Medicaid Reporting Errors in Four National Surveys: ACS, CPS, MEPS, and NHIS

• Activity Number: 548
• Type: Topic Contributed
• Date/Time: Wednesday, August 6, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Brett Fried*+ and Michel Boudreaux and Kathleen Thiede Call and Joanna Turner
• Companies: SHADAC/University of Minnesota and SHADAC/University of Minnesota and SHADAC/University of Minnesota and SHADAC/University of Minnesota
• Keywords: ACS ; Medicaid undercount ; validation study ; survey and administrative data ; health insurance coverage

Abstract

Measuring health coverage is challenging and every survey is said to undercount Medicaid enrollment. We use 2009 linked ACS and Medicaid Statistical Information System (MSIS) data to examine the extent to which Medicaid enrollment is misreported. Drawing on work from the SNACC team, a project examining the Medicaid undercount in federal surveys, we compare the magnitude of the undercount and factors associated with misreporting in the ACS, the NHIS, CPS and MEPS. We examine two sources of the discrepancy between administrative and survey counts of enrollment: concept alignment and measurement error. Comparing survey and raw MSIS counts after aligning concepts to the extent possible, we find that the CPS, NHIS and MEPS undercount Medicaid enrollment whereas the ACS overcounts by 15.8%. This is likely a function of the ACS question that asks about Medicaid and any kind of government plan whereas the other surveys have a stand-alone Medicaid question. Using the linked ACS/MSIS data we find that 21.6% of known Medicaid enrollees are not coded as such in the survey. This is comparable to the 17.5% undercount in the 2003 MEPS, and lower than the 2005 CPS (42.9%) and 2002 NHIS (32.2%).