Your independent source for Harvard news since 1898 | SUBSCRIBE

Your independent source for Harvard news since 1898

Right Now | COVID’s Collateral Color Line

How the Pandemic Killed the Uninfected

May-June 2022

Illustration showing different quality masks with different levels of exposure to pathogens indicated by red dots on the masks

Illustration by Giovanni Da Rae


Illustration by Giovanni Da Rae

When the COVID-19 pandemic laid bare health disparities among racial and ethnic groups in the United States, it appeared that a persistent public-health problem would finally get the increased attention it deserves. Population data from early in the pandemic revealed that black and Latino patients were more likely than whites to get infected and die from COVID. Such findings “were essential for setting the policy agenda and bringing health equity to the forefront,” explains Zirui Song, associate professor of health care policy at Harvard Medical School and a primary-care physician at Massachusetts General Hospital (MGH). “This raised awareness among clinicians, hospitals, and the public, which motivated efforts to address racial and ethnic disparities.”

For hospitalized non-COVID patients, “health disparities…worsened during the pandemic….” 

Yet a recent study coauthored by Song offers a more detailed picture of how COVID has affected racial and ethnic minority patients, revealing that health disparities have worsened during the pandemic for such patients receiving hospital treatment for any condition.

A health economist with expertise in insurance and provider payment, Song and colleagues analyzed claims data from traditional Medicare patients admitted to a hospital between January 2019 and February 2021: more than 14 million hospitalizations. (Medicare Advantage patients, whose health care is provided by private companies, were not included.) This offered the researchers an especially large sample of people, most of whom were 65 or older, a population more at risk for the negative effects of COVID, including death. The team found that in the first year of the pandemic, Hispanic patients averaged a 22 percent chance of dying when admitted to the hospital for COVID, compared with an average 17 percent mortality rate for hospitalized black and white COVID patients.

The researchers also examined how patients fared if they were hospitalized for non-COVID reasons during that time, comparing their mortality rates with those of patients who had pre-pandemic hospital stays. There was already a race and ethnicity gap in hospital mortality before the pandemic, Song says, but once COVID took hold, “We found that the gap actually widened,” driven by the differences in health outcomes between black and white patients.

In their analysis, he and his colleagues adjusted for factors such as age, sex, and disability, but chose a different tack with social factors such as household income and education. Such characteristics often are viewed by researchers as “confounding factors” that could cloud the results. Yet research also shows that household income and education can be affected by race and ethnicity, which in turn shape health outcomes. So the team left those “mediating” factors unadjusted with the hope that they could capture more accurately how race and ethnicity might contribute to hospital mortality.

Song had a particular interest in the health outcomes of non-COVID patients. He had seen MGH and other hospitals race to reallocate staff, supplies, and beds to care for the surge of COVID patients, affecting resources for other types of care. “While COVID rightfully grabbed many people’s attention and was at the top of the priority list in terms of hospital response across the country,” he explains, about 90 percent of U.S. hospitalizations were still for non-COVID illnesses. Accordingly, he wondered about “spillover effects”: the unintended consequences for non-COVID patients during this chaotic surge in COVID cases.

Patients who did not have COVID were less likely to be admitted to hospitals during this period of strained medical resources, and white patients were less likely to be hospitalized than minority patients relative to the pre-pandemic period. Notably, black patients admitted for non-COVID reasons were more likely than white patients to die in the hospital or shortly after they were discharged, relative to the pre-pandemic period. During the pandemic, this relative increase in the rate of death among black patients was almost 18 percent higher than their pre-pandemic mortality rate.

Song stresses that the study was not designed to isolate the mechanisms behind this increased mortality gap between white and minority patients, but he has multiple theories, including changes in the mix of patients and the quality of care. Perhaps the attention on COVID patients left “less resources and bandwidth for non-COVID patients,” he says. In that environment, “It’s not hard to imagine that clinical decisions may be based more on heuristics or that it might be more challenging to check our biases,” he adds, potentially widening racial and ethnic disparities. Song adds that future studies might consider the well-documented mental-health declines during the pandemic, which might have played a role in increasing mortality among minority patients. In time, researchers will accrue more data about the pandemic, including details of the more recent Delta and Omicron waves. Song expects future analyses to yield new information about the mechanisms that drive health disparities. Although he conducts mostly quantitative research, Song believes that qualitative work including efforts to interview groups of doctors and patients about their experiences in hospitals during the pandemic “might actually reveal more of the underlying story that a quantitative study can’t.”

In the meantime, “The awareness of racial and ethnic health disparities has undoubtedly been elevated by the pandemic,” Song says. He hopes this study will expand the way clinicians and policymakers think about the pandemic’s impact. “It’s a reminder,” he says, “that COVID can affect the care and the outcomes of patients who don’t have COVID.” 

You Might Also Like:

Illustration of a floating particle of the virus that causes Eastern Equine Encephalitis, the mosquito vector, and bird, horse, and human hosts

Illustration by Davide Bonazzi

Decoding the Alphviruses

Illustration of an eye superimposed on iconography from the national flag of China

Illustration by Taylor Callery

Authoritarian Regimes’ AI Innovation Advantage

black semi automatic pistol on gray surface

A new study of gun injuries quantifies the impacts on survivors—and their families.

Photograph by Unsplash

Gun Violence

You Might Also Like:

Illustration of a floating particle of the virus that causes Eastern Equine Encephalitis, the mosquito vector, and bird, horse, and human hosts

Illustration by Davide Bonazzi

Decoding the Alphviruses

Illustration of an eye superimposed on iconography from the national flag of China

Illustration by Taylor Callery

Authoritarian Regimes’ AI Innovation Advantage

black semi automatic pistol on gray surface

A new study of gun injuries quantifies the impacts on survivors—and their families.

Photograph by Unsplash

Gun Violence