Racial and Ethnic Disparities in COVID-19 Cases and Deaths in Nursing Homes


Data Sources

This analysis draws on federal data published by the Center for Medicare and Medicaid Services (CMS)and additional facility-level data from Brown School of Public Health’s LTCfocus.org data, which includes summary-level information derived from 2017 MDS (Minimum Dataset) data.

data on cases and deaths

Data on coronavirus cases and deaths in nursing homes is from federal data published by the Center for Medicare and Medicaid Services (CMS). CMS requires all Medicare and Medicaid certified nursing facilities to report data on suspected and confirmed coronavirus cases and deaths for residents and staff. These data collection standards make the federal data more comparable across states than state-published data. Data in this paper reflects nursing home cases and deaths as of October 11, 2020. A major limitation of this data is the time period included: facilities were only required to report data starting May 8th, 2020 and reporting cases or deaths prior to that date was not required. Thus, data may not be counting cases or deaths in facilities or states that experienced outbreaks earlier in spring 2020. For more information on the federal data, see KFF analysis comparing the state-reported data to the federal data.

In addition, some nursing homes in the CMS data report a greater number of cases and/or deaths than the total number of beds in the nursing home. We dropped approximately 1,000 nursing homes for which this was the case, leading to a final sample of 13,982 nursing homes. This sample is a subset of the approximately 15,000 nursing homes in the US.

Data on Race/Ethnicity of Residents

CMS does not require nursing facilities to report cases and deaths by race/ethnicity. Therefore, facility-level data on resident race/ethnicity was pulled from Brown School of Public Health’s LTCfocus.org data, which includes summary-level information derived from 2017 MDS (Minimum Dataset) data. MDS assessments are completed for all residents in nursing homes to identify each resident’s functional capabilities and help nursing home staff identify health problems. Demographic data, including race/ethnicity, is collected during this process as well. Brown School of Public Health provides summary data of MDS from 2010-2017, including facility-level data on the share of nursing home residents who are Black, White, or Hispanic. This analysis does not include data on staff cases and deaths since MDS does not collect data on racial makeup of nursing home staff, and we were unable to identify a data source for race/ethnicity of nursing home staff at the facility level.

The LTCfocus.org data censors data points representing fewer than 11 residents due to privacy standards set forth by CMS. Due to the small number of Black or Hispanic residents in many nursing homes, this created missing data points in nearly 9,000 nursing homes. For these “missing” data points, we imputed a value of 5.5 residents and calculated shares of residents based on the newly imputed value. We conducted a sensitivity analysis that showed that results were largely unchanged when the sample included and excluded the nursing homes with imputed data.

Classifying Facilities by Resident Race/Ethnicity

Overall, within the 13,982 nursing homes included in this analysis, 12% of all nursing home residents are Black and 6% are Hispanic. Nursing homes have a disproportionately low share of Hispanic residents when compared to total US population distribution by Race/Ethnicity. We categorized nursing homes where 20% or more of residents were Black or Hispanic as “High Share of Black Residents” or “High Share of Hispanic Residents.” This definition is based on the distribution of share of Black and Hispanic residents among the 13,982 nursing homes in our sample and a need to have adequate sample within each group. Many nursing homes reported extremely low shares of Black or Hispanic residents or zero Black or Hispanic residents. The following are the Ns for the groups reported in the paper:

Number of Facilities Included in Analysis, by Race/Ethnicity Group
  Category Definition Total Facilities
Total US Facilities 13,982
Low Share of Black Residents Under 20% 10,995
High Share of Black Residents 20% or greater 2,987
Low Share of Hispanic Residents Under 20% 12,995
High Share of Hispanic Residents 20% or greater 987
Low Share of White Residents Under 80% 5,276
High Share of White Residents 80% or greater 8,706

Residents may be double counted in race/ethnicity data (for example, a resident could be counted as both “White” and “Hispanic”). Therefore, some facilities with a “high” share of Black or Hispanic residents may overlap with facilities with a “high” share White residents.

Choosing States for State-Level Analysis

State-level analysis was limited to the states with 50 or more facilities in the high share of Black or Hispanic residents categories, based on our definitions of “high share” described above. In total, we were able to provide state-level data on 21 states. California, Florida, New York, and Texas each had 50 or more nursing homes with a high share of both Black and Hispanic residents, and the remaining 17 states in the analysis had only a sufficient number of nursing homes with a high share of Black residents.

Data Note Table

KFF Headquarters: 185 Berry St., Suite 2000, San Francisco, CA 94107 | Phone 650-854-9400
Washington Offices and Barbara Jordan Conference Center: 1330 G Street, NW, Washington, DC 20005 | Phone 202-347-5270

www.kff.org | Email Alerts: kff.org/email | facebook.com/KFF | twitter.com/kff

The independent source for health policy research, polling, and news, KFF is a nonprofit organization based in San Francisco, California.