National Drug Abuse Treatment System Survey, Waves II-IV (ICPSR 4146)
Alternate Title: NDATSS
Principal Investigator(s): D'Aunno, Thomas, University of Michigan. Institute for Social Research; Price, Richard, University of Michigan. Institute for Social Research
The National Drug Abuse Treatment System Survey (NDATSS) is a longitudinal program of research into organizational structures, operating characteristics, and treatment modalities of outpatient drug treatment programs in the United States. This is done through interviews with program directors and clinical supervisors. In some publications, this research is referred to as the Outpatient Drug Abuse Treatment Studies (ODATS). Data being released include Wave II (1988), Wave III (1990), and Wave IV (1995).
These data are freely available.
Substance Abuse & Mental Health Data Archive
This study is maintained and distributed by the Substance Abuse & Mental Health Data Archive (SAMHDA). SAMHDA is supported by the Substance Abuse & Mental Health Services Administration (SAMHSA), Center for Behavioral Health Statistics and Quality (CBHSQ), formerly the Office of Applied Studies.
D'Aunno, Thomas, and Richard Price. National Drug Abuse Treatment System Survey, Waves II-IV. ICPSR04146-v1. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2009-07-30. http://doi.org/10.3886/ICPSR04146.v1
Persistent URL: http://doi.org/10.3886/ICPSR04146.v1
This study was funded by:
- United States Department of Health and Human Services. National Institutes of Health. National Institute on Drug Abuse (5ROI-DA03272)
Scope of Study
Subject Terms: AIDS, alcohol abuse, drug abuse, drug treatment, health care services, HIV, intervention, methadone maintenance, substance abuse, substance abuse treatment, treatment facilities, treatment programs
Geographic Coverage: United States
Date of Collection:
Unit of Observation: facility
Universe: Outpatient Drug Abuse Units in the United States
Data Types: survey data
Data Collection Notes:
The data were collected by Survey Research Operations, Institute for Social Research, University of Michigan.
Wave I is not planned for public release because it had a significantly different sample design than the other waves.
Study Purpose: To understand the organizational determinants of treatment practices in a national sample of substance abuse treatment units.
Study Design: Stratified Random Sample
Though Wave I is not a part of this public release, its sample design is relevant due to the longitudinal design of the study. A stratified random sample of n = 413 outpatient drug treatment units were selected. The sample was selected from national databases of federally supported treatment programs and was restricted to outpatient drug treatment programs that were drug-free (i.e., they did not use methadone or related treatments). Alcohol treatment programs were excluded. The primary stratification criteria included: location (inner city, suburban, rural), funding agency (National Institute of Mental Health (NIMH) or National Institute on Drug Abuse (NIDA)), and program classification (freestanding versus those based in a community mental health clinic (CMHC)).
Wave II: As mentioned previously, the survey population for Wave II was expanded to include methadone-based treatment programs. The Survey Research Center's (SRC) sampling section developed a comprehensive national frame of approximately 8,149 program listings to meet the expanded sample requirements. The primary stratification criteria of the National Sample Frame of Drug Treatment Programs (NSFDTP) had: public versus private program, methadone versus drug-free treatment, and program type (CMHC, hospital, or freestanding). After selecting a stratified random sample, screening for eligibility, and selecting a stratified subsample, 438 programs were chosen for inclusion in the survey. In addition to these programs, an additional 207 programs from Wave I were deemed to be eligible programs and were added as a recontact sample, A total of 575 programs provided a program director interview, a clinical supervisor interview, or both.
Wave III: No sample of new programs was included in Wave III, which was a longitudinal follow-up of the programs that were successfully interviewed in Wave II. In Wave III, 550 programs were found to be still operating and eligible, and 481 total programs provided a program director interview, a clinical supervisor interview, or both.
Wave IV: This wave included both a longitudinal sample of those programs that completed waves II and III, and was also supplemented with a new stratified random sample from a completely updated National Frame of Substance Abuse Treatment Programs (NFSATP).
Interviews were attempted with the 429 programs that participated in Waves II and III, and were also still operating and eligible. In Wave IV, 387 of these programs provided a program director interview, a clinical supervisor interview, or both.
The sample frame for the new supplemental cross sectional sample is a subset of the NFSATP. The full NFSATP database is the integration of four drug treatment program databases:
- National Facilities Register (NFR), 1992 NDATUS (a predecessor to the National Survey of Substance Abuse Treatment Services, N-SSATS) -- provided to the SRC by the Substance Abuse and Mental Health Services Administration (SAMHSA).
- The 1994 American Hospital Association (AHA) Survey -- added to ensure comprehensive coverage of hospital-based programs.
- 1994 Food and Drug Admisistration (FDA) List of Licensed Methadone Treatment Programs -- printed list provided to SRC by the FDA at the request of SAMHSA.
- The Survey Sampling, Inc. (SSI) Business Data Base -- purchased in order to ensure comprehensive coverage of privately operated programs.
The Wave IV sample was assigned to 12 stratum codes as defined by the same primary stratification criteria as existed for the Wave II sample (public vs. private program, methadone vs. drug-free treatment, and program -- CMHC, hospital, or freestanding). However, the database components of the NFSATP did not always supply all the necessary information needed to classify each program to one of the 12 strata. These programs with incomplete information for stratum assignments were, as a rule, included in the private, nonhospital, nonmethadone stratum.
It was determined that the outpatient drug treatment program sector that had experienced the most growth correlated with the private, nonhospital, nonmethadone stratum. Thus it was decided that the supplemental sample would be a subsample of the NFSATP programs that were known to be private, nonhospital, nonmethadone programs or for which there was incomplete information for proper classification into one of the 12 stratum as described above. After initial screening of this subsample for eligibility, 270 programs were determined to be eligible for interviews. Of these, 231 total programs provided a program director interview, a clinical supervisor interview, or both. Based on the more complete information gathered from the interviews, 81 programs were reclassified into one of the other 11 stratum.
Each of the three data files come with two weight variables, a trimmed weight and a centered weight. Both of these weight variables are derived from an original weight variable (not included in the public data file). For each wave of the data, the original weight variable was produced by the principal investigator in order to reduce bias, which may be associated with unequal sample selection probabilities and nonresponse. Trimmed and centered weights are defined as follows:
Trimmed Weight: One disadvantage of weighting to decrease bias is that it might increase the variance of weighted estimates. When it appears that a few weights are disproportionately affecting the variance, then outlier weights are often trimmed. With trimming it is implicitly assumed that the potential increase in bias is small relative to the reduction in variance. The ultimate goal of weight trimming is to reduce the sampling variance more than enough to compensate for the possible increase in bias and, thereby, to reduce the mean squared error. For this study, the trimming consisted of reducing outlier weights to a value slightly larger than the next largest value within the same stratum. To maintain the total sum of all weight values, the excess "trimmed" weight was then redistributed to the other weight values within the same stratum.
Centered Weight: When a typical weight is applied, the weighted estimates are inflated to the full population value from which the sample was drawn. This results in standard errors that are inappropriately deflated. Centered weights are commonly used to avoid this problem. The centered weights for this study are calculated by taking the trimmed weight for each case and dividing by the mean of the trimmed weight variable.
Mode of Data Collection: telephone interview
Description of Variables: The data include variables from the following major areas of investigation: referral sources, client demographics, types of problems treated, unit staffing, diagnostic and assessment patterns, treatment goals, services provided, end-of-treatment outcomes, quality improvement efforts, funding, licensing and accreditation, collaboration and competition with other treatment organizations, and recent programmatic changes.
1988: 86 percent
1990: 88 percent
1995: 87 percent
Extent of Processing: ICPSR data undergo a confidentiality review and are altered when necessary to limit the risk of disclosure. ICPSR also routinely creates ready-to-go data files along with setups in the major statistical software formats as well as standard codebooks to accompany the data. In addition to these procedures, ICPSR performed the following processing steps for this data collection:
- Created variable labels and/or value labels.
- Standardized missing values.
- Created online analysis version with question text.
- Performed recodes and/or calculated derived variables.
- Checked for undocumented or out-of-range codes.
Original ICPSR Release: 2009-07-30
- View publications for the study (~50)
Most Recent Publications
Browse Matching Variables
DS1: Wave II, 1988
- Citations exports are provided above.
Export Study-level metadata (does not include variable-level metadata)