Disclosure Analyses Undertaken by SAMHDA

The Substance Abuse and Mental Health Data Archive (SAMHDA) at ICPSR is engaged in formal disclosure analyses for the sponsor of the SAMHDA project, the Office of Applied Studies at the Substance Abuse and Mental Health Services Administration. The disclosure analysis for the Treatment Episode Data Set (TEDS) has been completed. TEDS provides annual administrative data on approximately 1.5 million treatment admissions per year. For TEDS, a record swapping procedure was used for a small percentage of records in TEDS that posed a risk of re-identification of an individual whose record was included in the file. For more information on the TEDS data protection procedures, see the 1998 TEDS codebook on the TEDS Series page, under "Data and Documentation."

For the Alcohol and Drug Services Study (ADSS) a disclosure protection technique called micro-aggregation was used. Micro-aggregation involves identifying problematic variables (e.g., those that could be linked to other files or that contain data that are publicly available and could be used to identify an individual or organization), grouping records by three for each problematic variable, and then applying the average of the three records for each grouping to each record in the group. ADSS links treatment facility data to abstracted client records and client follow-up interviews. Two facility-level variables posing a risk of identification were micro-aggregated and 191 additional variables were recalculated based on the micro-aggregated variables. This procedure allowed the linkage between the files to remain and nearly all of the client data to be released. The disclosure techniques are summarized in each ADSS codebook on the ADSS page, under "Data and Documentation."