Beginning in 2025, the U.S. Census Bureau revised the health insurance question used in the American Community Survey (ACS) based on results from the 2022 ACS Content Test. Since its implementation in 2008, the ACS health insurance question underwent ongoing research on measurement quality indicating three areas for survey question improvement: an over-reporting of direct-purchase plans, an under-reporting of Medicaid coverage, and an over-reporting of multiple plan-types. Additionally, the answer choices did not mention Marketplace plans, since they did not exist when the question was originally designed.
To address these issues, the response categories for the health insurance question were changed in a number of ways. Several answer choices were reordered, such that direct purchase coverage was moved down in the list (after Medicare and Medicaid), while TRICARE and Veteran’s health care changed places. Instructions were also added to exclude single-service health plans, to avoid inflating estimates of direct purchase coverage as well as multiple plan-types.
There were also improvements to the wording of several of the answer choices. The words “Children’s Health Insurance Program (CHIP)” were added to the Medicaid option to help with Medicaid underreporting, while the terms “union or professional association” were added to the employer sponsored health coverage option. And the terms “State or Federal Marketplace” and “HealthCare.gov” were added to the wording of the direct purchase coverage option.
Finally, based on recommendations from cognitive testing, the question format changed from a yes or no response for each individual insurance type, to a more efficient mark all that apply list of coverage types, with a new “Uninsured” checkbox.
Along with these question changes, the Census Bureau implemented health insurance data processing improvements. As the write-in fields may risk inflating direct purchase or multiple coverage types, the decision was made to only code write-ins when the respondent did not select a specific insurance type. Finally, data editing improvements were made to improve logic and prevent over-imputation of multiple coverage types.
Due to these changes to the health insurance question and its processing, the Census Bureau recommends data users compare health insurance estimates to previous years with caution. Data users should also be aware that the estimates for health insurance coverage in table CP03 are suppressed for the 2021, 2022, 2023, and 2024 periods within 2025 and subsequent 1-year Comparison Profiles.
The health insurance estimates will be suppressed in the 2021-2025 ACS 5-year tables and the health insurance variables in the Public Use Microdata Sample (PUMS) files will not be available until the research to bridge the estimates and variables can be completed. We expect to publish 5-year health insurance estimates with the 2022-2026 ACS 5-year data release.
Although health insurance estimates will be excluded from the 2021-2025 ACS 5-year products, four demonstration tables will be released that will provide unofficial 5-year estimates of insured and uninsured rates overall and by age and sex for the levels of geography covered by the standard ACS 5-year tables. These tables will not bridge the change in methodology, but rather assume that high-level reports of insured vs uninsured are consistent across the 5-year period, which contains data before and after the change to the health insurance question.
For more information, visit: