Who Counts and Who Gets Counted? Health Equity in Infectious Disease Surveillance

Infectious disease surveillance has long relied on a biomedical paradigm of disease risk, centering the human host and microbial pathogen without ample consideration of the social environment in which they interact. However, the risk of exposure to infectious pathogens, the susceptibility to infection once exposed, and the resulting effects of infection are inextricably tied to the social positions that individuals occupy. Regardless of the disease under surveillance, an individual’s education level, residential neighborhood, occupation, race, ethnicity, and other proxies for social position provide essential information about disease risk. This is true across a range of infectious diseases, from those we routinely survey (e.g., influenza, tuberculosis, HIV) to emerging pathogens (e.g., SARS-CoV-2). Consequently, if health equity is not at the core of our surveillance activities, inequities will inevitably arise, persist, and widen over time.

Today, the COVID-19 pandemic is disproportionately burdening racial/ethnic minority groups as a direct consequence of historical and contemporary injustices rooted in the social process of racialization. As Douglas et al. (p. 1141) highlight in this issue of AJPH, inadequate reporting of cases and deaths by race/ethnicity continues to mask the true magnitude of inequities owing to COVID-19. In this editorial, we further contextualize their findings by highlighting five shortcomings of current national surveillance activities that have hindered efforts to address racial/ethnic inequities in the impact of COVID-19:

1. Nearly a year into the pandemic, we still lack a funded and enforced federal mandate to report data on race/ethnicity.

2. We lack information about how data on race/ethnicity are collected across contexts.

3. We lack data on the full impact of the pandemic on population health disparities, beyond what can be captured in data on confirmed cases, hospitalizations, or deaths from COVID-19.

4. Suppressing or collapsing data across groups renders smaller racial/ethnic groups invisible.

5. Acknowledgment of the broader social and historical context is often missing from the analysis and interpretation of racial data.

FEDERAL MANDATE TO REPORT RACE/ETHNICITY DATA

The thoughtful collection of data on race/ethnicity has long been recognized as an important part of infectious disease surveillance.¹ Yet, not until June 4, 2020, more than four months after the first documented US case of COVID-19, was there a federal mandate to systematically collect data on race/ethnicity for all reported cases of COVID-19. Unfortunately, the mandate has done little to improve reporting. Of the 15 381 721 cases reported on the Centers for Disease Control and Prevention’s (CDC’s) COVID Data Tracker between its inception on August 28, 2020 and February 4, 2021, 7 912 371 (51.4%) were missing data on race/ethnicity.² Although a federal mandate is a necessary first step to ensure that data are collected, follow-up action is needed to ensure that such a mandate is consistently implemented across contexts. Lack of clear federal guidance and support for implementation reflects a systematic undercounting, and therefore devaluing, of Black and Brown lives.

DIFFERENCES IN COLLECTING RACIAL/ETHNIC DATA

Although the CDC’s COVID-19 case report form includes fields for race and ethnicity, it is unclear how this information is collected. Self-report is often the preferred method for ascertaining race/ethnicity.³ It is unclear, however, if or how individuals are asked to report their race or ethnicity at testing centers or hospitals. In addition, data on cases, hospitalizations, and deaths may be captured differently across different surveillance systems, with some relying on doctors or medical examiners to assign race/ethnicity and others relying on self-report. To add to this complexity, the racial/ethnic identities reported by health care providers or medical examiners may not match the racial/ethnic identities that are self-reported as part of the US Census. Differential classification of race/ethnicity across data sources may lead to under- or overestimates of disease risk. For example, Indigenous individuals are often misclassified as White in surveillance data, potentially leading to underestimation of the burden of COVID-19 in this population.⁴

POPULATION HEALTH DISPARITIES DATA

SARS-CoV-2 infections are likely to be underdetected in communities that face structural barriers to testing. This is particularly problematic because undetected infections may have long-term health effects. Moreover, our current surveillance activities have focused almost exclusively on the direct effects of the pandemic on population health, measured in terms of SARS-CoV-2 infections or deaths that can be directly attributed to COVID-19. The indirect effects of the pandemic on population health—through mechanisms such as social isolation, job loss, food insecurity, and delayed medical care—have received far less attention. Inequities in the indirect effects of the pandemic are likely to be substantial and will continue to play out long after infections and hospitalizations wane.^5,6

SUPPRESSING OR COLLAPSING DATA

The practice of suppressing small numbers of cases or deaths makes it difficult to investigate trends in smaller racial/ethnic groups or across multiple axes of identity. Collapsing smaller groups into an “other” category, which often includes cases or deaths missing information on race/ethnicity, does little to address the problem. When determining whether and how to release granular surveillance data, we must balance considerations of individual privacy with considerations of justice. Surveillance data play a central role in guiding the equitable allocation of resources, so the lack of sufficiently granular data inhibits efforts to achieve health equity for smaller groups and multiply marginalized individuals.

ACKNOWLEDGING SOCIAL AND HISTORICAL CONTEXT

The methodological choices we make when analyzing data can profoundly affect the conclusions we draw about the existence, direction, and magnitude of health inequities. Moreover, such choices are not purely objective and value-free; rather, they reflect one’s view of the world and judgments about what sources of variation in health status are permissible. Analyses that seek to advance health equity must acknowledge and make explicit the assumptions and values that guide methodological decisions. For example, statistical adjustment for covariates such as age and geography when comparing disease risk across racial/ethnic groups reflects the belief that different distributions of age or geography are not important components of racial disparities in disease risk.⁷ By contrast, an analytic approach that seeks to understand how racial health inequities are produced might stratify on age and place to assess the roles of age and geography—population characteristics that are themselves shaped by structural racism—in determining the distribution of disease across population groups. Because the analysis and interpretation of surveillance data have real consequences for the subsequent implementation of public health interventions, it is critical that analyses be grounded in an antiracist approach that acknowledges the role of social and historical forces in shaping the distribution of disease.

CONCLUSIONS

Existing inequities in the impact of COVID-19 are likely to be exacerbated by the roll-out of vaccines. Yet, inadequate reporting of race/ethnicity among those vaccinated continues to hamper efforts to measure and alleviate inequities. Of the nearly 13 million individuals who received a first dose of a COVID-19 vaccine between December 14, 2020 and January 14, 2021, data on race and ethnicity were available for only 52%. Among those with available data, 11.5% were classified as Hispanic/Latino and 5.4% non-Hispanic Black, despite these groups comprising 21% and 12% of all COVID-19 deaths, respectively.⁸

Continued gaps in the reporting of inequities in COVID-19 cases, deaths, and now vaccinations are a stark reminder of the structural changes needed in infectious disease surveillance. On the surface, decisions about what data to collect with regard to a nationally notifiable disease may seem purely biomedical in nature, but in practice such decisions reflect underlying power structures that determine who gets counted and for what purpose. It has perhaps never been more urgent for public health researchers and practitioners to reckon with how our surveillance systems, although powerful tools to improve population health, may insidiously promote inequities in health and perpetuate the notion that not all lives count equally.