both the ODATS and DATOS data pose limitations, however, for researchers addressing questions about the management of substance abuse treatment programs. While there is considerable information in ODATS about the organization of treatment programs, there are no patient-level data that would allow for analyses of the interrelationships between treatment service delivery and individual patient characteristics, (which define the primary clinical level), and institutional factors and administrative policies that operate at one (or more) levels above the clinical domain. The DATOS data are comparatively rich in patient-level information, but the limited amount of data collected about the programs (i.e., their structures, environments and other relevant characteristics) and linked to the patient-level data likewise diminishes their usefulness for exploring issues of program governance and management and their relationship to patient outcomes.
Still another study was the Services Research Outcome Study (SROS), based on data from 1,799 clients in a nationally representative sample of the population of substance abuse treatment program clients, September 1989-August 1990. Drug use and behavior of the clients in this sample were studied over a period of five years. As described by Schildhaus et al. (2000a:1851), “the strength of representativeness gives us a sample of ‘good’ and ‘bad’ treatment programs, programs that ‘match’ clients and treatment and programs that do