How the 2020 census miscount jeopardizes health in communities of color

Author(s)
Published on
May 3, 2022

The U.S. Census undercount of Black, Hispanic and American Indian residents in 2020 poses significant risks to the health of the nation, especially its most vulnerable.

In early March, the U.S. Census Bureau reported undercounting these populations nationally, while overcounting non-Hispanic White populations. In a recently published essay in The Philadelphia Inquirer, we described the ways in which these undercounts could threaten the health of Pennsylvanians. Stories like this can be done in cities and states across the country; all are susceptible to the health implications of census miscounts.

Miscounts are not new. The 2010 census also undercounted Black, Hispanic and American Indian populations. But the latest shortcomings are greater, especially for the Hispanic population, which had an undercount rate of 4.99% in 2020, compared with 1.54% 10 years earlier. Meanwhile, overcounting of non-Hispanic white residents increased.

As a result of the combination of over- and undercounting at the national level, the overall population estimate for United States, 331,449,281, is roughly on target. However, the counts for communities that have large Black, Hispanic or American Indian populations are likely to be underestimated in ways that jeopardize their health.

First, undercounting could lead to a loss of federal funding for Medicaid and the Children’s Health Insurance Program (CHIP), which cover more than 85.8 million people, as of April 1, 2022. Census undercounting results in a reduction in the Federal Medical Assistance Percentage (FMAP), the formula used to determine federal reimbursement for each state dollar spent on these programs. The minimum FMAP is 50%. Census miscounts put states with FMAPs greater than 50% at risk of funding cuts. 

This happened in 2015: Funding allocations based on 2010 census undercounting resulted in cuts to federal funding in the 37 states with FMAPs greater than 50%, according to a report by the GW Institute of Public Policy at George Washington University. These states lost a median of $1,091 for every person the census had missed.

The stakes are higher now. Kaiser Family Foundation reported in April that Medicaid and CHIP enrollment has increased by 20.5% since the start of the COVID-19 pandemic; 38 states will have FMAPs greater than 50% for fiscal year 2023. In other words, all but 12 states are susceptible to losing federal funding, even as demand for these programs is on the rise. 

Second, census miscounting can impact our response to public health threats. Since the start of the COVID-19 pandemic, for example, masking, vaccination, and testing policies in communities, schools and businesses have been based on the CDC’s metrics for assessing levels of community COVID-19 risk. These metrics include the number of new COVID-19 hospital admissions per 100,000 people — a denominator based on census counts.  Census miscounting in communities of color can result in flawed estimates of disease, leading to potentially misguided health policies and strategies. In 2020, the federal government allocated COVID-19 vaccines to states based only on census data (i.e., state populations) and, subsequently, states have been using census data to identify and target communities for vaccine distribution. 

The counts for communities that have large Black, Hispanic or American Indian populations are likely to be underestimated in ways that jeopardize their health.

Third, census miscounting can exacerbate longstanding racial and ethnic health disparities. In 2020, the COVID-19 pandemic quickly underscored the disparities experienced by communities of color. As of March 2022, the CDC reported that American Indian and Alaskan Native populations have died of COVID at more than twice the rate of the non-Hispanic white population, and Black and Hispanic populations have died at nearly twice the rate. Federal aid to support communities in responding to the pandemic, such as the American Rescue Plan, is largely reliant on population counts. Undercounting communities of color could mean that programs using 2020 census data shortchange the people and places disproportionately affected by poor health and fewer resources.  

Fourth, census miscounting hinders health research. Identifying critical health concerns in a community requires correctly characterizing the population. Census data allow physicians, researchers and health officials to know what health issues affect different communities For example, the “Stroke Belt” in the Southeastern U.S. is plainly visible on maps using census data. A flawed head count could produce misleading estimates that are problematic if they’re too high or too low. A census undercount can make a health problem seem worse than it is: If there are 10 stroke deaths per 100 people, as counted by the census, that’s a bigger problem than 10 per, say, 110 people, adding in the population the census missed. If stroke mortality actually improves or stays the same, but appears to be getting worse because of undercounting the population, a community may continue to receive resources needed elsewhere, or prevention programs that work may be ended because they look like they’re failing. Conversely, if stroke mortality is actually worsening but appears to be improving because of overcounting the population, the community may lose out on resources and prevention strategies may seem more effective than they are.

Fifth, other data sources on which researchers, journalists, and policymakers rely will be inaccurate. The census bureau conducts over 100 surveys each year, many of which are relevant for health and all of which rely on the census counts to enable representative samples. For example, the American Community Survey is used to help researchers identify health disparities and it guides plans for critical infrastructure, including hospitals and other services. During the pandemic, the Household Pulse Survey collected data on mental health, unemployment, vaccinations, vaccine hesitancy and other urgent issues. These data have provided near-real-time information on how the pandemic is affecting the country. Without correct population counts, estimates from these surveys become distorted and may not reflect the most pressing needs of a community. 

Journalists can track how the miscount plays out close to home. Here are four things to watch for:

Disparities across states: Right now we know about undercounting only at a national level. In the coming months, the U.S. Census Bureau will release estimates of undercounting at the state level. States with more extensive miscounting may be most vulnerable to the health consequences. 

Disparities within states: As government and public health agencies begin using data from the 2020 census, the impacts on health may become apparent in cities and municipalities where undercounting was most pronounced. These are likely to be communities with a higher proportion of Black, Hispanic, and American Indian residents, as well as college towns where students were required to move off campus at the start of the pandemic. Right now, for example, many states are closing COVID-19 testing and vaccination sites. If cases or hospitalizations were to spike again and sites were reopened based on 2020 census population estimates, places with substantial undercounting may see longer lines and other barriers to testing. 

Contested results: Tribal, state and local governments can contest limited aspects of the census results (e.g., errors related to geographic boundaries or housing placement) through the Census Count Question Resolution (CQR) operation through June 30, 2023. Later this year the Census Bureau is expected to release details of how governments can challenge population estimates through the Population Estimates Challenge Program. Multiple municipalities have already filed formal challenges

Potential policy changes: If states lose some Medicaid funding due to FMAP adjustment, they will have to account for the budget shortfall. Some may amend their Medicaid eligibility to reduce the number of enrollees, cut Medicaid benefits, or cut other programs that support residents’ health. Any of these measures would make health care less available for the most vulnerable residents in a state.  

Unfortunately, the ramifications of census miscounting have the potential to impact health for a long time. Staying vigilant to emerging stories and identifying the potential role of census miscounting in health data and policy decisions are critical to understanding, protecting and promoting the nation’s health.

Annemarie Hirsch is an epidemiologist and associate professor at Geisinger. Christine Gray is an epidemiologist and assistant research professor at Duke University. Drs. Hirsch and Gray are co-founders of HGEpidemiology, LLC.