Assessing Potential Precision Gains Using Imputation to Combine Administrative Data with Sampled Outcomes

• Activity Number: 150
• Type: Topic Contributed
• Date/Time: Monday, August 9, 2004 : 2:00 PM to 3:50 PM
• Sponsor: General Methodology
• Author(s): Thomas R. Belin*+ and Jun Xing
• Companies: University of California, Los Angeles and University of California, Los Angeles
• Keywords: multiple imputation ; quality of care ; health services

Abstract

Research on quality of care often investigates whether certain procedures or medications are being used in treating patients. Some relevant information may be available from administrative data sources, which is apt to be relatively inexpensive, while other information may only be available from more expensive forms of inquiry. A motivating application for the present work involved a comparison of two public health systems, a county system and a Veterans Administration system located in a large urban area in the United States, regarding quality of care for schizophrenic patients, where data were available from administrative records on medication prescriptions, medical chart reviews, and patient interviews. Because the latter two methods of data collection are more expensive, it was of interest to investigate whether the more expensive data sources could be collected on a sample basis, with missing-data techniques used to support inferences comparing health systems. We investigate this question using both an available sample of 224 schizophrenic patients and a simulation evaluation that allows for alternative sample sizes and correlation structures.