OIG: Medicare Race, Ethnicity Data Is Inaccurate, Thwarts Health Equity

The Office of Inspector General (OIG) has discovered that Medicare’s race and ethnicity data was less accurate for minority groups—particularly for those who identify as American Indian or Alaska Native, Asian or Pacific Islander, or Hispanic—and that these discrepancies could have real impacts on health equity efforts.

“The Centers for Medicare & Medicaid Services (CMS) has made advancing health equity a top priority. Ensuring that Medicare is able to assess disparities is key to this goal,” the report explained. “The ability to assess health disparities hinges on the quality of the underlying race and ethnicity data.”

OIG used Medicare’s enrollment database which draws on Social Security Administration data and a CMS algorithm to assess the accuracy of Medicare’s race and ethnicity data. The agency compared the results from this source to the results of self-reported data from Medicare beneficiaries in nursing homes as well as federal race and ethnicity data collection standards.

OIG uncovered that Medicare’s data did not match the accuracy of the other data sets, despite the algorithm, nor did Medicare’s data collection comply with federal standards.

Almost half of all beneficiaries who were identified as American Indian/Alaska Native in the Medicare enrollment data (46 percent) did not identify themselves as such in the nursing home assessment. 

Additionally, nearly three in ten respondents in the nursing home assessment who were identified as Hispanic did not identify themselves as Hispanic (28 percent). And 17 percent of those identified as Asian/Pacific Islander did not ascribe to that group in the self-reported data.

Not only were certain beneficiary populations more likely to be identified as a different race or ethnicity, but they were also more likely to not be identified as their correct race or ethnicity.

Among beneficiaries who self-reported their ethnicity as Hispanic, 13 percent were not identified as Hispanic in the Medicare database. Thirty-five percent of American Indian/Alaska Native beneficiaries and nearly a quarter of the beneficiaries who identified as Asian/Pacific Islander were not identified as such in the Medicare data. 

A history of inadequate race and ethnicity categories in the Social Security Administration’s data was the key cause behind the inconsistencies in Medicare’s database, along with incompleteness as 3.3 million Medicare beneficiaries’ racial or ethnic identities are recorded as “unknown,” “other,” or remain blank.

Moreover, federal race and ethnicity data collection standards require that entities ask about race and ethnicity separately, permit beneficiaries to select multiple racial and ethnic categories, and provide a comprehensive list of racial and ethnic categories. However, Medicare has not complied with these standards.

The agency recognized that an effort to improve race and ethnicity data would be challenging, but urged CMS to take up this initiative, given the serious implications for health equity.

First, OIG recommended that CMS create its own source of race and ethnicity data. Currently, CMS takes the data from the Social Security Administration and runs it through an algorithm called the Research Triangle Institute algorithm. Instead, CMS could collect this information as part of enrollment, OIG suggested.

The agency said that CMS did not explicitly agree to comply with this recommendation.

Second, the agency suggested that CMS rely on self-reported data for race and ethnicity information. The report stated that self-reported data tends to be more reliable. Existing self-reported assessments should be incorporated into the Medicare dataset. CMS can also create new opportunities for collecting self-reported race and ethnicity information.

CMS agreed to this recommendation.

Third, CMS should standardize its race and ethnicity data collection processes. CMS agreed to this recommendation.

And finally, CMS should ensure that beneficiaries are aware of these efforts to pursue health equity through better data collection.

“Beneficiaries may be reluctant to provide this information to CMS if they do not understand how it will be used. CMS should develop and implement a plan to address these concerns that includes education and communication strategies,” OIG explained.

CMS agreed to this recommendation.

In 2021 before OIG conducted this report, Better Medicare Alliance called on CMS to standardize data collection processes for race and ethnicity information as well as gender and social determinants of health data.

CMS is not the only major payer that struggles to assess enrollees’ racial and ethnicity data. Both the federal and state health insurance marketplaces across the US are seeking solutions for collecting this information, with 22.6 percent of enrollees on the state health insurance marketplace declining to reveal race or ethnicity data.

The agency has indicated its intentions to improve race and ethnicity data collection processes as part of its overall health equity strategy, released in April 2022. CMS identified community-based organizations as key partners in gathering this data.