Preparing Medicare Claim Data for Stochastic Statistical Disclosure Limitation Treatment

• Activity Number: 332
• Type: Topic Contributed
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Health Policy Statistics
• Author(s): Tzy-Chyi Yu*+ and Elizabeth C. Hair and Beth A. Virnig
• Companies: NORC at the The University of Chicago and NORC at the The University of Chicago and University of Minnesota
• Keywords: Comparative effectiveness research ; Medicare claims ; statistical disclosure limitation ; data utility ; inside intrusion

Abstract

Comparative effectiveness research (CER) analysts prefer data at the individual level. However, data sets containing information required by CER researchers often risk disclosing sensitive information about individuals in the data through indirect identifiers. Statistical disclosure limitation (SDL) methods are used to reduce this risk, while maintaining analytic utility. When preparing data for SDL it is often necessary to aggregate data from multiple records, within numerous tables to create a single record describing the disc losable entity in detail. In the case of Medicare claims this requires an approach that scans records from multiple claim tables (inpatient, outpatient, prescription drug, etc.) to create a beneficiary profile representative of what an intruder could use to disclose sensitive information.