Uncovering the Missing Medicaid Cases and Assessing their Bias for Estimates of the Uninsured

Abstract

General population surveys of health insurance coverage are thought to undercount Medicaid enrollment, which may bias estimates of the uninsured. This article describes the results of an experiment undertaken in conjunction with a general population survey in Minnesota. Responses to health insurance questions by a known sample of public program enrollees are analyzed to determine possible reasons for the undercount and the amount of bias introduced in estimates of uninsured people. While public program enrollees often misreport the type of coverage they have, the impact on estimates of those without insurance is negligible. Restrictions to generalizing the finding beyond this study are discussed.

Accuracy in Self-Reported Health Insurance Coverage among Medicaid Enrollees

“…It is also possible that Medicaid programs need to do a better job of informing people of their enrollment status and benefits. However, together with the results from our original Minnesota experiment (Call et al 2001), the findings from this study in three additional states suggest that the Medicaid undercountwhile a source of some bias in state general population survey estimates of specific types of coverage and uninsurance-should not undermine the confidence with which policymakers use and rely on uninsurance estimates. An important goal of this project, which we hope has been accomplished by these results, is to restore policymakers' trust in the use of point-in-time survey data to monitor health insurance coverage, eligibility for coverage, and characteristics of the uninsured, and how these change over time.…”

The Accuracy of Reported Insurance Status in the MEPS

“…3 Respondents who did not report their Medicaid coverage reported other coverage or uninsurance. Studies of the CPS and statespecific surveys found that many of those who did not report their Medicaid coverage reported other public programs or private coverage (Call et al 2001;. Among nonelderly CPS sample members with Medicaid, 24.6% reported other public or private coverage , much more than the 6.4% to 16.7% of Medicaid enrollees reported in three state surveys ).…”

“…2 Call et al (2001), however, found higher underreporting, but they used the Minnesota Health Access Survey, which did not ask about Medicaid if the person reported Medicare, military programs, veterans programs, or the Indian Health Service. 3 Similarly, Cohen and Martinez (2006) found that adding state program names to the NHIS increased reported Medicaid and SCHIP coverage.…”

On The Road To Universal Coverage: Impacts Of Reform In Massachusetts At One Year

“…5 Although it is believed that most people accurately report in surveys whether they have insurance coverage, there is evidence of some misreporting of coverage type. 6 In Massachusetts, where several of the public programs have similar names, respondents often reported being enrolled in multiple programs (for example, CommCare and CommChoice) or having both nongroup and public coverage. Because this raises concerns about the accuracy of the reporting of coverage type for the various public programs and nongroup coverage, the analysis of source of coverage is limited to those with employer coverage and those with all other types of insurance.…”