Depending on the source used, estimates of uninsurance rates and disparities in these rates between population groups can differ substantially, according to this analysis comparing data from three national surveys regarding the American Indian/Alaska Native (AIAN) population.

AIANs continue to have significant health disparities relative to the rest of the nation but the population has received limited attention and researchers already have identified data gaps and weaknesses. This analysis examines whether three national surveys produced “trustworthy and consistent” estimates of disparities and rates of uninsurance among AIANs; and highlights the impact the choice of data source could have on analytic conclusions. The data examined related to health insurance coverage of AIANs younger than age 65 from the Current Population Survey (CPS), the National Health Interview Survey (NHIS) and the Medical Expenditure Panel Survey (MEPS).

Key Findings:

• Estimates of AIAN sociodemographic characteristics differed significantly across the sources.
• Estimates of all-year uninsurance rates for AIAN children varied from three to 23 percent and, for AIAN adults, from 18 to 35 percent.
• While the CPS estimated that AIAN children were 2.9 times more likely to be uninsured all-year compared to non-Hispanic white children, the NHIS and MEPS detected no significant disparities.
• Variations among estimates of uninsurance rates for partial years were even greater: the NHIS estimates AIAN children were 3.7 times more likely to be uninsured than non-Hispanic white children; the MEPS found no difference.

This analysis demonstrates the choice of data source matters to research results. The authors postulate the causes include differences in data collection and processing. But this analysis may reveal the inadequacy of national surveys for collecting data for small sub-populations. It is critical researchers understand the limitations of data sources when evaluating health care disparities.