Imputation in a National Health Survey: Balancing Data Quality with Respondent Burden in the Medical Expenditure Panel Survey (MEPS)

• Activity Number: 416 - Nonresponse Errors and Fixes
• Type: Contributed
• Date/Time: Tuesday, July 30, 2019 : 2:00 PM to 3:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Emily Mitchell* and Jerrod Anderson and Samuel H Zuvekas
• Companies: Agency for Healthcare Research and Quality and Agency for Healthcare Research and Quality and Agency for Healthcare Research and Quality
• Keywords: Medical Expenditure Panel Survey; Survey Data; Imputation; Measurement Error; Respondent Burden

Abstract

The Medical Expenditure Panel Survey (MEPS) collects payment data from households for reported medical events. This household-reported data, however, is often incomplete and inaccurate. More accurate data for medical services are collected from households’ medical providers. Provider-reported payment information is applied to medical events when available. For events without provider data, but with complete household data, household payment data are applied. For the remaining cases, incomplete payment data are imputed. In this analysis, we assess the impact of reducing data collected from the household that is most likely to be inaccurate. Doing so would reduce the burden on the household respondent, but would rely more heavily on imputation to complete the missing data. We assess the accuracy of household-reported data by comparing it to provider-reported data. We then compare imputed household data to actual provider data to assess the performance of the imputation algorithm. Finally, we simulate imputing additional household expenditures to assess the impact of reducing payment data collected from the household on the accuracy of MEPS expenditure estimates.