print share

CHIS Data Quality

CHIS 2003: Comparing Medicaid Estimates from CHIS with other Data Sources

One of the key health care indicators used by policy analysts and researchers are measures of health insurance coverage. Although there are no "gold standards" with which to judge the accuracy of survey-based estimates of private health insurance coverage, Medicaid administrative data provide such a standard to compare survey estimates of the number of persons enrolled in Medicaid.

Estimates of Medicaid enrollment based on population surveys typically estimate a lower number of persons enrolled in Medicaid compared to administrative data. If surveys undercount the Medicaid enrolled population they likely over-count the number of uninsured persons. Accurate estimates of Medicaid enrollment are therefore critical for estimating the size of the population that health care reforms should attempt to cover. This is especially true in a large state like California where a percentage point difference can impact a large number of persons and have significant cost implications.

In order to assess potential Medicaid undercount in CHIS, we developed a study to compare CHIS 2003 estimates of Medi-Cal (California's Medicaid program) enrollment with estimates drawn from two federal surveys and administrative data. We compared estimates from CHIS with estimates based on the California samples of the National Health Interview Survey (NHIS, the nation's "gold standard" health survey conducted by the National Center for Health Statistics) and the Current Population Survey (CPS, conducted by the U.S. Census Bureau and one of the most widely cited sources of health insurance coverage estimates). Estimates from these three population-based surveys were then compared with administrative enrollment data from the California Department of Health Care Services. Findings suggest that the CHIS estimates closely match enrollment counts based on administrative data, while both NHIS and CPS estimates undercount Medi-Cal enrollment. (The results of this study have been submitted to a peer-review journal and will be made available on this web site when the paper is published.)

See further documentation on this (PDF)