Urban Wire Three Steps to Improving Data to Help Combat the Public Health Emergency of Structural Racism
Sonia Torres Rodríguez
Display Date

A nurse stares at the camera wearing a mask and face shield

Amid the COVID-19 pandemic that is disproportionately killing Black people, more than 20 cities, at least 3 states, the American Public Health Association, and the American Medical Association have declared racism a public health emergency. Public health data glaringly highlight these disparities. The maternal mortality rate is three times higher for Black women than for white women, a reality that could worsen under COVID-19. And the average life expectancy for Black people is four years lower than for white people, in part because of medical racism and police targeting of Black communities.

In their statements, jurisdictions note that such declarations encourage doctors, public officials, and communities to center racial health disparities in their collaborative work. But a surprising amount of national data, especially at granular levels, are not disaggregated by race or ethnicity. To tackle racism as a public health emergency, the field needs to address limitations in health data, such as morbidity and disease prevalence, and the social determinants of health, such as employment, wealth, education, and housing. Advocates, researchers, policymakers, health professionals, and others should work together to fill information gaps todisrupt how structural oppression operates. The first step to doing so is disaggregating data by race and ethnicity to enable access to key data to inform public policy decisions that can transform lives. Here are three ways to advance this goal.

1. Require race and ethnicity data in ongoing data collection

COVID-19 has widened racial and ethnic disparities, underscoring how violently and quickly anti-Blackness and structural racism can shorten Black and brown lives. Black, Indigenous, and Latinx people are two to three times more likely to die from COVID-19 and are more likely to experience adverse economic effects during the pandemic than white people.

Without cross-sector advocacy at the beginning of the pandemic, we would have been in the dark about the extent of the pandemic’s disparate impacts. In March and April, US researchers, doctors, and policymakers (PDF) suggested that an egregious lack of racial and ethnic data on COVID-19 cases and deaths was likely masking disproportionate case and death rates among communities of color. Thanks to their efforts, the Trump administration provided new guidance in June on COVID-19 case and death reporting, requiring all state and public health departments to report racial, ethnic, and other demographic data. These data, albeit incomplete, allowed organizations like Data for Black Lives and the Urban Institute’s Racial Equity Analytics Lab to produce maps and visualizations that validated targeted efforts to improve COVID-19 outcomes in BIPOC communities.

The work is not done. Research shows that this guidance had less-than-perfect compliance in several states, suppressing reported racial and ethnic health disparities in COVID-19 cases and deaths. Expanding major racial identifiers to more closely match our population’s diversity would help better serve populations’ needs. For example, the racial category of “Asian American and Pacific Islander” contains more than 27 ethnicities.

2. Advocate for new research questions about racism

Thanks to the momentum built by declarations of racism as a public health crisis, Mariko Toyoji, an epidemiologist at Public Health Seattle and King County’s Assessment, Policy Development, and Evaluation Unit, was a part of a recent successful campaign to include new questions about experiences with racial discrimination in the Washington State Behavioral Risk Factor Surveillance System module. Toyoji helped build a coalition of county health jurisdictions united by their interest in better understanding racial discrimination in their state. “We wanted to add the questions so we could have an idea of the prevalence of racial discrimination and link health disparities and outcomes, whether physical and mental, back to the experiences of discrimination among BIPOC,” Toyoji said. Answers to these questions will help inform funding and resources in local health equity and social justice initiatives.

3. Advance ethical use of data

When collecting demographic data is not practical in the short term, researchers can adopt alternative estimation methods, such as data imputation, or the practice of replacing missing values with predicted ones. This method requires statistical estimation (PDF) and may not be accessible to all data practitioners. But it can help close reporting gaps. Nearly a quarter of confirmed COVID-19 cases nationally are missing race and ethnicity data, and a study by the Virginia Department of Health (PDF) estimated race and ethnicity for more than 85 percent of those missing values.

Urban’s Racial Equity Analytics Lab is exposing patterns of structural racism against BIPOC communities by imputing race and ethnicity onto credit bureau data. This will help policymakers and practitioners identify, monitor, and build tools that address racial disparities in wealth and financial well-being and dismantle institutions that perpetuate oppressive systems. Before jumping into the statistical methods, however, the team is assessing the ethical implications of imputation and related methods, exploring the benefits and potential harms to BIPOC and developing tools to help researchers and data scientists avoid these harms. This work will help answer new questions about the wealth and financial health gaps, especially post pandemic, and highlight the institutions that perpetuate oppressive systems.

All approaches to data analysis should apply an equity lens

These three approaches to data collection and analysis alone will not disrupt racial and ethnic health disparities. Absent a commitment to equity, they could even perpetuate racist outcomes. Careless data collection could cause harmful violations of informed consent, privacy, and confidentiality; machine learning used in data imputation could replicate stereotypes through algorithm bias; and reporting racial disparities could cause racist data interpretations, as we have seen with the blaming of Black communities for their disproportionate number of COVID-19 cases and deaths.

To ensure disaggregated data inform responses that mitigate, rather than further entrench, racial and ethnic disparities, researchers, policymakers, and health professionals should apply an equity lens to their data practice. This involves centering the people who bear the greatest risks, primarily BIPOC communities, in research design, implementation and policy advocacy. With a multipronged strategy on data collection, data advocacy, and the latest data science guided by equity principles, we can make significant progress on the data available to combat structural racism.

Research Areas Race and equity Health and health care
Tags COVID-19 Structural racism
Policy Centers Metropolitan Housing and Communities Policy Center
Research Methods Research methods and data analytics