student essay the disproportional impact of covid

STUDENT ESSAY The Disproportional Impact of COVID-19 on African Americans

Volume 22/2, December 2020, pp 299-307

Maritza Vasquez Reyes

Introduction

We all have been affected by the current COVID-19 pandemic. However, the impact of the pandemic and its consequences are felt differently depending on our status as individuals and as members of society. While some try to adapt to working online, homeschooling their children and ordering food via Instacart, others have no choice but to be exposed to the virus while keeping society functioning. Our different social identities and the social groups we belong to determine our inclusion within society and, by extension, our vulnerability to epidemics.

COVID-19 is killing people on a large scale. As of October 10, 2020, more than 7.7 million people across every state in the United States and its four territories had tested positive for COVID-19. According to the New York Times database, at least 213,876 people with the virus have died in the United States. [1] However, these alarming numbers give us only half of the picture; a closer look at data by different social identities (such as class, gender, age, race, and medical history) shows that minorities have been disproportionally affected by the pandemic. These minorities in the United States are not having their right to health fulfilled.

According to the World Health Organization’s report Closing the Gap in a Generation: Health Equity through Action on the Social Determinants of Health , “poor and unequal living conditions are the consequences of deeper structural conditions that together fashion the way societies are organized—poor social policies and programs, unfair economic arrangements, and bad politics.” [2] This toxic combination of factors as they play out during this time of crisis, and as early news on the effect of the COVID-19 pandemic pointed out, is disproportionately affecting African American communities in the United States. I recognize that the pandemic has had and is having devastating effects on other minorities as well, but space does not permit this essay to explore the impact on other minority groups.

Employing a human rights lens in this analysis helps us translate needs and social problems into rights, focusing our attention on the broader sociopolitical structural context as the cause of the social problems. Human rights highlight the inherent dignity and worth of all people, who are the primary rights-holders. [3] Governments (and other social actors, such as corporations) are the duty-bearers, and as such have the obligation to respect, protect, and fulfill human rights. [4] Human rights cannot be separated from the societal contexts in which they are recognized, claimed, enforced, and fulfilled. Specifically, social rights, which include the right to health, can become important tools for advancing people’s citizenship and enhancing their ability to participate as active members of society. [5] Such an understanding of social rights calls our attention to the concept of equality, which requires that we place a greater emphasis on “solidarity” and the “collective.” [6] Furthermore, in order to generate equality, solidarity, and social integration, the fulfillment of social rights is not optional. [7] In order to fulfill social integration, social policies need to reflect a commitment to respect and protect the most vulnerable individuals and to create the conditions for the fulfillment of economic and social rights for all.

Disproportional impact of COVID-19 on African Americans

As noted by Samuel Dickman et al.:

economic inequality in the US has been increasing for decades and is now among the highest in developed countries … As economic inequality in the US has deepened, so too has inequality in health. Both overall and government health spending are higher in the US than in other countries, yet inadequate insurance coverage, high-cost sharing by patients, and geographical barriers restrict access to care for many. [8]

For instance, according to the Kaiser Family Foundation, in 2018, 11.7% of African Americans in the United States had no health insurance, compared to 7.5% of whites. [9]

Prior to the Affordable Care Act—enacted into law in 2010—about 20% of African Americans were uninsured. This act helped lower the uninsured rate among nonelderly African Americans by more than one-third between 2013 and 2016, from 18.9% to 11.7%. However, even after the law’s passage, African Americans have higher uninsured rates than whites (7.5%) and Asian Americans (6.3%). [10] The uninsured are far more likely than the insured to forgo needed medical visits, tests, treatments, and medications because of cost.

As the COVID-19 virus made its way throughout the United States, testing kits were distributed equally among labs across the 50 states, without consideration of population density or actual needs for testing in those states. An opportunity to stop the spread of the virus during its early stages was missed, with serious consequences for many Americans. Although there is a dearth of race-disaggregated data on the number of people tested, the data that are available highlight African Americans’ overall lack of access to testing. For example, in Kansas, as of June 27, according to the COVID Racial Data Tracker, out of 94,780 tests, only 4,854 were from black Americans and 50,070 were from whites. However, blacks make up almost a third of the state’s COVID-19 deaths (59 of 208). And while in Illinois the total numbers of confirmed cases among blacks and whites were almost even, the test numbers show a different picture: 220,968 whites were tested, compared to only 78,650 blacks. [11]

Similarly, American Public Media reported on the COVID-19 mortality rate by race/ethnicity through July 21, 2020, including Washington, DC, and 45 states (see figure 1). These data, while showing an alarming death rate for all races, demonstrate how minorities are hit harder and how, among minority groups, the African American population in many states bears the brunt of the pandemic’s health impact.

Approximately 97.9 out of every 100,000 African Americans have died from COVID-19, a mortality rate that is a third higher than that for Latinos (64.7 per 100,000), and more than double than that for whites (46.6 per 100,000) and Asians (40.4 per 100,000). The overrepresentation of African Americans among confirmed COVID-19 cases and number of deaths underscores the fact that the coronavirus pandemic, far from being an equalizer, is amplifying or even worsening existing social inequalities tied to race, class, and access to the health care system.

Considering how African Americans and other minorities are overrepresented among those getting infected and dying from COVID-19, experts recommend that more testing be done in minority communities and that more medical services be provided. [12] Although the law requires insurers to cover testing for patients who go to their doctor’s office or who visit urgent care or emergency rooms, patients are fearful of ending up with a bill if their visit does not result in a COVID test. Furthermore, minority patients who lack insurance or are underinsured are less likely to be tested for COVID-19, even when experiencing alarming symptoms. These inequitable outcomes suggest the importance of increasing the number of testing centers and contact tracing in communities where African Americans and other minorities reside; providing testing beyond symptomatic individuals; ensuring that high-risk communities receive more health care workers; strengthening social provision programs to address the immediate needs of this population (such as food security, housing, and access to medicines); and providing financial protection for currently uninsured workers.

Social determinants of health and the pandemic’s impact on African Americans’ health outcomes

In international human rights law, the right to health is a claim to a set of social arrangements—norms, institutions, laws, and enabling environment—that can best secure the enjoyment of this right. The International Covenant on Economic, Social and Cultural Rights sets out the core provision relating to the right to health under international law (article 12). [13] The United Nations Committee on Economic, Social and Cultural Rights is the body responsible for interpreting the covenant. [14] In 2000, the committee adopted a general comment on the right to health recognizing that the right to health is closely related to and dependent on the realization of other human rights. [15] In addition, this general comment interprets the right to health as an inclusive right extending not only to timely and appropriate health care but also to the determinants of health. [16] I will reflect on four determinants of health—racism and discrimination, poverty, residential segregation, and underlying medical conditions—that have a significant impact on the health outcomes of African Americans.

Racism and discrimination

In spite of growing interest in understanding the association between the social determinants of health and health outcomes, for a long time many academics, policy makers, elected officials, and others were reluctant to identify racism as one of the root causes of racial health inequities. [17] To date, many of the studies conducted to investigate the effect of racism on health have focused mainly on interpersonal racial and ethnic discrimination, with comparatively less emphasis on investigating the health outcomes of structural racism. [18] The latter involves interconnected institutions whose linkages are historically rooted and culturally reinforced. [19] In the context of the COVID-19 pandemic, acts of discrimination are taking place in a variety of contexts (for example, social, political, and historical). In some ways, the pandemic has exposed existing racism and discrimination.

Poverty (low-wage jobs, insurance coverage, homelessness, and jails and prisons)

Data drawn from the 2018 Current Population Survey to assess the characteristics of low-income families by race and ethnicity shows that of the 7.5 million low-income families with children in the United States, 20.8% were black or African American (while their percentage of the population in 2018 was only 13.4%). [20] Low-income racial and ethnic minorities tend to live in densely populated areas and multigenerational households. These living conditions make it difficult for low-income families to take necessary precautions for their safety and the safety of their loved ones on a regular basis. [21] This fact becomes even more crucial during a pandemic.

Low-wage jobs: The types of work where people in some racial and ethnic groups are overrepresented can also contribute to their risk of getting sick with COVID-19. Nearly 40% of African American workers, more than seven million, are low-wage workers and have jobs that deny them even a single paid sick day. Workers without paid sick leave might be more likely to continue to work even when they are sick. [22] This can increase workers’ exposure to other workers who may be infected with the COVID-19 virus.

Similarly, the Centers for Disease Control has noted that many African Americans who hold low-wage but essential jobs (such as food service, public transit, and health care) are required to continue to interact with the public, despite outbreaks in their communities, which exposes them to higher risks of COVID-19 infection. According to the Centers for Disease Control, nearly a quarter of employed Hispanic and black or African American workers are employed in service industry jobs, compared to 16% of non-Hispanic whites. Blacks or African Americans make up 12% of all employed workers but account for 30% of licensed practical and licensed vocational nurses, who face significant exposure to the coronavirus. [23]

In 2018, 45% of low-wage workers relied on an employer for health insurance. This situation forces low-wage workers to continue to go to work even when they are not feeling well. Some employers allow their workers to be absent only when they test positive for COVID-19. Given the way the virus spreads, by the time a person knows they are infected, they have likely already infected many others in close contact with them both at home and at work. [24]

Homelessness : Staying home is not an option for the homeless. African Americans, despite making up just 13% of the US population, account for about 40% of the nation’s homeless population, according to the Annual Homeless Assessment Report to Congress. [25] Given that people experiencing homelessness often live in close quarters, have compromised immune systems, and are aging, they are exceptionally vulnerable to communicable diseases—including the coronavirus that causes COVID-19.

Jails and prisons : Nearly 2.2 million people are in US jails and prisons, the highest rate in the world. According to the US Bureau of Justice, in 2018, the imprisonment rate among black men was 5.8 times that of white men, while the imprisonment rate among black women was 1.8 times the rate among white women. [26] This overrepresentation of African Americans in US jails and prisons is another indicator of the social and economic inequality affecting this population.

According to the Committee on Economic, Social and Cultural Rights’ General Comment 14, “states are under the obligation to respect the right to health by, inter alia , refraining from denying or limiting equal access for all persons—including prisoners or detainees, minorities, asylum seekers and illegal immigrants—to preventive, curative, and palliative health services.” [27] Moreover, “states have an obligation to ensure medical care for prisoners at least equivalent to that available to the general population.” [28] However, there has been a very limited response to preventing transmission of the virus within detention facilities, which cannot achieve the physical distancing needed to effectively prevent the spread of COVID-19. [29]

Residential segregation

Segregation affects people’s access to healthy foods and green space. It can also increase excess exposure to pollution and environmental hazards, which in turn increases the risk for diabetes and heart and kidney diseases. [30] African Americans living in impoverished, segregated neighborhoods may live farther away from grocery stores, hospitals, and other medical facilities. [31] These and other social and economic inequalities, more so than any genetic or biological predisposition, have also led to higher rates of African Americans contracting the coronavirus. To this effect, sociologist Robert Sampson states that the coronavirus is exposing class and race-based vulnerabilities. He refers to this factor as “toxic inequality,” especially the clustering of COVID-19 cases by community, and reminds us that African Americans, even if they are at the same level of income or poverty as white Americans or Latino Americans, are much more likely to live in neighborhoods that have concentrated poverty, polluted environments, lead exposure, higher rates of incarceration, and higher rates of violence. [32]

Many of these factors lead to long-term health consequences. The pandemic is concentrating in urban areas with high population density, which are, for the most part, neighborhoods where marginalized and minority individuals live. In times of COVID-19, these concentrations place a high burden on the residents and on already stressed hospitals in these regions. Strategies most recommended to control the spread of COVID-19—social distancing and frequent hand washing—are not always practical for those who are incarcerated or for the millions who live in highly dense communities with precarious or insecure housing, poor sanitation, and limited access to clean water.

Underlying health conditions

African Americans have historically been disproportionately diagnosed with chronic diseases such as asthma, hypertension and diabetes—underlying conditions that may make COVID-19 more lethal. Perhaps there has never been a pandemic that has brought these disparities so vividly into focus.

Doctor Anthony Fauci, an immunologist who has been the director of the National Institute of Allergy and Infectious Diseases since 1984, has noted that “it is not that [African Americans] are getting infected more often. It’s that when they do get infected, their underlying medical conditions … wind them up in the ICU and ultimately give them a higher death rate.” [33]

One of the highest risk factors for COVID-19-related death among African Americans is hypertension. A recent study by Khansa Ahmad et al. analyzed the correlation between poverty and cardiovascular diseases, an indicator of why so many black lives are lost in the current health crisis. The authors note that the American health care system has not yet been able to address the higher propensity of lower socioeconomic classes to suffer from cardiovascular disease. [34] Besides having higher prevalence of chronic conditions compared to whites, African Americans experience higher death rates. These trends existed prior to COVID-19, but this pandemic has made them more visible and worrisome.

Addressing the impact of COVID-19 on African Americans: A human rights-based approach

The racially disparate death rate and socioeconomic impact of the COVID-19 pandemic and the discriminatory enforcement of pandemic-related restrictions stand in stark contrast to the United States’ commitment to eliminate all forms of racial discrimination. In 1965, the United States signed the International Convention on the Elimination of All Forms of Racial Discrimination, which it ratified in 1994. Article 2 of the convention contains fundamental obligations of state parties, which are further elaborated in articles 5, 6, and 7. [35] Article 2 of the convention stipulates that “each State Party shall take effective measures to review governmental, national and local policies, and to amend, rescind or nullify any laws and regulations which have the effect of creating or perpetuating racial discrimination wherever it exists” and that “each State Party shall prohibit and bring to an end, by all appropriate means, including legislation as required by circumstances, racial discrimination by any persons, group or organization.” [36]

Perhaps this crisis will not only greatly affect the health of our most vulnerable community members but also focus public attention on their rights and safety—or lack thereof. Disparate COVID-19 mortality rates among the African American population reflect longstanding inequalities rooted in systemic and pervasive problems in the United States (for example, racism and the inadequacy of the country’s health care system). As noted by Audrey Chapman, “the purpose of a human right is to frame public policies and private behaviors so as to protect and promote the human dignity and welfare of all members and groups within society, particularly those who are vulnerable and poor, and to effectively implement them.” [37] A deeper awareness of inequity and the role of social determinants demonstrates the importance of using right to health paradigms in response to the pandemic.

The Committee on Economic, Social and Cultural Rights has proposed some guidelines regarding states’ obligation to fulfill economic and social rights: availability, accessibility, acceptability, and quality. These four interrelated elements are essential to the right to health. They serve as a framework to evaluate states’ performance in relation to their obligation to fulfill these rights. In the context of this pandemic, it is worthwhile to raise the following questions: What can governments and nonstate actors do to avoid further marginalizing or stigmatizing this and other vulnerable populations? How can health justice and human rights-based approaches ground an effective response to the pandemic now and build a better world afterward? What can be done to ensure that responses to COVID-19 are respectful of the rights of African Americans? These questions demand targeted responses not just in treatment but also in prevention. The following are just some initial reflections:

First, we need to keep in mind that treating people with respect and human dignity is a fundamental obligation, and the first step in a health crisis. This includes the recognition of the inherent dignity of people, the right to self-determination, and equality for all individuals. A commitment to cure and prevent COVID-19 infections must be accompanied by a renewed commitment to restore justice and equity.

Second, we need to strike a balance between mitigation strategies and the protection of civil liberties, without destroying the economy and material supports of society, especially as they relate to minorities and vulnerable populations. As stated in the Siracusa Principles, “[state restrictions] are only justified when they support a legitimate aim and are: provided for by law, strictly necessary, proportionate, of limited duration, and subject to review against abusive applications.” [38] Therefore, decisions about individual and collective isolation and quarantine must follow standards of fair and equal treatment and avoid stigma and discrimination against individuals or groups. Vulnerable populations require direct consideration with regard to the development of policies that can also protect and secure their inalienable rights.

Third, long-term solutions require properly identifying and addressing the underlying obstacles to the fulfillment of the right to health, particularly as they affect the most vulnerable. For example, we need to design policies aimed at providing universal health coverage, paid family leave, and sick leave. We need to reduce food insecurity, provide housing, and ensure that our actions protect the climate. Moreover, we need to strengthen mental health and substance abuse services, since this pandemic is affecting people’s mental health and exacerbating ongoing issues with mental health and chemical dependency. As noted earlier, violations of the human rights principles of equality and nondiscrimination were already present in US society prior to the pandemic. However, the pandemic has caused “an unprecedented combination of adversities which presents a serious threat to the mental health of entire populations, and especially to groups in vulnerable situations.” [39] As Dainius Pūras has noted, “the best way to promote good mental health is to invest in protective environments in all settings.” [40] These actions should take place as we engage in thoughtful conversations that allow us to assess the situation, to plan and implement necessary interventions, and to evaluate their effectiveness.

Finally, it is important that we collect meaningful, systematic, and disaggregated data by race, age, gender, and class. Such data are useful not only for promoting public trust but for understanding the full impact of this pandemic and how different systems of inequality intersect, affecting the lived experiences of minority groups and beyond. It is also important that such data be made widely available, so as to enhance public awareness of the problem and inform interventions and public policies.

In 1966, Dr. Martin Luther King Jr. said, “Of all forms of inequality, injustice in health is the most shocking and inhuman.” [41] More than 54 years later, African Americans still suffer from injustices that are at the basis of income and health disparities. We know from previous experiences that epidemics place increased demands on scarce resources and enormous stress on social and economic systems.

A deeper understanding of the social determinants of health in the context of the current crisis, and of the role that these factors play in mediating the impact of the COVID-19 pandemic on African Americans’ health outcomes, increases our awareness of the indivisibility of all human rights and the collective dimension of the right to health. We need a more explicit equity agenda that encompasses both formal and substantive equality. [42] Besides nondiscrimination and equality, participation and accountability are equally crucial.

Unfortunately, as suggested by the limited available data, African American communities and other minorities in the United States are bearing the brunt of the current pandemic. The COVID-19 crisis has served to unmask higher vulnerabilities and exposure among people of color. A thorough reflection on how to close this gap needs to start immediately. Given that the COVID-19 pandemic is more than just a health crisis—it is disrupting and affecting every aspect of life (including family life, education, finances, and agricultural production)—it requires a multisectoral approach. We need to build stronger partnerships among the health care sector and other social and economic sectors. Working collaboratively to address the many interconnected issues that have emerged or become visible during this pandemic—particularly as they affect marginalized and vulnerable populations—offers a more effective strategy.

Moreover, as Delan Devakumar et al. have noted:

the strength of a healthcare system is inseparable from broader social systems that surround it. Health protection relies not only on a well-functioning health system with universal coverage, which the US could highly benefit from, but also on social inclusion, justice, and solidarity. In the absence of these factors, inequalities are magnified and scapegoating persists, with discrimination remaining long after. [43]

This current public health crisis demonstrates that we are all interconnected and that our well-being is contingent on that of others. A renewed and healthy society is possible only if governments and public authorities commit to reducing vulnerability and the impact of ill-health by taking steps to respect, protect, and fulfill the right to health. [44] It requires that government and nongovernment actors establish policies and programs that promote the right to health in practice. [45] It calls for a shared commitment to justice and equality for all.

Maritza Vasquez Reyes, MA, LCSW, CCM, is a PhD student and Research and Teaching Assistant at the UConn School of Social Work, University of Connecticut, Hartford, USA.

Please address correspondence to the author. Email: [email protected].

Competing interests: None declared.

Copyright © 2020 Vasquez Reyes. This is an open access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original author and source are credited.

[1] “Coronavirus in the U.S.: Latest map and case count,” New York Times (October 10, 2020). Available at https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html.

[2] World Health Organization Commission on the Social Determinants of Health, Closing the gap in a generation: Health equity through action on the social determinants of health (Geneva: World Health Organization, 2008), p. 1.

[3] S. Hertel and L. Minkler, Economic rights: Conceptual, measurement, and policy issues (New York: Cambridge University Press, 2007); S. Hertel and K. Libal, Human rights in the United States: Beyond exceptionalism (Cambridge: Cambridge University Press, 2011); D. Forsythe, Human rights in international relations , 2nd edition (Cambridge: Cambridge University Press, 2006).

[4] Danish Institute for Human Rights, National action plans on business and human rights (Copenhagen: Danish Institute for Human Rights, 2014).

[5] J. R. Blau and A. Moncada, Human rights: Beyond the liberal vision (Lanham, MD: Rowman and Littlefield, 2005).

[6] J. R. Blau. “Human rights: What the United States might learn from the rest of the world and, yes, from American sociology,” Sociological Forum 31/4 (2016), pp. 1126–1139; K. G. Young and A. Sen, The future of economic and social rights (New York: Cambridge University Press, 2019).

[7] Young and Sen (see note 6).

[8] S. Dickman, D. Himmelstein, and S. Woolhandler, “Inequality and the health-care system in the USA,” Lancet , 389/10077 (2017), p. 1431.

[9] S. Artega, K. Orgera, and A. Damico, “Changes in health insurance coverage and health status by race and ethnicity, 2010–2018 since the ACA,” KFF (March 5, 2020). Available at https://www.kff.org/disparities-policy/issue-brief/changes-in-health-coverage-by-race-and-ethnicity-since-the-aca-2010-2018/.

[10] H. Sohn, “Racial and ethnic disparities in health insurance coverage: Dynamics of gaining and losing coverage over the life-course,” Population Research and Policy Review 36/2 (2017), pp. 181–201.

[11] Atlantic Monthly Group, COVID tracking project . Available at https://covidtracking.com . 

[12] “Why the African American community is being hit hard by COVID-19,” Healthline (April 13, 2020). Available at https://www.healthline.com/health-news/covid-19-affecting-people-of-color#What-can-be-done?.

[13] World Health Organization, 25 questions and answers on health and human rights (Albany: World Health Organization, 2002).

[14] Ibid; Hertel and Libal (see note 3).

[17] Z. Bailey, N. Krieger, M. Agénor et al., “Structural racism and health inequities in the USA: Evidence and interventions,” Lancet 389/10077 (2017), pp. 1453–1463.

[20] US Census. Available at https://www.census.gov/library/publications/2019/demo/p60-266.html.

[21] M. Simms, K. Fortuny, and E. Henderson, Racial and ethnic disparities among low-income families (Washington, D.C.: Urban Institute Publications, 2009).

[23] Centers for Disease Control and Prevention, Health Equity Considerations and Racial and Ethnic Minority Groups (2020). Available at https://www.cdc.gov/coronavirus/2019-ncov/community/health-equity/race-ethnicity.html.

[24] Artega et al. (see note 9).

[25] K. Allen, “More than 50% of homeless families are black, government report finds,” ABC News (January 22, 2020). Available at https://abcnews.go.com/US/50-homeless-families-black-government-report-finds/story?id=68433643.

[26] A. Carson, Prisoners in 2018 (US Department of Justice, 2020). Available at https://www.bjs.gov/content/pub/pdf/p18.pdf.

[27] United Nations Committee on Economic, Social and Cultural Rights, General Comment No. 14, The Right to the Highest Attainable Standard of Health, UN Doc. E/C.12/2000/4 (2000).

[28] J. J. Amon, “COVID-19 and detention,” Health and Human Rights 22/1 (2020), pp. 367–370.

[30] L. Pirtle and N. Whitney, “Racial capitalism: A fundamental cause of novel coronavirus (COVID-19) pandemic inequities in the United States,” Health Education and Behavior 47/4 (2020), pp. 504–508.

[31] Ibid; R. Sampson, “The neighborhood context of well-being,” Perspectives in Biology and Medicine 46/3 (2003), pp. S53–S64.

[32] C. Walsh, “Covid-19 targets communities of color,” Harvard Gazette (April 14, 2020). Available at https://news.harvard.edu/gazette/story/2020/04/health-care-disparities-in-the-age-of-coronavirus/.

[33] B. Lovelace Jr., “White House officials worry the coronavirus is hitting African Americans worse than others,” CNBC News (April 7, 2020). Available at https://www.cnbc.com/2020/04/07/white-house-officials-worry-the-coronavirus-is-hitting-african-americans-worse-than-others.html.

[34] K. Ahmad, E. W. Chen, U. Nazir, et al., “Regional variation in the association of poverty and heart failure mortality in the 3135 counties of the United States,” Journal of the American Heart Association 8/18 (2019).

[35] D. Desierto, “We can’t breathe: UN OHCHR experts issue joint statement and call for reparations” (EJIL Talk), Blog of the European Journal of International Law (June 5, 2020). Available at https://www.ejiltalk.org/we-cant-breathe-un-ohchr-experts-issue-joint-statement-and-call-for-reparations/.

[36] International Convention on the Elimination of All Forms of Racial Discrimination, G. A. Res. 2106 (XX) (1965), art. 2.

[37] A. Chapman, Global health, human rights and the challenge of neoliberal policies (Cambridge: Cambridge University Press, 2016), p. 17.

[38] N. Sun, “Applying Siracusa: A call for a general comment on public health emergencies,” Health and Human Rights Journal (April 23, 2020).

[39] D. Pūras, “COVID-19 and mental health: Challenges ahead demand changes,” Health and Human Rights Journal (May 14, 2020).

[41] M. Luther King Jr, “Presentation at the Second National Convention of the Medical Committee for Human Rights,” Chicago, March 25, 1966.

[42] Chapman (see note 35).

[43] D. Devakumar, G. Shannon, S. Bhopal, and I. Abubakar, “Racism and discrimination in COVID-19 responses,” Lancet 395/10231 (2020), p. 1194.

[44] World Health Organization (see note 12).

Applications are open

• Skip to content
• Skip to search
• Accessibility Policy
• Report an Accessibility Issue

COVID-19 and the Disproportionate Impact on Black Americans

Q&A with Enrique Neblett

Professor of health behavior and health education.

July 1, 2020

Why is the coronavirus pandemic causing Black Americans to be disproportionately affected by COVID-19 and what can we do at the individual and community level to dismantle the systemic racism at the root of these health disparities? We spoke to Enrique Neblett, professor of Health Behavior and Health Education, to learn more. 

In what ways has the Black population in the United States been uniquely affected by COVID-19?

There are a few ways in which people have talked about how Black Americans have been affected by the pandemic.

One has to do with higher rates of hospitalization, infections and mortality rates. From very early on in the pandemic, we were appalled by data that showed Black Americans disproportionately represented for these outcomes relative to their percentage of the population. 

Milwaukee and Chicago are two examples that come to mind. African Americans in these cities are about a third of the population, but represented over 70% of the deaths. Similarly, in Georgia, African Americans make up a third of the population but represented 80% of hospitalizations. Unfortunately, we see the same things here in Michigan, where African Americans are roughly 14% of the population, yet they represent 33% of the cases, and 41% of deaths.

The second area is the psychological impact of the pandemic. The Black population is among a large group of the essential workers, and many have to make tough decisions with regard to staying home if they get sick or risking lost wages—or even unemployment—in addition to situations such as having to deal with the loss of family members or taking care of family who may be sick. A recent poll found that Black Americans are nearly three times as likely to personally know someone who has died from the virus than white Americans. In some cases, we’re talking about the emotional toll of having to make excruciating decisions about whether to risk getting sick or work while sick and other financial considerations, while also coping with premature and unexpected death and loss. 

Why is COVID-19 impacting the Black population disportionately? 

There are a complex set of factors that account for why the pandemic is disproportionately affecting Black Americans, but it is important that we name structural and systemic racism as drivers of COVID-19 disparities. There was speculation early on in the pandemic about chronic underlying conditions and how some are more likely to succumb to COVID-19 when they have diabetes or other underlying conditions. Unfortunately, we know that Black Americans are more likely to have high rates of cardiovascular disease and other chronic conditions than whites, in part, due to structural inequities in access to critical resources necessary to maintain health.

Another factor is occupational vulnerability. Black Americans are more likely than white Americans to hold jobs that are essential to the function of critical infrastructure. These are jobs that require continuous interaction with the public and, in some cases, don’t offer benefits such as paid vacation or the option to work from home.

Availability and access to testing is another important factor. In the initial stages of the pandemic, there were many places where testing was limited or unavailable, or there were significant delays in processing the test results. Lack of access to adequate testing and timely results can both be liabilities in getting urgent and needed medical care.

Poverty is another social determinant of health, structured by institutional and systemic racism, that has played a role in COVID-19 disparities. Lack of access to medical care to seek treatment, quality health insurance, healthy food, standard housing, and clean water are all factors that can indirectly contribute to heightened vulnerability to exposure and infection and lead to negative COVID-19 outcomes.

It is critical that we take a close look at how racism and longstanding structural inequities and practices—past and present—shape these factors and contribute to negative COVID-19 outcomes. 

If systemic racism is the root cause of COVID-19 related and other health disparities, how do we need to work together to end it?

There are several strategies that can be mobilized in working against systemic racism and, in turn, the impact of COVID-19 on Black Americans. A multi-pronged approach must inform the action steps that we can take as individuals and communities.

Listening to one another, self-educating, reading, and learning about systemic racism and how it operates are a great start. At the individual level, I’ve also seen people using their voices and privilege to raise awareness and propose concrete actions for eradicating racism by writing op-eds, letters, and making phone calls to lawmakers. Community groups and organizations possess valuable knowledge and expertise, represent critical assets, and are also well positioned to write letters and make calls.

Other strategies for mobilizing include investing in capacity building and helping communities to build their infrastructure in order to be able to respond to disasters like COVID-19. 

I've been really fortunate in my role as associate director of the Detroit Community-Academic Urban Research Center (Detroit URC) to discuss mobilization efforts with community partners in Detroit. It’s important that we work together to share resources and information with the residents who need them most. Also, it is important to remember that eradicating racism and promoting health equity will require the execution of concrete, specific and measurable actions that will lead to lasting systemic and structural change.

• Listen to Enrique Neblett on the Population Healthy podcast.
• Learn more about Health Behavior and Health Education.

Subscribe and listen to the Population Healthy podcast on Spreaker , Apple Podcasts , Spotify , YouTube , or wherever you listen to podcasts.

Sign up for the Population Healthy newsletter to receive the latest news from Michigan Public Health straight to your inbox.

Recent Posts

• April Zeoli appointed to the Michigan Gun Violence Prevention Task Force
• Gut reaction: Low levels of manganese can aggravate IBD
• College students' mental health improving, more finding support
• New federal rules to make mental health care more accessible, affordable

What We’re Talking About

• Adolescent Health
• Air Quality
• Alternative Therapies
• Alumni News and Networking
• Biostatistics
• Child Health
• Chronic Disease
• Community Partnership
• Computational Epidemiology and Systems Modeling
• Disaster Relief
• Diversity Equity and Inclusion
• Engaged Learning
• Environmental Health
• Epidemiologic Science
• Epidemiology
• Epigenetics
• First Generation Students
• Food Policy
• Food Safety
• General Epidemiology
• Global Health Epidemiology
• Global Public Health
• Health Behavior and Health Equity
• Health Care
• Health Care Access
• Health Care Management
• Health Care Policy
• Health Communication
• Health Disparities
• Health Informatics
• Health for Men
• Health for Women
• Heart Disease
• Hospital Administration
• Hospital and Molecular Epidemiology
• Immigration
• Industrial Hygiene
• Infectious Disease
• Internships
• LGBT Health
• Maternal Health
• Mental Health
• Mobile Health
• Occupational and Environmental Epidemiology
• Pain Management
• Pharmaceuticals
• Precision Health
• Professional Development
• Reproductive Health
• Research in Focus
• Scholarships
• Sexual Health
• Social Epidemiology
• Social Media
• Student Organizations
• Urban Health
• Urban Planning
• Value-Based Care
• Water Quality
• What Is Public Health?

Information For

• Prospective Students
• Current Students
• Alumni and Donors
• Community Partners and Employers
• Prospective Faculty (M-PACT)
• About Public Health
• How Do I Apply?
• Departments
• Findings magazine

Student Resources

• Career Development
• Certificates
• The Heights Intranet
• Update Contact Info
• Report Website Feedback

• Search Menu
• Sign in through your institution
• Advance articles
• Editor's Choice
• Supplement Archive
• Cover Archive
• IDSA Guidelines
• IDSA Journals
• The Journal of Infectious Diseases
• Open Forum Infectious Diseases
• Photo Quizzes
• State-of-the-Art Reviews
• Voices of ID
• Author Guidelines
• Open Access
• Why Publish
• IDSA Journals Calls for Papers
• Advertising and Corporate Services
• Advertising
• Journals Career Network
• Reprints and ePrints
• Sponsored Supplements
• Branded Books
• About Clinical Infectious Diseases
• About the Infectious Diseases Society of America
• About the HIV Medicine Association
• IDSA COI Policy
• Editorial Board
• Self-Archiving Policy
• For Reviewers
• For Press Offices
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

Covid-19 disparities through a biomedical lens, covid-19 disparities through social determinants of health lens, implications for the clinician.

• < Previous
• The Disproportionate Impact of COVID-19 on Racial and Ethnic Minorities in the United States

• Article contents
• Figures & tables
• Supplementary Data

Don Bambino Geno Tai, Aditya Shah, Chyke A Doubeni, Irene G Sia, Mark L Wieland, The Disproportionate Impact of COVID-19 on Racial and Ethnic Minorities in the United States, Clinical Infectious Diseases , Volume 72, Issue 4, 15 February 2021, Pages 703–706, https://doi.org/10.1093/cid/ciaa815

• Permissions Icon Permissions

The coronavirus disease 2019 (COVID-19) pandemic has disproportionately affected racial and ethnic minority groups, with high rates of death in African American, Native American, and LatinX communities. Although the mechanisms of these disparities are being investigated, they can be conceived as arising from biomedical factors as well as social determinants of health. Minority groups are disproportionately affected by chronic medical conditions and lower access to healthcare that may portend worse COVID-19 outcomes. Furthermore, minority communities are more likely to experience living and working conditions that predispose them to worse outcomes. Underpinning these disparities are long-standing structural and societal factors that the COVID-19 pandemic has exposed. Clinicians can partner with patients and communities to reduce the short-term impact of COVID-19 disparities while advocating for structural change.

( See the Editorial Commentary by Wilder on pages 707–9 .)

In the United States, coronavirus disease (COVID-19) caused by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) has disproportionately affected racial/ethnic minority and underserved groups, especially African American, LatinX, and Native American communities. The stark disparities underscore important medical, social, economic, environmental, and political contexts that predate the pandemic. In this paper, we describe potential factors contributing to COVID-19 disparities in the United States from both biomedical and social determinants of health perspectives. We also discuss the roles that clinicians can play in addressing these disparities.

As of June 2020, the Centers for Disease Control and Prevention (CDC) reported that 21.8% of COVID-19 cases in the United States were African Americans and 33.8% were LatinX, despite the fact that these groups comprise only 13% and 18% of the US population, respectively. These data are limited by underreporting of race (47% of cases) and ethnicity (43% of cases) [ 1 , 2 ]. In a report of hospitalized patients, 33% were African Americans, despite representing only 18% of a catchment area population [ 3 ]. The numbers have changed throughout time, but the disproportion has been consistent across the United States.

The mortality rate for COVID-19 among African Americans is more than 2-fold higher than whites. In the majority of the reporting states and Washington, DC, African Americans comprise a higher proportion of deaths relative to their percentage in those jurisdictions’ population [ 4 ]. The numbers have changed throughout time, but the signal of racial disproportion is clear.

In the hardest hit city in the United States, New York City, age-adjusted confirmed COVID-19 deaths were 220 and 236 per 100 000 for African American and LatinX patients, respectively. This is double compared to 110 and 102 per 100 000 for whites and Asians, respectively [ 5 ]. In Arizona, 13% of cases and 18% of deaths are Native Americans, who only make up 5.3% of the state’s population [ 6 , 7 ].

Models of COVID-19 mortality by race and ethnicity adjusted for medical complexity are not yet available, but a disproportionate burden of preexisting chronic medical problems likely contribute to disparities in COVID-19 outcomes. Preliminary data show that diabetes mellitus, hypertension, renal disease, and obesity increase a patient’s risk for severe COVID-19 disease and mortality [ 8 ]. African Americans have a disproportionately high prevalence of such comorbidities, including diabetes, hypertension, obesity, and coronary artery disease [ 9–11 ], and they are more likely to die prematurely compared to Whites due to all causes [ 12 ]. Therefore, these long-standing health disparities contribute to disproportionate deaths among African Americans with COVID-19.

The disproportionate burden of chronic medical conditions is compounded by lower access to healthcare among some racial and ethnic minority groups. Uninsured rates among nonelderly Americans are significantly higher for Native Americans (22%), Hispanics (19%), and African Americans (12%), compared with whites (8%) [ 13 ]. They also tend to live in areas where medical care is of poor quality or is underserved. Therefore, racial and ethnic minorities may receive lower quality care for COVID-19.

A wider societal lens is required to understand disparities in preexisting medical conditions, healthcare access, and other factors that may contribute to the disproportionate impact of COVID-19 on minority populations. Social determinants of health are the “conditions in the places where people live, learn, work, and play that affect a wide range of health risks and outcomes” [ 14 ].

Before the pandemic and associated economic fallout, poverty rates in the United States were 24% for Native Americans, 22% for African Americans, and 19% for Hispanics, compared to 9% for whites [ 15 ]. Across all income brackets, the median wealth of white households is 10 times the wealth of African American households [ 16 ]. Therefore, these minority groups have less financial capacity to make healthful decisions in the midst of the financial hardships that have accompanied the pandemic.

It is not surprising then that minority groups comprise a disproportionate percentage of workers in essential industries during the pandemic. Furthermore, only 20% of African American workers have the privilege of working from home compared to 30% of whites [ 17 ]. A report by New York City’s comptroller showed that 75% of frontline workers in the city are people of color. African Americans make up 40% of the transit workers. More than half use public transportation [ 18 ], and across the country, African Americans are more likely to use public transportation to commute to work compared to whites, 34% versus 14%, respectively [ 19 ]. These occupational hazards are compounded by the fact that only 55% of essential workers in the food service industry have access to paid sick leave [ 20 ]. These working conditions undoubtedly contribute to the disproportionate impact of COVID-19 on minority communities.

Living conditions in some minority communities further increase risk for SARS-CoV-2 infection and/or transmission. Communities with higher racial and ethnic minority populations have higher housing density, more housing insecurity, scarcity of potable water, and more multigenerational households that makes social distancing harder [ 21–23 ]. Likewise, there is often less access to healthy foods, which makes chronic disease management more difficult [ 24 ]. Furthermore, communities with higher minority populations are more likely to be targeted for marketing of unhealthful products like alcohol, cigarettes, and fast food that may negatively influence chronic medical conditions [ 25 ]. Finally, there is emerging evidence that air pollution, which is higher in minority communities, may play a role in COVID-19 severity [ 26 , 27 ].

During a pandemic, it is essential that credible, accurate health information is disseminated from health and healthcare institutions to the public in real time. Minority groups are more likely to have communication gaps due to issues of health literacy, socioeconomic disadvantage, and limited English language proficiency [ 28 ]. These gaps are exacerbated by justifiable mistrust of health institutions in some minority communities [ 29 ]. The result is a relative lack of credible COVID-19 information reaching marginalized communities, thereby elevating risk of disease contraction and transmission.

Both the disproportionate biomedical risk factors and social determinants that contribute to COVID-19 health disparities may be traced, in part, to a foundation of structural racism. The legacy of redlining and housing segregation, a policy that made African American communities pay more in mortgage but with less return in investment, is just one example of the ways in which these inequities were put in place by design [ 30 ]. Achieving health equity will require deconstruction of the legacy of structural racism (see Figure 1 ).

Pathogen, host, environment interplay in racial disparities in COVID-19. Abbreviations: COVID-19, coronavirus disease 2019; SARS-CoV-2, severe acute respiratory syndrome coronavirus-2.

As clinicians who are already overwhelmed by tending to the medical needs of COVID-19 patients and the challenges of resuming routine clinical care, it can be overwhelming to consider the root causes of disparities that often fall outside our realm of expertise. At the most basic level, clinicians should focus on providing high quality standard of care for all patients, regardless of background.

The CDC recommended interventions that address the disproportionate impact of COVID-19 to African American, LatinX, and Native American communities [ 31 ]. Standardized treatment and management protocols should be in place for all patients. Quality improvement endeavors must include health equity as a focus. A vital resource in health care facilities is adequately staffed and integrated support services (eg, navigation and professional interpreter service).

In addition, clinicians may advocate within their institutions and work units to foster sustainable partnerships with community-based organizations who serve vulnerable populations and to empower patients and communities by sharing resources. Community empowerment is especially critical in the context of a pandemic. Healthcare institutions are usually prominent member of the communities where they are located and should have a role in improving the communities’ health across all domains (physical, social, and psychological).

Implicit bias among clinicians and other healthcare workers can compound socioeconomic disadvantage for patients; contribute to poorer communication with patients, mistrust, and lower quality of care received; and potentially affect the outcomes of care [ 32 , 33 ]. Although much work is needed to understand the ways to effectively address implicit bias through educational or training programs, clinicians are encouraged to practice self-awareness on racial bias and stigma in the delivery of care [ 34 ]. Of particular relevance is evidence that equitable access to the same high-quality care can eliminate health disparities, but underserved and socioeconomically disadvantaged people need additional levels of support to remove structural barriers and benefit equitably from the same intervention strategy.

The COVID-19 crisis provides an opportunity for clinicians to collectively act on the root causes of these fundamental inequities that have been flagrantly demonstrated by the pandemic. Clinicians have a powerful voice at the health policy table. We can reflect to institutional and professional societies and national decision makers what we are seeing when caring for our patients. The disproportionate impacts of the pandemic on marginalized communities can be captured powerfully by stories of clinicians and their patients. Because many systemic inequities were put in motion by design, clinicians and other front-line healthcare workers can be important catalysts in designing a new more equitable system that promotes health for all Americans irrespective of social or economic background.

Potential conflicts of interest. The authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.

Centers for Disease Control and Prevention. Coronavirus Disease 2019 (COVID-19) in the US. Available at: https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/cases-in-us.html . Published 2020. Accessed 8 June 2020 .

US Census Bureau QuickFacts: United States. Census Bureau QuickFacts. Available at: https://www.census.gov/quickfacts/fact/table/US/PST045218 . Published 2018. Accessed 21 April 2020 .

Garg S , Kim L , Whitaker M , et al.  Hospitalization rates and characteristics of patients hospitalized with laboratory-confirmed coronavirus disease 2019—COVID-NET, 14 States, March 1–30, 2020 . MMWR Morb Mortal Wkly Rep 2020 ; 69 : 458 – 64 .

Google Scholar

Egbert A . COVID-19 deaths analyzed by race. APM Research Lab. Available at: https://www.apmresearchlab.org/covid/deaths-by-race . Published 11 May 2020. Accessed 8 June 2020 .

COVID-19: Data–NYC Health. Www1.nyc.gov. Available at: https://www1.nyc.gov/site/doh/covid/covid-19-data.page . Published 2020. Accessed 8 June 2020 .

AZDHS | COVID-19 Dashboards. Arizona department of health services. Available at: https://www.azdhs.gov/preparedness/epidemiology-disease-control/infectious-disease-epidemiology/covid-19/dashboards/index.php . Published 2020. Accessed 8 June 2020 .

US Census Bureau QuickFacts: Arizona. Census Bureau QuickFacts. Available at: https://www.census.gov/quickfacts/fact/table/AZ/PST045218 . Published 2018. Accessed 28 April 2020 .

Richardson S , Hirsch JS , Narasimhan M , et al.  Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in the New York city area . JAMA . Published online 22 April 2020 . doi: 10.1001/jama.2020.6775 .

Arias E , Xu JQ United States life tables, 2017: national vital statistics reports ; vol 68 no 7 . Hyattsville, MD : National Center for Health Statistics, 2019 .

Google Preview

Centers for Disease Control and Prevention. National diabetes statistics report, 2020 . Atlanta, GA : Centers for Disease Control and Prevention, US Dept of Health and Human Services ; 2020 .

National Center for Health Statistics. Health, United States, 2018 . Hyattsville, MD : National Center for Health Statistics, 2019 .

Cunningham TJ , Croft JB , Liu Y , Lu H , Eke PI , Giles WH . Vital signs: racial disparities in age-specific mortality among blacks or African Americans—United States, 1999–2015 . MMWR Morb Mortal Wkly Rep 2017 ; 66 : 444 – 56 .

Artiga S , Orgera K , Pham O . Disparities in health and health care: five key questions and answers. The Henry J. Kaiser Family Foundation. Available at: https://www.kff.org/disparities-policy/issue-brief/disparities-in-health-and-health-care-five-key-questions-and-answers/ . Published 2020. Accessed 28 April 2020 .

Social Determinants of Health | Healthy People 2020. Healthypeople.gov. Available at: https://www.healthypeople.gov/2020/topics-objectives/topic/social-determinants-of-health . Published 2020. Accessed 27 April 2020 .

Poverty Rate by Race/Ethnicity. The Henry J. Kaiser Family Foundation. Available at: https://www.kff.org/other/state-indicator/poverty-rate-by-raceethnicity . Published 2018. Accessed 28 April 2020 .

Kochhar R , Cilluffo A . How US wealth inequality has changed since great recession. Pew Research Center. Available at: https://www.pewresearch.org/fact-tank/2017/11/01/how-wealth-inequality-has-changed-in-the-u-s-since-the-great-recession-by-race-ethnicity-and-income/ . Published 2017. Accessed 22 April 2020 .

Bureau of Labor Statistics. The employment situation March 2020 . Washington, DC : Bureau of Labor Statistics ; 2020 .

New York City Comptroller. New York city’s frontline workers . New York : Bureau of Policy & Research ; 2020 .

Anderson M . Who relies on public transit in the US Pew Research Center. Available at: https://www.pewresearch.org/fact-tank/2016/04/07/who-relies-on-public-transit-in-the-u-s/ . Published 2016. Accessed 24 April 2020 .

Schneider D , Harknett K . Essential and vulnerable: service-sector workers and paid sick leave. Berkeley, CA: Shift Project; 2020. Available at: https://shift.berkeley.edu/files/2020/04/Essential_and_Vulnerable_Service_Sector_Workers_and_Paid_Sick_Leave.pdf . Accessed 28 April 2020 .

Taylor J . Racism, inequality, and health care for African Americans. The Century Foundation. Available at: https://tcf.org/content/report/racism-inequality-health-care-african-americans . Published 2019. Accessed 24 April 2020 .

Khunti K , Singh AK , Pareek M , Hanif W . Is ethnicity linked to incidence or outcomes of COVID-19? BMJ 2020 ; 369 : m1548 .

Branigin A . Black communities are on the “Frontline” of the COVID-19 pandemic: here’s why. The Root. Available at: https://www.theroot.com/black-communities-are-on-the-frontline-of-the-covid-19-1842404824 . Published 2020. Accessed 24 April 2020 .

Dutko P , Ver Ploeg M , Farrigan T Characteristics and influential factors of food deserts, ERR-140 . Washington, DC: US Department of Agriculture, Economic Research Service , August 2012.

Grier SA , Kumanyika S . Targeted marketing and public health . Annu Rev Public Health 2010 ; 31 : 349 – 69 .

Tessum C , Apte J , Goodkind A et al.  Inequity in consumption of goods and services adds to racial-ethnic disparities in air pollution exposure . Proc Nat Acad Sci USA 2019 ; 116 : 6001 – 6 .

Setti L , Passarini F , De Gennaro G et al.  SARS-Cov-2 RNA found on particulate matter of Bergamo in Northern Italy: first preliminary evidence. 2020 . doi: 10.1101/2020.04.15.20065995 .

Blumenshine P , Reingold A , Egerter S , Mockenhaupt R , Braveman P , Marks J . Pandemic influenza planning in the United States from a health disparities perspective . Emerg Infect Dis 2008 ; 14 : 709 – 15 .

Bergstresser SM . Health communication, public mistrust, and the politics of “rationality.” Am J Bioeth 2015 ; 15 : 57 – 9 .

Perry A . Black Americans were forced into “social distancing” long before the coronavirus. Brookings Institution. Available at: https://www.brookings.edu/blog/the-avenue/2020/03/20/black-americans-were-forced-into-social-distancing-long-before-the-coronavirus/ . Published 2020. Accessed 23 April 2020 .

Coronavirus Disease 2019 (COVID-19). Centers for Disease Control and Prevention. Available at: https://www.cdc.gov/coronavirus/2019-ncov/need-extra-precautions/racial-ethnic-minorities.html . Published 2020. Accessed 4 June 2020 .

FitzGerald C , Hurst S . Implicit bias in healthcare professionals: a systematic review . BMC Med Ethics 2017 ; 18 : 19 .

Hoffman KM , Trawalter S , Axt JR , Oliver MN . Racial bias in pain assessment and treatment recommendations, and false beliefs about biological differences between blacks and whites . Proc Natl Acad Sci U S A 2016 ; 113 : 4296 – 301 .

Hagiwara N , Kron FW , Scerbo MW , Watson GS . A call for grounding implicit bias training in clinical and translational frameworks . Lancet 2020 ; 395 : 1457 – 60 .

• minority groups
• african american
• health disparity
• native americans

Email alerts

More on this topic, related articles in pubmed, citing articles via, looking for your next opportunity.

• Recommend to your Library

Affiliations

• Online ISSN 1537-6591
• Print ISSN 1058-4838
• Copyright © 2024 Infectious Diseases Society of America
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

The PMC website is updating on October 15, 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• JAMA Network

Disparities in COVID-19 Outcomes by Race, Ethnicity, and Socioeconomic Status

Shruti magesh.

1 Division of Otolaryngology–Head and Neck Surgery, Department of Surgery, University of California, San Diego

2 Research Service, VA San Diego Healthcare System, San Diego, California

Daniel John

Aidan mattingly-app, sharad jain.

3 The University of California Davis School of Medicine, Sacramento

Eric Y. Chang

4 Department of Radiology, University of California, San Diego

5 Radiology Service, VA San Diego Healthcare System, San Diego, California

Weg M. Ongkeko

Accepted for Publication: September 12, 2021.

Published: November 11, 2021. doi:10.1001/jamanetworkopen.2021.34147

Open Access: This is an open access article distributed under the terms of the CC-BY License . © 2021 Magesh S et al. JAMA Network Open .

Author Contributions: Dr Ongkeko had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Ms Magesh, Mr John, and Mr W. Li contributed equally and are co–first authors.

Concept and design: Magesh, John, W. Li, Mattingly-app, Ongkeko.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Magesh, John, Mattingly-app, Ongkeko.

Critical revision of the manuscript for important intellectual content: Magesh, John, W. Li, Y. Li, Jain, Chang, Ongkeko.

Statistical analysis: Magesh, John, W. Li, Y. Li, Mattingly-app.

Obtained funding: Chang, Ongkeko.

Administrative, technical, or material support: W. Li, Y. Li, Jain, Ongkeko.

Supervision: W. Li, Ongkeko.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by grant R00RG2369 from the University of California, Office of the President/Tobacco-Related Disease Research Program Emergency COVID-19 Research Seed Funding to Dr Ongkeko.

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Associated Data

eTable 2. Study Summary Characteristics for Comorbidities

eTable 3. Adjustment of Relative Risk Ratios (RRs) for Additional Variables

eTable 4. Adjustment of Odds Ratios (ORs) for Additional Variables

eTable 5. Combined Prevalence of Cohort and Cross-sectional Studies

eTable 6. Summary of Q and I 2 Statistics for Study Variables

eFigure 1. PRISMA Workflow for Studies Included in Analysis

eFigure 2. Funnel Plots for Deceased Individuals in Cohort and Cross-sectional Studies

eFigure 3. Funnel Plots for Patients Admitted to ICU or Hospitalized in Cohort Studies

eFigure 4. Funnel Plots for COVID-19 Positive Patients in Cohort and Cross-sectional Studies

eFigure 5. Forest Plot for COVID-19 Positivity in Cohort and Cross-sectional Studies

eFigure 6. Forest Plot for Patients Admitted to ICU or Hospitalized in Cohort Studies

eFigure 7. Forest Plot for Deceased Individuals in Cohort and Cross-sectional Studies

eFigure 8. Metaregression for County Median Income

eFigure 9. Spearman Correlations for Measures of Clinical Care Quality

eFigure 10. Metaregression for Clinical Care Measures

eFigure 11. Leave-One-Out Sensitivity Analysis for Deceased Individuals in Cohort and Cross-sectional Studies

eFigure 12. Leave-One-Out Sensitivity Analysis for Patients Admitted to ICU or Hospitalized in Cohort Studies

eFigure 13. Leave-One-Out Sensitivity Analysis for COVID-19 Positive Patients in Cohort and Cross-sectional Studies

eFigure 14. Forest Plots for Deceased Patients After Removing Dominating Studies

eFigure 15. Forest Plots for Positive Individuals After Removing Dominating Studies

eMethods 1. Methods Pertaining to Search Criteria and Data Collection

eMethods 2. Citations of Articles that Appeared to Meet Inclusion Criteria but Were Excluded

eMethods 3. Description of Statistical Methods Used in Analyses

Are race and ethnicity–based COVID-19 outcome disparities in the United States associated with socioeconomic characteristics?

In this systematic review and meta-analysis of 4.3 million patients from 68 studies, African American, Hispanic, and Asian American individuals had a higher risk of COVID-19 positivity and ICU admission but lower mortality rates than White individuals. Socioeconomic disparity and clinical care quality were associated with COVID-19 mortality and incidence in racial and ethnic minority groups.

In this study, members of racial and ethnic minority groups had higher rates of COVID-19 positivity and disease severity than White populations; these findings are important for informing public health decisions, particularly for individuals living in socioeconomically deprived communities.

This systematic review and meta-analysis examines the association between race, ethnicity, COVID-19 outcomes, and socioeconomic determinants.

COVID-19 has disproportionately affected racial and ethnic minority groups, and race and ethnicity have been associated with disease severity. However, the association of socioeconomic determinants with racial disparities in COVID-19 outcomes remains unclear.

To evaluate the association of race and ethnicity with COVID-19 outcomes and to examine the association between race, ethnicity, COVID-19 outcomes, and socioeconomic determinants.

Data Sources

A systematic search of PubMed, medRxiv, bioRxiv, Embase, and the World Health Organization COVID-19 databases was performed for studies published from January 1, 2020, to January 6, 2021.

Study Selection

Studies that reported data on associations between race and ethnicity and COVID-19 positivity, disease severity, and socioeconomic status were included and screened by 2 independent reviewers. Studies that did not have a satisfactory quality score were excluded. Overall, less than 1% (0.47%) of initially identified studies met selection criteria.

Data Extraction and Synthesis

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed. Associations were assessed using adjusted and unadjusted risk ratios (RRs) and odds ratios (ORs), combined prevalence, and metaregression. Data were pooled using a random-effects model.

Main Outcomes and Measures

The main measures were RRs, ORs, and combined prevalence values.

A total of 4 318 929 patients from 68 studies were included in this meta-analysis. Overall, 370 933 patients (8.6%) were African American, 9082 (0.2%) were American Indian or Alaska Native, 101 793 (2.4%) were Asian American, 851 392 identified as Hispanic/Latino (19.7%), 7417 (0.2%) were Pacific Islander, 1 037 996 (24.0%) were White, and 269 040 (6.2%) identified as multiracial and another race or ethnicity. In age- and sex-adjusted analyses, African American individuals (RR, 3.54; 95% CI, 1.38-9.07; P  = .008) and Hispanic individuals (RR, 4.68; 95% CI, 1.28-17.20; P  = .02) were the most likely to test positive for COVID-19. Asian American individuals had the highest risk of intensive care unit admission (RR, 1.93; 95% CI, 1.60-2.34, P  < .001). The area deprivation index was positively correlated with mortality rates in Asian American and Hispanic individuals ( P  < .001). Decreased access to clinical care was positively correlated with COVID-19 positivity in Hispanic individuals ( P  < .001) and African American individuals ( P  < .001).

Conclusions and Relevance

In this study, members of racial and ethnic minority groups had higher risks of COVID-19 positivity and disease severity. Furthermore, socioeconomic determinants were strongly associated with COVID-19 outcomes in racial and ethnic minority populations.

Introduction

As of August 19, 2021, more than 209 million people across the world had been infected by COVID-19, with the United States accounting for more than 36 million cases and 618 000 deaths. 1 COVID-19 disproportionately affects racial and ethnic minority groups. 2 To reduce exposure and mortality rates, it is critical to identify the disparities associated with greater occurrences of COVID-19 among different populations. 3

In a meta-analysis of 50 articles, 4 it was shown that African American and Asian American patients were at a higher risk of intensive care unit (ICU) admission because of COVID-19 than White patients. A separate meta-analysis examining 45 articles 5 indicated that race may be associated with worse COVID-19 outcomes because of the increased occurrence of comorbidities in racial and ethnic minority groups. However, these studies did not examine the role of socioeconomic determinants, which disproportionately affect racial and ethnic minority populations. Another study 6 explored underlying factors for COVID-19 outcomes in racial and ethnic minority groups but did not integrate data from external sources, such as county median income. As such, current meta-analyses lack investigations assessing how socioeconomic determinants may be associated with COVID-19 disease severity in minority populations.

Individual cross-sectional and cohort studies have found that COVID-19 infection rates in racial and ethnic minority groups are associated with low socioeconomic status and income. 7 , 8 Specifically, studies have found that there is a positive association between COVID-19 risk and area deprivation index (ADI). 9 Past studies have also demonstrated that 11.7% of African Americans individuals are uninsured, compared with 7.5% of White individuals, thus potentially leading to more severe disease outcomes because of lack of access to medical care. 10 Geographic variation may also play a role in COVID-19 disease severity, as rural hospitals and communities often lack resources. 11 Therefore, it is plausible that these social determinants might be associated with COVID-19 disease severity in racial and ethnic minority populations.

In this study, we examine the associations of race and ethnicity with COVID-19 positivity rates, mortality, hospitalization, and ICU admission in the United States. We then associate these outcomes with various social determinants through adjusted and unadjusted relative risk ratio (RR) and odds ratio (OR) calculations and metaregression analysis. To our knowledge, we are the first to examine social determinants of health in racial disparities of COVID-19 outcomes through a systematic review and meta-analysis, which provides a more accurate understanding than results published in single-site studies.

Database Search and Inclusion Criteria

We conducted a systematic search of studies published from January 1, 2020, to January 6, 2021, in the PubMed, medRxiv, bioRxiv, Embase, and the World Health Organization COVID-19 databases. We used search terms pertaining to COVID-19 and disparities (eMethods 1 in the Supplement ) and only included studies that reported data on race and ethnicity as well as the following variables: socioeconomic status, COVID-19 positivity, hospitalization, ICU admission, mortality, and location/geography. All included studies were conducted in the United States.

Two independent reviewers (S.M. and D.J.) screened the titles, abstracts, and full text of each eligible study from the selected databases. Disagreements were resolved through discussion with a third reviewer (Y.L.). We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses ( PRISMA ) guideline for selection of papers in this meta-analysis (eFigure 1 and eMethods 2 in the Supplement ). The Joanna Briggs Institute critical appraisal tools were used to assess the quality of evidence from all studies in respect to study design. Studies were not included in our analysis if they scored lower than a 6 of 8 (75%) for cohort studies and 9 of 11 (82%) for cross-sectional studies. The complete search and inclusion strategy can be found in the eMethods 1 and 2 in the Supplement .

Data Extraction

Data were extracted from the 68 studies screened with the PRISMA guidelines. We collected details from studies regarding study setting and type and patient demographic characteristics, comorbidities, and outcomes (eMethods 1 in the Supplement ), using the same independent reviewer design as during study selection. Following the initial data review, socioeconomic variables quantifying disparities in health, income, and geography were extracted from external sources using zip code and congressional district location (eMethods 1 in the Supplement ). 12 , 13 , 14 , 15 External measures of socioeconomic disparities were not extracted for studies that occurred at a statewide level or that included data from patients across the United States, as the specific tools we used to determine these values were limited by units of analysis at the county, congressional district, and/or geographic address level.

COVID-19 has been strongly associated with lower socioeconomic status in racial/ethnic minorities. 16 Accordingly, ADI was used as a quantitative measure of socioeconomic disadvantage, and it accounts for several factors, such as income, education, employment, and housing quality. The Urban Core Opportunity Index (UOI) measures the urbanicity of geographic location, through the characterization of factors such as the amount of renters and households without vehicles. 15

We also examined the association of clinical care quality with COVID-19 positivity, mortality, ICU admission, and hospitalization through metaregression analysis. Specifically, we investigated the following measures of clinical care quality: preventable hospital stays, ratio of the population to primary care physicians, and percentage of uninsured individuals. A higher rate of preventable hospital stays represents a lower quality of available medical care, and a higher ratio of the population to physicians refers to a larger population with access to only 1 primary care physician. 14

Statistical Analysis

All data analysis was conducted using R Studio version 4.1.1 (R Project for Statistical Computing). Analyses were conducted separately for each racial and ethnic group in the following cohorts: COVID-19 positivity, ICU admission, hospitalization, and mortality. Studies with missing data for a particular cohort or variable were excluded from the respective analysis. The following analyses were conducted to investigate the association of race and ethnicity with COVID-19 outcomes. Combined prevalence refers to the incidence of COVID-19 outcomes in a certain population per 1000 patients. Metaregression analysis was conducted to assess associations between study effect size and socioeconomic variables extracted by study location. Relative risk ratios (RRs) and odds ratios (ORs) were also used to assess the associations of race and ethnicity with COVID-19 outcomes, with White individuals as the reference group. Both RR and OR values were adjusted for several key confounders using a linear mixed-effect model (eMethods 3 in the Supplement ). Statistical significance was set at P  < .05, and all tests were 2-tailed. The Egger test was used to assess publication bias, with P  < .05 as the level of statistical significance (eFigures 2-4 in the Supplement ). Information for all the studies is reported in detail in eTables 1 and 2 in the Supplement .

Study Characteristics

A total of 4 318 929 patients from 68 studies 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 , 62 , 63 , 64 , 65 , 66 , 67 , 68 , 69 , 70 , 71 , 72 , 73 , 74 , 75 , 76 , 77 , 78 , 79 , 80 , 81 , 82 , 83 , 84 were included in this meta-analysis ( Table ). Overall, 370 933 patients (8.6%) were African American, 9082 (0.2%) were American Indian or Alaska Native, 101 793 (2.4%) were Asian American, 851 392 identified as Hispanic/Latino (19.7%), 7417 (0.2%) were Pacific Islander, 1 037 996 (24.0%) were White, and 269 040 (6.2%) identified as multiracial or of other racial or ethnic group. The studies were separated into cohort and cross-sectional studies for data analysis. All unadjusted and adjusted RR and OR values are reported in Figure 1 and Figure 2 and eTables 3 and 4 in the Supplement .

Abbreviations: ICU, intensive care unit; NA, not applicable.

COVID-19 Positivity Rates

In age- and sex-adjusted analyses, we found that African American and Hispanic individuals were significantly more likely to test positive for COVID-19 than White individuals (African American: RR, 3.54; 95% CI, 1.38-9.07; P  = .008; Hispanic: RR, 4.68; 95% CI, 1.28-17.20, P  = .02) ( Figure 1 ). There was a lack of data to calculate age- and sex-adjusted RR and OR values for Asian American individuals. Following adjustment for ADI, African American and Hispanic individuals were almost 2 times as likely to test positive for COVID-19 as White individuals (African American: RR, 2.01; 95% CI, 1.04-3.88; P  = .04; Hispanic: RR, 2.09; 95% CI, 1.13-3.88; P  = .02), followed by Asian American individuals (RR, 1.12; 95% CI, 1.04-1.21; P  = .003) ( Figure 1 ). After adjustment for clinical care quality, we found that African American individuals were still the most likely to test positive for COVID-19 (RR, 1.79; 95% CI, 1.11-3.17; P  = .03), followed by Asian American individuals (RR, 1.16; 95% CI, 1.03-1.31; P  = .02) ( Figure 1 ). Hispanic individuals did not exhibit significant results following adjustment for clinical care quality. Interestingly, adjustment for the UOI demonstrated that Asian American individuals face the highest risk of COVID-19 positivity (RR, 1.13; 95% CI, 1.07-1.19, P  < .001) (eTable 3 in the Supplement ). We did not observe significant results in African American and Hispanic individuals following adjustment for UOI. Combined prevalence values demonstrated similar trends, with African American individuals having the highest prevalence of COVID-19 positivity (eFigure 5 in the Supplement ). In summary, with some exceptions, adjusting for ADI and clinical care quality significantly decreased the risk of COVID-19 infection in African American and Hispanic individuals when compared with White individuals. However, the risk still remained high in these populations following adjustment.

Risk of ICU Admission

COVID-19 disease severity was assessed through ICU admission and hospitalization rates among various racial and ethnic groups (eFigure 6 in the Supplement ). Following adjustment for sex, Asian American individuals had a significant RR of 1.93 (95% CI, 1.60-2.34; P  < .001) compared with White individuals ( Figure 1 ).

Mortality Rates in Cohort and Cross-sectional Studies

The combined prevalence of COVID-19 mortality rates in cohort studies was highest among White individuals (161.12 per 1000 patients), followed by African American individuals (143.99 per 1000 patients), Hispanic/Latino individuals (130.51 per 1000 patients), and Asian American individuals (42.99 per 1000 patients) (eTable 5 in the Supplement ). In cross-sectional studies, the combined prevalence of mortality rates were highest among African American individuals (277.15 per 1000 patients), followed by Hispanic individuals (213.34 per 1000 patients), White individuals (173.38 per 1000 patients), and Asian individuals (80.4 per 1000 patients) (eTable 5 and eFigure 7 in the Supplement ).

The ADI-adjusted RR for cross-sectional studies found that Hispanic individuals were at a lower risk of COVID-19 mortality than White individuals (RR, 0.44; 95% CI, 0.31-0.61; P  < .001). Similarly, the county median income–adjusted RR showed that Hispanic and Asian American individuals were at a lower risk of COVID-19 mortality than White individuals (Hispanic: RR, 0.43; 95% CI, 0.41-0.46; P  < .001; Asian American: RR, 0.44; 95% CI, 0.36-0.54; P  = .001) ( Figure 1 ).

ADI and Racial Disparities in COVID-19 Mortality

We further investigated the association of ADI with COVID-19 positivity and disease severity by race and ethnicity through metaregression analysis. A higher ADI corresponds to worse socioeconomic status. Accordingly, we found that an increase in ADI was positively associated with the mortality rates of Asian American and Hispanic individuals in cross-sectional studies ( P  < .001) ( Figure 3 ). Interestingly, an increase in ADI was negatively associated with mortality rates of Hispanic individuals in cohort studies ( P  = .03) ( Figure 3 ).

The solid line represents the association between the 2 variables. The dashed lines represent the 95% CI. The circles represent the different studies this particular graph is accounting for, while the sizes of the circles represent the weight of each of the studies.

Metaregression With County Median Income

We conducted metaregression analysis to assess the association of county median income with COVID-19 outcomes by race and ethnicity. Although ADI is a more comprehensive measure of socioeconomic deprivation, we also analyzed county median income because it provided more significant results for RR/OR adjustment in comparison with ADI. Therefore, we determined that we should further examine any associations with income, as it may have been more strongly associated with COVID-19 outcomes than other socioeconomic measures included in ADI. In cohort studies, we found that county median income was negatively associated with mortality rates in Asian American populations ( P  < .001). In cross-sectional studies, higher county median income was associated with lower mortality rates in Hispanic and African American individuals ( P  < .001). County median income was also negatively associated with the proportion of White individuals admitted to the ICU ( P  = .02) ( Figure 4 ; eFigure 8 in the Supplement ).

Through further metaregression analysis, we determined that Hispanic individuals had a positive association of increasing income and positivity rates ( P  = .03). However, African American individuals displayed a negative association between income and positivity rates ( P  = .02).

We additionally conducted Spearman correlations to assess the degree of association between these studied determinants. We observed a strong, positive correlation between county median income and area deprivation index ( R  = 0.61; P  < .001), as county median income was among the measures included in developing the ADI (eFigure 9 in the Supplement ). We found there was a lesser degree of association between county median income and measures of clinical care quality (eFigure 9 in the Supplement ).

Metaregression With Clinical Care Variables

In cohort studies, we found that an increase in number of preventable hospital stays ( P  = .04) and the population served by 1 primary care physician ( P  = .009) were associated with a decrease in positivity among Asian American individuals (eFigure 10 in the Supplement ). Conversely, the population served by 1 primary care physician was positively associated with COVID-19 positivity among Hispanic individuals ( P  < .001). In cross-sectional studies, we found that the ratio of the population served to primary care physicians was positively correlated with mortality among White individuals ( P  < .001). The percentage of uninsured individuals was positively associated with positivity among African American ( P  < .001) and White ( P  = .01) individuals in cohort and cross-sectional studies.

Risk of Bias Across Studies

We found that cohort studies detailing the proportion of Asian American and Hispanic individuals who tested positive for COVID-19, cohort studies detailing the proportion of Asian American individuals admitted to the ICU, cross-sectional studies detailing the proportion of Asian American individuals who died, and cohort studies detailing the proportion of African American individuals who died exhibited publication bias. To evaluate the association of study heterogeneity with summary proportions, we conducted leave-one-out sensitivity analysis to measure the effects of outliers (eTable 6 and eFigures 11-13 in the Supplement ). We found that summary proportions were not significantly altered by the removal of these outliers (eFigures 14 and 15 in the Supplement ). However, we observed that following the removal of the outlier in the COVID-19 mortality group (ie, cohort studies), African American individuals had the highest rate of mortality followed by Asian American individuals. Prior to removal of the outlier, we found that mortality rates were highest among White individuals. We additionally observed high heterogeneity statistics in our results, indicating that there may be variability in the studies included.

In our meta-analysis, we found that COVID-19 positivity and ICU admission rates were higher in African American, Hispanic, and Asian American individuals compared with White individuals, with some exceptions. Our results are consistent with previous findings that suggest that racial and ethnic minority groups face a higher risk of ICU admission and COVID-19 positivity but a lower risk of mortality than White populations. 65 , 85 , 86 , 87 , 88 , 89

However, current meta-analyses do not provide associations with socioeconomic variables, which are highly implicated in COVID-19 outcomes. Therefore, in this study, we aimed to investigate both racial and ethnic disparities in COVID-19 outcomes as well as their associations with socioeconomic variables.

Following adjustment for ADI and clinical care quality, we found that risk of COVID-19 positivity in African-American and Hispanic individuals substantially decreased. However, the risk for COVID-19 positivity following adjustment remained higher in these minoritized populations when compared with Whites. As such, this occurrence may be because of the overrepresentation of members of racial and ethnic minority groups in essential jobs, which increase exposure to COVID-19. Furthermore, comorbidities, such as hypertension or obesity, are prevalent among minority populations, thus contributing to worsened disease outcomes. 8 , 90 , 91 To our knowledge, we are the first to adjust RRs and ORs of race-associated COVID-19 outcomes using health care quality and access.

We further examined the association of socioeconomic determinants with COVID-19 positivity rates, mortality rates, hospitalization, and ICU admission in racial and ethnic minority groups through metaregression analysis. Increased deprivation was found to be associated with increased mortality in Asian American individuals. Paradoxically, an increase in county median income was associated with increased mortality rates in Asian American individuals. This result suggests that factors other than income that contribute to ADI, such as education, housing equality, and employment, could affect Asian American populations. One hypothesis is that a large number of Asian American individuals work in health care settings, which can lead to increased mortality rates that do not reflect the general population of the surrounding community. 92

An increase in deprivation was also found to be associated with decreased mortality rates in Hispanic individuals in cohort studies, although the opposite result was seen in cross-sectional studies. This inconsistency suggests that further research is needed to establish conclusively the association between mortality rates and deprivation in Hispanics.

We additionally assessed associations between measures of clinical care quality and COVID-19 outcomes. Curiously, we found that an increase in preventable hospital stays and the population served by 1 primary physician were associated with a decrease in the percentage of Asian American individuals who tested positive for COVID-19, suggesting again that other variables may be affecting COVID-19 positivity rates in this population.

Conversely, we observed a positive association between lack of primary care physician access (ie, increased ratio of population to physician) and COVID-19 positivity among Hispanic individuals. Past studies have demonstrated that Hispanic individuals are less likely to delay care if the primary care physician to patient ratio is improved. 93

An increase in the number of uninsured individuals was also positively associated with COVID-19 positivity among African American individuals. African American individuals are less likely to have health insurance coverage compared with White individuals. 94 Members of racial and ethnic minority groups who are uninsured may also not have access to COVID-19 tests. 10

Collectively, our findings demonstrate that racial and ethnic minority groups have faced higher risk of COVID-19 positivity and ICU admission. Public health policies should address socioeconomic and racial disparities to reduce exposure to and fatality from COVID-19 in underrepresented populations. Increasing equitable access to health care and improving resources for underserved populations may reduce exposure to COVID-19 in racial/ethnic minorities.

Limitations

Our study has several limitations. First, we found high heterogeneity statistics, indicating that there may be variation in the effect sizes of the studies. Second, a number of publications that were included had incomplete or missing data on mortality, positivity, ICU admission, and hospitalization rates. Moreover, there were limited data on several racial and ethnic groups. There was also a lack of information on comorbidities in some studies, which limited our ability to adjust for these variables. Additionally, several study cohorts exhibited publication bias. As publication bias reduces the accuracy of results, the validity of results in these particular study cohorts may be limited.

Conclusions

In this study, African American, Hispanic, and Asian American individuals were at considerably higher risk of COVID-19 positivity and ICU admission compared with White individuals. Adjustment for social determinants of health and socioeconomic factors decreased risks of COVID-19 positivity in racial and ethnic minority groups; however, several factors were not accounted for by these variables. We also observed that decreased access to clinical care was positively associated with COVID-19 positivity in Hispanic and African American individuals. In conclusion, we found that racial and ethnic disparities in COVID-19 outcomes could be accounted for by socioeconomic determinants in some populations, such as African American, Hispanic, and Asian American individuals.

Supplement.

eTable 1. Detailed Study Summary Characteristics of All 68 Included Studies

This website uses cookies to ensure you get the best experience. Learn more about DOAJ’s privacy policy.

Hide this message

You are using an outdated browser. Please upgrade your browser to improve your experience and security.

The Directory of Open Access Journals

Quick search.

Health and Human Rights (Dec 2020)

The Disproportional Impact of COVID-19 on African Americans

• Maritza Vasquez Reyes

Affiliations

Read online

WeChat QR code

• Publications
• Conferences & Events
• Professional Learning
• Science Standards
• Awards & Competitions
• Instructional Materials
• Free Resources
• For Preservice Teachers
• NCCSTS Case Collection
• Science and STEM Education Jobs
• Interactive eBooks+
• Digital Catalog
• Regional Product Representatives
• e-Newsletters
• Browse All Titles
• Bestselling Books
• Latest Books
• Popular Book Series
• Submit Book Proposal
• Web Seminars
• National Conference • New Orleans 24
• Leaders Institute • New Orleans 24
• National Conference • Philadelphia 25
• Exhibits & Sponsorship
• Submit a Proposal
• Conference Reviewers
• Past Conferences
• Latest Resources
• Professional Learning Units & Courses
• For Districts
• Online Course Providers
• Schools & Districts
• College Professors & Students
• The Standards
• Teachers and Admin
• eCYBERMISSION
• Toshiba/NSTA ExploraVision
• Junior Science & Humanities Symposium
• Teaching Awards
• Climate Change
• Earth & Space Science
• New Science Teachers
• Early Childhood
• Middle School
• High School
• Postsecondary
• Informal Education
• Journal Articles
• Lesson Plans
• e-newsletters
• Science & Children
• Science Scope
• The Science Teacher
• Journal of College Sci. Teaching
• Connected Science Learning
• NSTA Reports
• Next-Gen Navigator
• Science Update
• Teacher Tip Tuesday
• Trans. Sci. Learning

MyNSTA Community

• My Collections

Middle School High School    |    Daily Do

What Causes the Disproportionate Impact of COVID-19 on Racial and Ethnic Minority Groups?

By Todd Campbell and Okhee Lee

Share Discuss

Biology Crosscutting Concepts Disciplinary Core Ideas Is Lesson Plan Life Science NGSS Phenomena Science and Engineering Practices Three-Dimensional Learning Middle School High School

Sensemaking Checklist

Teachers and families across the country are facing a new reality of providing opportunities for students to do science through distance and home learning. The Daily Do is one of the ways NSTA is supporting teachers and families with this endeavor. Each weekday, NSTA will share a sensemaking task teachers and families can use to engage their students in authentic, relevant science learning. We encourage families to make time for family science learning (science is a social process!) and are dedicated to helping students and their families find balance between learning science and the day-to-day responsibilities they have to stay healthy and safe.

Interested in learning about other ways NSTA is supporting teachers and families? Visit the NSTA homepage .

Sensemaking is actively trying to figure out how the world works (science) or how to design solutions to problems (engineering). Students do science and engineering through the science and engineering practices . Engaging in these practices necessitates that students be part of a learning community to be able to share ideas, evaluate competing ideas, give and receive critique, and reach consensus. Whether this community of learners is made up of classmates or family members, students and adults build and refine science and engineering knowledge together.

Introduction

In today’s task, students answer the following questions: What are the causes of the disproportionate impact of COVID-19 on racial and ethnic minority groups? Why is the Centers for Disease Control (CDC) and Prevention guidance for how to slow the spread of COVID-19 necessary, but insufficient? What kinds of system-level solutions can our society implement to address the disproportionate impact of COVID-19 on racial and ethnic minority groups?

Students use science and engineering practices alongside disciplinary core ideas and crosscutting concepts to identify and explain the causes of the disproportionate impact of COVID-19 on racial and ethnic minority groups. Then they consider why the CDC guidance for slowing the spread of COVID-19 is necessary, but insufficient to address the causes that have led to the disproportionate impact of COVID-19. Finally, they propose system-level solutions for addressing the disproportionate impact of COVID-19.

Today’s task builds on ideas introduced in the following Daily Dos: How Can We Make Informed Decisions to Keep Ourselves and Our Communities Safe During the COVID-19 Pandemic? by Todd Campbell and Okhee Lee, and Are There Differences in How People Are Affected by the COVID-19 Pandemic in the United States? If So, Why Are There Differences, and What Should We Do About the Disproportionate Impact of COVID-19? by Todd Campbell, Okhee Lee, Eileen Murray, and John Russell.

Daily Do Playlist: Tracking COVID-19 in the United States

What causes the disproportionate impact of COVID-19 on racial and ethnic minority groups? is a stand-alone task. However, it can be taught as part of an instructional sequence in which students are provided authentic opportunities to develop and employ the science and engineering practice Using Mathematics and Computational Thinking to make sense of the spread of COVID-19 through the U.S. population and the disproportionate number of cases and deaths in non-white communities. Students further use data to support them in identifying actions they can take to keep their families and communities safe and in implementing their proposed solutions to ending health disparities.

View Playlist

Part 1. What are the indicators of the disproportionate impact of COVID-19 on racial and ethnic minority groups? [Data Science and Critical Consciousness]

The purpose of Part 1 is to help students see the evidence for the disproportionate impact of COVID-19 on racial and ethnic minority groups. Share the accompanying Student Journal with students before you begin.

What do you notice and wonder? What questions or comments do you have? Figure 1 presents the CDC data on the disproportionate impact of COVID-19 cases, hospitalizations, and deaths by race and ethnicity on certain racial and ethnic populations in the United States.

To begin, in small groups of 2–3 or individually, students share their noticings and wonderings about Figure 1.

Note: From Risk for COVID-19 Infection, Hospitalization, and Death by Race/Ethnicity, by Centers for Disease Control and Prevention, 2021 ( https://www.cdc.gov/coronavirus/2019-ncov/covid-data/investigations-discovery/hospitalization-death-by-race-ethnicity.html ).

Next, students write the following in their journals:

• Write an argument about which racial and ethnic groups have been most affected by COVID-19.
• Provide evidence for this argument by using the data in Figure 1.

Part 2. What is the CDC guidance for slowing the spread of COVID-19? What are the scientific explanations for the CDC guidance? [Simulations/Computer Science]

The purpose of Part 2 is to help students understand the scientific explanations behind the CDC guidance for slowing the spread of COVID-19. The CDC guidance is presented at the bottom of Figure 1 and explored in more depth through the activities below.

Activity A: Wear a Mask. In this activity, students learn about the role that masks play in slowing the spread of COVID-19.

Students explore this simulation, published by Live Science: To Mask or Not to Mask: This Simulation Shows Why It’s a Good Idea to Wear a Mask . [To play the simulation, scroll down to the last video embedded in the article and click on the white arrow in the middle of the black box.] In their journals, students write a scientific explanation for how masks slow the spread of COVID-19.

Activity B: Stay 6 Feet Apart. In this activity, students learn about the role that staying 6 feet apart plays in slowing the spread of COVID-19.

Students explore this simulation, published in The Washington Post article “ Military-Grade Camera Shows Risks of Airborne Coronavirus Spread .” In their journals, students write a scientific explanation for how staying 6 feet apart slows the spread of COVID-19.

Activity C: Wash Your Hands. In this activity, students learn about the role that handwashing plays in slowing the spread of COVID-19.

Students watch the following video, available at the Health Matters website:  How Soap Suds Kill the Coronavirus .  In their journals, students write a scientific explanation for how handwashing slows the spread of COVID-19.

Part 3. What are the societal causes of the disproportionate impact of COVID-19 on racial and ethnic minority groups? How are these societal causes connected to the CDC guidance for slowing the spread of COVID-19? [Critical Consciousness]

The purpose of Part 3 is twofold. First, students recognize societal causes of the disproportionate impact of COVID-19 on racial and ethnic minority groups (e.g., individuals from certain racial and ethnic minority groups are overrepresented in the populations of essential workers and those who are incarcerated). Second, students connect these causes to challenges these groups might experience when trying to follow the CDC guidance (e.g., staying 6 feet apart).

Students read four news articles:

• Article 1. " Coronavirus Infection by Race: What’s Behind the Health Disparities? "
• Article 2. " How COVID-19 Is Highlighting Racial Disparities in Americans’ Health "
• Article 3. " In California, the Pandemic Hits Latinos Hard "
• Article 4. “ Exclusive: Indigenous Americans Dying From Covid at Twice the Rate of White Americans ”

After reading each article, students answer these questions in their journals:

• Identify the racial and ethnic minority groups referenced in the article.
• List societal causes identified in the article.
• Connect these societal causes to the CDC guidance that the identified racial and ethnic minority groups might not be able to follow.

Part 4. How does the CDC guidance for slowing COVID-19's spread (wearing a mask, staying 6 feet apart, washing your hands), which is based on science, fail to address how systemic racism is causing inequities related to COVID-19 for racial and ethnic minority groups? [Critical Conciousness]

The purpose of Part 4 is to help students recognize that while following the CDC guidance for slowing the spread of COVID-19 is essential, not all citizens are able to follow this guidance due to the impact of systemic racism in the United States, which has resulted in significant differences among racial and ethnic groups regarding employment opportunities, wages, living arrangements, access to medical care, access to running water, and incarceration rates.

Students combine what they have learned in Part 1 (to what extent racial and ethnic minority groups are disproportionately impacted by COVID-19), Part 2 (scientific explanations in the CDC guidance for slowing the spread of COVID-19), and Part 3 (societal causes that have led to the disproportionate impact of COVID-19 on racial and ethnic minority groups).

In their journals, students write an essay to answer this question:

How does the CDC guidance for slowing the spread of COVID-19 (wearing a mask, staying 6 feet apart, washing your hands), which is based on science, fail to address systemic racism that has created inequities for racial and ethnic minority groups?

Part 5. Beyond individual decision-making related to CDC guidance, what system-level solutions can our society implement to address the disproportionate impact of COVID-19 on racial and ethnic minority groups? [Critical Consciousness, Media Literacy]

The purpose of Part 5 is to help students recognize that beyond following the CDC guidance for slowing the spread of COVID-19, system-level solutions are necessary if the United States is to address the disproportionate impact of COVID-19 on racial and ethnic minority groups. While students should understand the need for individual action to slow the spread of COVID-19 in their communities, they should also understand that without equitable access to resources in society, ethnic and racial minority groups will continue to be disproportionately harmed by COVID-19.

Students watch two videos and explore a website that describe different approaches to addressing systemic racism in the United States.

• Video 1. Inside California Politics: Former Stockton Mayor, Oakland Mayor Talk Guaranteed Income for Residents
• Video 2. Biden Signs Four Executive Actions Focused on Racial Equity
• Website. Racial Equity Tools (Students select one additional system-level solution they believe could help address systemic racism.)

After each activity, students return to their journals to identify (1) what approach is proposed/implemented, (2) how this approach addresses systemic racism, and (3) how this approach could reduce the disproportionate impact of COVID-19 on racial and ethnic minority groups.

Part 6. What system-level solutions can we propose to local, state, or national leaders to address the disproportionate impact of COVID-19 on racial and ethnic minority groups? [Critical Consciousness and Civic Engagement]

Working in groups of 2–3, students propose system-level solutions to the disproportionate impact of COVID-19 on racial and ethnic minority groups and share their ideas with local, state, or national leaders.

Using the Common Cause: Find Your Representatives website, students identify a local, state, or national leader with whom to share their system-level solutions.

( Note: To identify leaders, the site asks for a local address. Students can enter their own address, or the teacher can provide the school's address. After students enter an address, national, state, and local leaders are identified and links to their homepages are provided. Each homepage contains that leader’s contact information [e.g., e-mail and mailing address]. )

Students write a letter proposing one or two approaches for mitigating the disproportionate impact of COVID-19 on racial and ethnic minority groups. Students use this resource for helpful guidance on writing a letter to political leaders: Guidelines: How to Write a Letter to a Politician .

NSTA has created a What causes the disproportionate impact of COVID-19 on racial and ethnic minority groups? collection of resources to support teachers and families using this task. If you're an NSTA member, you can add this collection to your library by clicking on Add to My Library (near top of page).

The NSTA Daily Do is an open educational resource (OER) and can be used by educators and families providing students distance and home science learning. Access the entire collection of NSTA Daily Dos .

You may also like

Reports Article

Web Seminar

Join us on Thursday, November 21, 2024, from 7:00 PM to 8:00 PM ET, to learn about strategies educators can use when applying for a science teaching p...

The Disproportional Impact of COVID-19 on African Americans.

Chat with paper, disparities in covid-19 outcomes by race, ethnicity, and socioeconomic status: a systematic-review and meta-analysis., l-arginine and covid-19: an update., community-based, cluster-randomized pilot trial of a cardiovascular mhealth intervention: rationale, design, and baseline findings of the faith trial, structural modeling of covid-19 spread in relation to human mobility, “what will we do if we get infected”: an interview-based study of the covid-19 pandemic and its effects on the health and safety of sex workers in the united states, inequality and the health-care system in the usa, racial capitalism: a fundamental cause of novel coronavirus (covid-19) pandemic inequities in the united states., the neighborhood context of well-being, racial and ethnic disparities in health insurance coverage: dynamics of gaining and losing coverage over the life-course, applying siracusa: a call for a general comment on public health emergencies., related papers (5), covid-19 and human rights: a new inseparable relationship, the social integration of individuals with disabilities, hre in the era of global aging: the human rights of older persons in contemporary europe, gender, health, and human rights., guest editorial: is health a fundamental right for migrants.

Research roundup: How COVID-19 impacts African Americans

New studies offer data, discussion, and guidance for psychologists during the dual pandemics of systemic racism and COVID-19.

• Race and Ethnicity

Discrimination and racism have long contributed to negative emotional, mental, and physical health outcomes in African American communities. The COVID-19 pandemic has highlighted this fact, with recent data showing that 1 in 1,000 Black individuals have died from the coronavirus (APM Research Lab, 2020).

In this installment of “Research Roundup,” we look at studies that explore how discrimination has impacted the overall health of African Americans and ways psychologists can support Black communities now and when the pandemic subsides.

Learning from previous crises

Novacek, D. M., Hampton-Anderson, J. N., Ebor, M. T., Loeb, T. B., and Wyatt, G. E. (2020). Mental health ramifications of the COVID-19 pandemic for Black Americans: Clinical and research recommendations . Psychological Trauma: Theory, Research, Practice, and Policy, 12 (5), 449–451.

Sneed, R. S., Key, K., Bailey, S., and Johnson-Lawrence, V. (2020). Social and psychological consequences of the COVID-19 pandemic in African-American communities: Lessons from Michigan . Psychological Trauma: Theory, Research, Practice, and Policy, 12 (5), 446–448.

Two articles in the journal Psychological Trauma: Theory, Research, Practice, and Policy ® give an overview of the trauma African American communities have faced during previous public health crises and call attention to the differential impact of COVID-19 on communities of color. The studies show that Hurricane Katrina, the HIV/AIDS epidemic, and the Flint water crisis in Michigan all provide lessons for reducing negative outcomes among Black Americans during current times of crisis.

Discrimination and racism, along with economic and environmental disparities, already reduce access to education, health care, and resources in Black communities. And during the COVID-19 health emergency, government restrictions on public activities have further reduced or completely eliminated access to the care (Lund, 2020). Studies show that even when African American individuals do seek care, the assumed bias and lack of trust in providers plays a large role in continuation and quality of care. Because of the Flint water crisis, for example, many members of the Black community continue to experience adverse health effects and have significant distrust in government and public health officials.

Black communities also experience high rates of posttraumatic stress disorder (PTSD) compared to other ethnic/racial groups. When crises result in exclusion or isolation, adverse mental health issues like depression and PTSD are only exacerbated. These outcomes are also expected to occur as a result of the COVID-19 pandemic.

Early COVID-19 rates in Michigan show that Black Americans make up 41% of deaths, while only making up 13% of the population (Sneed, Key, Bailey, and Johnson-Lawrence, 2020). Given the ongoing community concerns in Flint, the state government has taken a multitude of steps to engage the community and local leaders in order to prevent adverse outcomes during the COVID-19 pandemic. These efforts are targeted at increasing community engagement, for example, through creation of diverse, multisector task forces and ongoing, virtual videoconferences and webinars that involve local organizations. Community engagement and support increase resilience within African American communities.

Early impact of COVID-19 in Black counties

Millett, G. A., Jones, A. T., Benkeser, D., Baral, S., Mercer, L., Beyrer, C., ... and Sullivan, P. S. (2020). Assessing differential impacts of COVID-19 on black communities . Annals of Epidemiology, 47 , 37–44.

Early, national data show that counties with a predominantly African American population are more heavily impacted by COVID-19, both in rates of diagnosis and death. While the COVID-19 pandemic is cited as one of the most unprecedented periods in recent history, this disparate impact of crises on Black communities regularly occurs.

Researchers gathered public, county-level data across the U.S. to assess which counties made up a disproportionate population of African Americans, compared to the national average (greater than or equal to 13%). This data was then compared to the Centers for Disease Control and Prevention (CDC) rates of COVID-19 cases and deaths, as reported by each state. Counties with less than a 13% population of Black Americans were used as a control group. Demographics, preexisting comorbidities, social, and environmental factors were also included, and the Bayesian hierarchical model was used to assess whether any of these factors confounded COVID-19 rates and deaths in Black counties.

At the time of this study, states largely did not account for or did not disaggregate ethnicity/race when submitting COVID-19 data, a common issue in public health data collection. The study concluded just over 20% of U.S. counties met criteria for Black communities (≥13%), yet African Americans made up 52% of COVID-19 cases and 58% of COVID-19 deaths. Higher rates of diagnosis and death persisted after controlling for confounding factors.

Family support mediates stress effects

Priest, J. B., McNeil Smith, S., Woods, S. B., and Roberson, P. N. E. (2020). Discrimination, family emotional climate, and African American health: An application of the BBFM . Journal of Family Psychology, 34 (5), 598–609.

For African Americans, family support appears to buffer the impact of negative stress on health brought on by discrimination. The biobehavioral family model (BBFM) has been used to explain the connection between social relationships and biobehavioral reactivity since 2008. This study extends the research on the BBFM to Black communities while additionally controlling for discrimination.

Researchers looked at the support and strain of relationships for African Americans as it intersects with racial discrimination and their effects on health. Between 2004–2006, 592 eligible African American adults from Milwaukee County, Wisconsin, were interviewed twice: first in-person using computer assistance and then by self-guided questionnaire. Four scales were used to assess aspects of support and strain in close family and intimate partner relationships (about 40% of participants met criteria for intimate partner relationships). Participants’ experience of discrimination was measured both across the lifetime and as a daily occurrence, and then participants were asked how much discrimination impacts quality of life. Biobehavioral reactivity was operationalized as depression and anxiety.

Research in African American communities shows the interdependent nature of relationships and its influence on health vis-à-vis stress. This study concluded that family support mediates the negative stress effects from discrimination on health. Intimate partner relationships in this study (and other studies done with BBFM) did not have significant associations with support or strain on health outcomes.

Clinical implications

Discrimination and racism impact all areas of Black lives, individually and collectively, including where people live, education, types of jobs, access to health care, insurance coverage, and a host of other factors. Public health crises like the COVID-19 pandemic magnify this impact and highlight the disproportionate negative effect in Black communities.

Recognizing the differential impact of COVID-19 on Black communities and individuals and understanding why it occurs is essential in order to mitigate this situation. Psychologists can help flatten the curve of the emotional and mental health trauma associated with racial discrimination before, during, and after the COVID-19 pandemic. Research highlights the importance of belonging and community support for African Americans. Collaborating with local groups may help individuals and families to build trust with government and public health officials, likely increasing access and quality of care. Highlighting challenging experiences and how they were overcome helps to foster resilience and may benefit care. Including families in psychological care and building support networks may be especially critical for Black individuals given their powerful role in mitigating the negative impact of discrimination on health.

Steps are being taken to improve access and resources during the COVID-19 pandemic for underserved ethnic and racial groups in the United States. For example, the Boston University Center for Antiracist Research has developed a COVID Racial Data Tracker in collaboration with the National Institutes of Health to improve COVID-19 testing in these communities. APA also submitted a letter to the Senate Special Committee on Aging (PDF, 299KB) to support legislation that would advance care and services for older minority adult populations , in addition to creating a resource to build trust to improve participation (PDF, 108KB) of marginalized communities in COVID-19 testing and contact tracing.

Further reading

• APA Guidelines on Race and Ethnicity in Psychology (PDF, 577KB) , 2019
• APA Multicultural Guidelines : An Ecological Approach to Context, Identity, and Intersectionality, 2017
• APM Research Lab. (2020, October 15). COVID-19 deaths analyzed by race and ethnicity .
• COVID-19: Bias, Discrimination, and Equity Resources . Information, resources and support for behavioral and social scientists, advocates, activists, and community serving practitioners during COVID-19.
• Lund, E. (2020). Even more to handle: Additional sources of stress and trauma for clients from marginalized racial and ethnic groups in the United States during the COVID-19 pandemic . Counselling Psychology Quarterly .
• New guidance on race and ethnicity for psychologists . Guidelines urge psychologists to consider how their backgrounds shape their behavior.

Related and recent

Recommended reading, contact apa services.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings
• My Bibliography
• Collections
• Citation manager

Save citation to file

Email citation, add to collections.

• Create a new collection
• Add to an existing collection

Add to My Bibliography

Your saved search, create a file for external citation management software, your rss feed.

• Search in PubMed
• Search in NLM Catalog
• Add to Search

Disproportionate Impact of COVID-19 on Racial and Ethnic Minority Groups in the United States: a 2021 Update

Affiliations.

• 1 Division of Infectious Diseases, Department of Medicine, Mayo Clinic, 200 First Street SW, Rochester, MN, 55905, USA. [email protected].
• 2 Division of Infectious Diseases, Department of Medicine, Mayo Clinic, 200 First Street SW, Rochester, MN, 55905, USA.
• 3 Center for Health Equity and Community Engagement Research, Mayo Clinic, 200 First Street SW, Rochester, MN, 55905, USA.
• 4 Department of Family Medicine, Mayo Clinic, Rochester, MN, USA.
• 5 Center for Health Equity and Community Engagement Research, Mayo Clinic, 200 First Street SW, Rochester, MN, 55905, USA. [email protected].
• 6 Division of Community Internal Medicine, Department of Medicine, Mayo Clinic, Rochester, MN, USA. [email protected].
• PMID: 34647273
• PMCID: PMC8513546
• DOI: 10.1007/s40615-021-01170-w

The COVID-19 pandemic's disproportionate impact on people from some racial and ethnic groups in the U.S. persisted throughout 2021. Black, Latinx, and American Indian persons have been hospitalized and died at a higher rate than White persons consistently from the start of the pandemic. Early data show that hospitalization and mortality rates for Black, Latinx, and American Indian children are higher than White children in a worrying trend. The pandemic has likely worsened the gaps in wealth, employment, housing, and access to health care: the social determinants of health that caused the disparities in the first place. School closures will have a long-lasting impact on the widening achievement gaps between Black and Latinx students and White students. In the earlier vaccination phase, Black and Latinx persons were being vaccinated at a lower rate than their proportion of cases due to vaccine hesitancy, misinformation, and barriers to access. Vaccine hesitancy rates among these groups have since decreased and are now comparable to White persons. Aggregated data make it challenging to paint a picture of the actual impact of COVID-19 on Asian Americans as they are a diverse group with significant disparities. All of this highlights that we have much work to do in dismantling systemic racism, engaging communities we serve, and advancing health equity to prepare us for future pandemics and a more just healthcare system.

Keywords: COVID-19 impact; Race and ethnicity; Racial disparities; Systemic racism.

© 2021. W. Montague Cobb-NMA Health Institute.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Temporal trends in 30-day running…

Temporal trends in 30-day running averages in COVID-19 cases, severe cases, and mortality…

Similar articles

• Improving Pathways to Care for Ethnic Minority Communities. Fuller H, King-Okoye M. Fuller H, et al. Adv Exp Med Biol. 2024;1458:157-173. doi: 10.1007/978-3-031-61943-4_11. Adv Exp Med Biol. 2024. PMID: 39102196 Review.
• Racial and Ethnic Disparities in Rates of COVID-19-Associated Hospitalization, Intensive Care Unit Admission, and In-Hospital Death in the United States From March 2020 to February 2021. Acosta AM, Garg S, Pham H, Whitaker M, Anglin O, O'Halloran A, Milucky J, Patel K, Taylor C, Wortham J, Chai SJ, Kirley PD, Alden NB, Kawasaki B, Meek J, Yousey-Hindes K, Anderson EJ, Openo KP, Weigel A, Monroe ML, Ryan P, Reeg L, Kohrman A, Lynfield R, Bye E, Torres S, Salazar-Sanchez Y, Muse A, Barney G, Bennett NM, Bushey S, Billing L, Shiltz E, Sutton M, Abdullah N, Talbot HK, Schaffner W, Ortega J, Price A, Fry AM, Hall A, Kim L, Havers FP. Acosta AM, et al. JAMA Netw Open. 2021 Oct 1;4(10):e2130479. doi: 10.1001/jamanetworkopen.2021.30479. JAMA Netw Open. 2021. PMID: 34673962 Free PMC article.
• Changing the Narrative: Structural Barriers and Racial and Ethnic Inequities in COVID-19 Vaccination. Njoku A, Joseph M, Felix R. Njoku A, et al. Int J Environ Res Public Health. 2021 Sep 20;18(18):9904. doi: 10.3390/ijerph18189904. Int J Environ Res Public Health. 2021. PMID: 34574827 Free PMC article.
• Racial and ethnic disparities in excess mortality among U.S. veterans during the COVID-19 pandemic. Feyman Y, Avila CJ, Auty S, Mulugeta M, Strombotne K, Legler A, Griffith K. Feyman Y, et al. Health Serv Res. 2023 Jun;58(3):642-653. doi: 10.1111/1475-6773.14112. Epub 2022 Dec 30. Health Serv Res. 2023. PMID: 36478574 Free PMC article.
• Racial and ethnic disparities in cardiometabolic disease and COVID-19 outcomes in White, Black/African American, and Latinx populations: Social determinants of health. Grosicki GJ, Bunsawat K, Jeong S, Robinson AT. Grosicki GJ, et al. Prog Cardiovasc Dis. 2022 Mar-Apr;71:4-10. doi: 10.1016/j.pcad.2022.04.004. Epub 2022 Apr 28. Prog Cardiovasc Dis. 2022. PMID: 35490870 Free PMC article. Review.
• Strategies to address COVID-19 vaccine hesitancy in First Nations peoples: a systematic review. Tinessia A, Clark K, Randell M, Leask J, King C. Tinessia A, et al. Glob Health Action. 2024 Dec 31;17(1):2384497. doi: 10.1080/16549716.2024.2384497. Epub 2024 Sep 4. Glob Health Action. 2024. PMID: 39230093 Free PMC article.
• A Randomized Controlled Trial of a Culturally Adapted, Community-Based, Remotely Delivered Mindfulness Program for Latinx Patients With Breast Cancer is Acceptable and Feasible While Reducing Anxiety. Juarez-Reyes M, Martinez E, Xiao L, Goldman Rosas L. Juarez-Reyes M, et al. Glob Adv Integr Med Health. 2024 Aug 16;13:27536130241274240. doi: 10.1177/27536130241274240. eCollection 2024 Jan-Dec. Glob Adv Integr Med Health. 2024. PMID: 39157776 Free PMC article.
• The effect of adding physician recommendation in digitally-enabled outreach for COVID-19 vaccination in socially/economically disadvantaged populations. Sumar K, Blue L, Fatahi G, Sumar M, Alvarez S, Cons P, Valencia N, Williams Z, Bhatti A, Parthasarathy S, Doubeni CA. Sumar K, et al. BMC Public Health. 2024 Jul 18;24(1):1933. doi: 10.1186/s12889-024-18648-x. BMC Public Health. 2024. PMID: 39026196 Free PMC article. Clinical Trial.
• Vietnamese Americans' level of trust in sources of information and willingness to participate in COVID-19 clinical trials. Vargas S, Siddiqi S, King B, Nguyen C, Diep J, Gilbert L, Nguyen BM. Vargas S, et al. J Clin Transl Sci. 2024 Apr 25;8(1):e88. doi: 10.1017/cts.2024.495. eCollection 2024. J Clin Transl Sci. 2024. PMID: 38784109 Free PMC article.
• Tai DBG, Shah A, Doubeni CA, et al. The disproportionate impact of COVID-19 on racial and ethnic minorities in the United States. Clin Infect Dis. 2021;72(4):703–706. doi: 10.1093/cid/ciaa815. - DOI - PMC - PubMed
• Jones CP. Systems of power, axes of inequity: parallels, intersections, braiding the strands. Med Care. 2014;52:S71–S75. doi: 10.1097/MLR.0000000000000216. - DOI - PubMed
• COVID Data Tracker Available at: https://covid.cdc.gov/covid-data-tracker/#demographics . Accessed 4 Oct 2021.
• Racial and Ethnic Diversity in the United States: 2010 Census and 2020 Census. Available at: https://www.census.gov/library/visualizations/interactive/racial-and-eth... . Accessed 22 Aug 2021.
• Baker D.W. Breaking links in the chain of racial disparities for COVID-19. JAMA Network Open 2021; 4(6): e2112879-e. 10.1001/jamanetworkopen.2021.12879. - PubMed

Publication types

• Search in MeSH

Related information

Linkout - more resources, full text sources.

• Europe PubMed Central
• PubMed Central
• MedlinePlus Health Information
• Citation Manager

NCBI Literature Resources

MeSH PMC Bookshelf Disclaimer

The PubMed wordmark and PubMed logo are registered trademarks of the U.S. Department of Health and Human Services (HHS). Unauthorized use of these marks is strictly prohibited.

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here .

Loading metrics

Open Access

Peer-reviewed

Research Article

Disproportionate impacts of COVID-19 on marginalized and minoritized early-career academic scientists

Roles Data curation, Formal analysis, Investigation, Methodology, Project administration, Visualization, Writing – original draft, Writing – review & editing

¶ ‡ Co-first authors

Affiliation Department of Psychology, University of Michigan, Ann Arbor, Michigan, United States of America

Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Visualization, Writing – original draft, Writing – review & editing

Affiliations Department of Psychology, University of Michigan, Ann Arbor, Michigan, United States of America, Department of Afroamerican and African Studies, University of Michigan, Ann Arbor, Michigan, United States of America

Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Writing – original draft, Writing – review & editing

Affiliations Department of Sociology, University of Michigan, Ann Arbor, Michigan, United States of America, Department of Mechanical Engineering, University of Michigan, Ann Arbor, Michigan, United States of America

Roles Conceptualization, Funding acquisition, Methodology, Writing – original draft, Writing – review & editing

* E-mail: [email protected]

Affiliations Lyman Briggs College, Michigan State University, East Lansing, Michigan, United States of America, Department of History, Michigan State University, East Lansing, Michigan, United States of America

Roles Visualization, Writing – review & editing

Affiliation Lyman Briggs College, Michigan State University, East Lansing, Michigan, United States of America

Affiliation Department of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan, United States of America

Roles Data curation, Investigation, Methodology, Project administration, Writing – review & editing

Affiliation Division of Social Sciences and History, Delta State University, Cleveland, Mississippi, United States of America

Roles Data curation, Investigation, Methodology, Writing – review & editing

Affiliations Lyman Briggs College, Michigan State University, East Lansing, Michigan, United States of America, Department of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan, United States of America, Department of Philosophy, Michigan State University, East Lansing, Michigan, United States of America

Roles Conceptualization, Funding acquisition, Methodology, Project administration, Supervision, Writing – original draft, Writing – review & editing

Affiliations Lyman Briggs College, Michigan State University, East Lansing, Michigan, United States of America, Department of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan, United States of America

• Hannah M. Douglas, 
• Isis H. Settles, 
• Erin A. Cech, 
• Georgina M. Montgomery, 
• Lexi R. Nadolsky, 
• Arika K. Hawkins, 
• Guizhen Ma, 
• Tangier M. Davis, 
• Kevin C. Elliott, 
• Kendra Spence Cheruvelil

• Published: September 13, 2022
• https://doi.org/10.1371/journal.pone.0274278
• Peer Review
• Reader Comments

Early research on the impact of COVID-19 on academic scientists suggests that disruptions to research, teaching, and daily work life are not experienced equally. However, this work has overwhelmingly focused on experiences of women and parents, with limited attention to the disproportionate impact on academic work by race, disability status, sexual identity, first-generation status, and academic career stage. Using a stratified random survey sample of early-career academics in four science disciplines ( N = 3,277), we investigated socio-demographic and career stage differences in the effect of the COVID-19 pandemic along seven work outcomes: changes in four work areas (research progress, workload, concern about career advancement, support from mentors) and work disruptions due to three COVID-19 related life challenges (physical health, mental health, and caretaking). Our analyses examined patterns across career stages as well as separately for doctoral students and for postdocs/assistant professors. Overall, our results indicate that scientists from marginalized (i.e., devalued) and minoritized (i.e., underrepresented) groups across early career stages reported more negative work outcomes as a result of COVID-19. However, there were notable patterns of differences depending on the socio-demographic identities examined. Those with a physical or mental disability were negatively impacted on all seven work outcomes. Women, primary caregivers, underrepresented racial minorities, sexual minorities, and first-generation scholars reported more negative experiences across several outcomes such as increased disruptions due to physical health symptoms and additional caretaking compared to more privileged counterparts. Doctoral students reported more work disruptions from life challenges than other early-career scholars, especially those related to health problems, while assistant professors reported more negative changes in areas such as decreased research progress and increased workload. These findings suggest that the COVID-19 pandemic has disproportionately harmed work outcomes for minoritized and marginalized early-career scholars. Institutional interventions are required to address these inequalities in an effort to retain diverse cohorts in academic science.

Citation: Douglas HM, Settles IH, Cech EA, Montgomery GM, Nadolsky LR, Hawkins AK, et al. (2022) Disproportionate impacts of COVID-19 on marginalized and minoritized early-career academic scientists. PLoS ONE 17(9): e0274278. https://doi.org/10.1371/journal.pone.0274278

Editor: Quinn Grundy, University of Toronto, CANADA

Received: March 14, 2022; Accepted: August 24, 2022; Published: September 13, 2022

Copyright: © 2022 Douglas et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: The de-identified data associated with this manuscript is available on openICPSR. To protect participant confidentiality, the data file does not include demographic information. [ https://doi.org/10.3886/E172961V1 ] The full study dataset will be under an embargo up to 3 years following the end date of the grant funding (2023). Following this period, the non-identifiable data will be deposited with the Inter-university Consortium for Political and Social Research (ICPSR) at the University of Michigan. You may contact the corresponding author for more information and for data analysis code.

Funding: This research was funded by National Science Foundation ECR-EHR core research grants #2000579 (University of Michigan) and #1954767 (Michigan State University) awarded to Isis H. Settles, Kendra Spence Cheruvelil, Erin A. Cech, Georgina M. Montgomery, and Kevin C. Elliott. https://beta.nsf.gov/funding/opportunities/ehr-core-research-ecrcore The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

The COVID-19 pandemic has produced striking disruptions to the day-to-day work of academic scientists and has required many scientists to shift their time away from research and toward increased teaching, mentoring, and service [ 1 – 3 ]. However, accumulating evidence suggests that these negative pandemic-related outcomes are not shouldered equally. Early research suggests that the COVID-19 pandemic has exacerbated existing socio-demographic inequalities such as lower representation and lack of inclusion among faculty in academic science [ 4 ]. As a result, scholars have expressed concerns that the work disruptions of the pandemic will result in even greater long-term underrepresentation of people from marginalized groups in the academy [ 5 – 8 ].

Research to date on the inequitable impact of COVID-19 on scientists’ work has overwhelmingly focused on women and parents in STEM, particularly in regard to research productivity as measured by the number of submitted research articles. Studies have found that compared to men, women faculty report increased workload from additional time spent engaged in teaching, mentoring, and service [ 4 ] and have experienced greater declines in research productivity (e.g., submitting fewer papers than men) [ 3 , 4 , 9 – 12 ], likely as a result of the greater service and caregiving demands they face. Among parents, it has been academic women, especially those with younger children, who have experienced the most significant declines in overall work time and their research productivity during COVID-19 [ 4 , 11 ].

Yet, beyond these initial insights from studies of gender differences, we have an incomplete picture of the landscape of disadvantages caused by COVID-19. The potential disparate impact on other groups that experience social minoritization (i.e., underrepresentation) and marginalization (i.e., devaluation)–particularly along race and ethnicity, sexual identity, and disability status–have been largely overlooked. Given that faculty of color already have a greater workload from visible and invisible service demands [ 13 ], the combined toll of the COVID-19 pandemic and the social pandemic of racial injustice likely has exacerbated these inequities through additional student support and service work [ 14 , 15 ]. The pandemic may similarly have disproportionately negative effects on scholars from other marginalized communities, such as sexual minorities, first-generation scholars, and scholars with disabilities. Research about work-related demands for people within multiple marginalized groups is critical because greater demands may be linked to the higher levels of stress, exhaustion, and burnout reported by women and faculty of color during the pandemic [ 3 , 16 – 19 ].

Such divergent impacts are likely to be especially problematic for those early in their careers. One study found that psychological distress has increased among graduate students since the pandemic, with higher rates among students from marginalized groups [ 20 ]. Other research has found differences in the impact of COVID-19 by faculty rank. For example, assistant professors, who are more likely to have young children, have reported more substantial declines in research progress than tenured faculty [ 21 ]. Assistant professors have reported feeling insufficient support from their colleagues, a loss of networks and collaborations, and fears about their ability to gain tenure [ 3 , 15 ]. Because research on the impact of the pandemic on academic science has focused primarily on faculty ranks, we know little about the differential impacts of COVID-19 on academic work among different groups of early-career scholars (e.g., graduate students, postdoctoral scholars, assistant professors). COVID-19 related impacts on these early-career scholars are especially important for the academic pipeline, as experiences during these formative years may affect scholars’ research trajectory and their interests in academic careers.

We addressed these knowledge gaps by examining differences in seven work outcomes due to the COVID-19 pandemic along six axes of socio-demographic variation (gender, parental caregiving, race, sexual identity, first-generation college student status, disability status) and three career stages. We conducted a representative survey of doctoral students, postdoctoral scholars, and assistant professors ( N = 3,243) using a stratified random sample of 124 departments of biology, physics, economics, and psychology in US institutions. Specifically, in April-May of 2021, we compared perceived changes over the past year in four work outcomes (research progress, workload, concern about career advancement, and support from mentors), as well as work disruptions due to three pandemic-related life challenges (physical health, mental health, and caregiving responsibilities) by each aforementioned socio-demographic status. Overall, we hypothesized that early-career scholars from marginalized groups would be more likely to report negative COVID-19 related work outcomes.

Materials and method

Participants were doctoral students ( n = 2,687), postdoctoral scholars ( n = 335), and assistant professors ( n = 221) who completed an online survey administered in April and May 2021. Participants were recruited from four STEM fields (biology, economics, physics, and psychology) at 124 different departments that were randomly selected and stratified across 2011 National Research Council S-rankings of departments within each field (see Participant Database and S1 Table in S1 File for more details). Participants provided informed consent online by selecting ‘Agree’ on the informed consent page. If participants did not consent to the study procedures, they were redirected out of the survey. All study materials were approved and deemed exempt from full review by the institutional review boards at both the University of Michigan (HUM00193386) and Michigan State University (STUDY00004853).

We asked participants how the COVID-19 pandemic impacted their work during the past year across seven work outcomes. We asked about change in research progress, workload, concern about career advancement, and support from mentors (1 = Greatly decreased to 5 = Greatly increased ). We also asked whether their work was disrupted in the past year by three COVID-19 related life challenges: physical health problems (e.g., sleep problems, headaches), mental health problems (e.g., mood problems, stress), and additional caretaking responsibilities at home ( 1 = Did not disrupt my work at all to 5 = Disrupted my work a great deal ).

Participants self-reported their membership in seven socio-demographic statuses: gender (man, woman, or non-binary [including genderqueer or gender fluid]); race (underrepresented minority [URM; Black or African American, Hispanic or Latina/o/x, Middle Eastern or North African, Native American, American Indian, or other Indigenous group], Asian/Asian American, or White); sexual identity (sexual minority [lesbian, gay, bisexual, pansexual, queer, asexual, demisexual, or other] or heterosexual); being a first-generation college student; having a mental, physical, or learning disability; and career stage (doctoral student, postdoc, or assistant professor). For gender, because the number of individuals identifying as gender non-binary, genderqueer, and/or gender fluid was small and their responses to work outcomes tended to be in the same direction as women’s responses, they were combined with women. We also combined Black or African American, Hispanic or Latina/o/x, Middle Eastern or North African, Native American, American Indian, or other Indigenous group into URM due to their underrepresented status and stigmatization in the academy. See SI Socio-demographic Variables for additional details and S2 Table in S1 File for sample size of socio-demographic groups.

Analytic plan

To assess differences in each work outcome described above, we ran seven separate multiple regressions on the full sample using the seven socio-demographic statuses as predictor variables: gender (women/non-binary or men), parental caregiving (primary caregiver, non-primary caregiver, or non-caregiver), race (URM, Asian, or White), disability status (at least one physical or mental disability or no disability), sexual identity (sexual minority or heterosexual), first-generation status (first-generation college student or not first-generation college student), and career stage (doctoral student, postdoctoral scholar, or assistant professor) (S3-S9 Tables in S1 File ). Analyses controlled for academic field (biology, psychology, economics, or physics) and departmental ranking (tier 1, tier 2, or tier 3) (see SI Participant Database for details on covariates). All analyses were conducted using STATA 17 software.

Due to the larger size of the sample of doctoral students compared to postdoctoral scholars and assistant professors, we also ran each of the seven multiple regressions separately for doctoral students and a combined group of postdoctoral scholars and assistant professors (hereafter referred to as “postdoc/asst prof”); we included these two groups in the same subgroup analyses due to their smaller sample sizes and because participants in both groups are more advanced among early-career scholars. The subsample analyses were identical to those for the full sample except that for the doctoral student analyses, academic rank was not included and for the postdoc/asst prof analyses, academic rank compared postdoctoral scholars and assistant professors. The results of each set of regressions (full sample including all three career stages, doctoral student subsample, and postdoc/asst prof subsample) are described below; we first describe findings that are consistent across the full sample and both subsamples, and then highlight when there were subsample differences in the patterns of relationships.

To confirm that the variability in COVID-19 outcomes were not driven by the distribution of participants in departments, we ran seven additional multilevel mixed-effects regression models with respondents nested in departments. Department-level variation did not significantly account for the variability in any of the seven work outcomes; results for these models are in the (S10 and S11 Tables in S1 File ). Finally, we conducted supplemental multiple regressions to capture the negative impact that COVID-19 disruptions had on job satisfaction, professional role confidence, turnover intentions, burnout disengagement, and burnout exhaustion; these are presented in the supplement for the overall sample as well as for the doctoral student and postdoc/asst prof subsamples (see S12—S16 Tables in S1 File for results).

Means, standard deviations, and correlations among COVID-19 impact variables are presented in S17 Table in S1 File for the full sample and each subsample.

Work disruptions due to COVID-19

Change in research progress ( fig 1 )..

Controlling for all other socio-demographic variables, we found that all participants, regardless of group membership, reported decreased research progress. Across samples, individuals with a disability reported a greater decrease in their research progress due to the pandemic compared to those without a disability. In the full sample, assistant professors reported a significant decrease in their research progress compared to doctoral students. However, differences in the impact of COVID-19 on research progress by gender, race, sexual identity, or first-generation status did not emerge across samples.

• PPT PowerPoint slide
• PNG larger image
• TIFF original image

Estimated marginal means of reported change in research progress due to the COVID-19 pandemic for the full sample (1a) and the doctoral student (1b) and postdoc/asst prof (1c) subsamples. Participants were asked to rate the degree to which research progress changed on a Likert scale from 1 = Greatly decreased to 5 = Greatly increased . Error bars represent the 95% CI. Numeric values in each bar are the estimated marginal mean and standard error. * indicates p < .05, ** indicates p < .01, and *** indicates p < .001. See S3 Table in S1 File for results of each model.

https://doi.org/10.1371/journal.pone.0274278.g001

When examining subsamples, we found differences in the effect of parental caregiving on research progress. For all groups, parents who were primary caregivers reported a greater decrease in research progress than both non-primary caregivers and non-parents. However, for the full sample and the postdoc/asst prof subsample, non-primary caregivers also reported a significant decrease in research progress compared to non-parents.

Change in workload ( Fig 2 ).

Overall, all groups indicated that their workload had increased to some extent. We found that assistant professors reported a significantly greater increase in their workload compared to doctoral students and postdoctoral scholars. However, across the three samples, there were no differences in workload by parental caregiving or sexual identity.

Estimated marginal means of reported change in workload due to the COVID-19 pandemic for the full sample (2a) and the doctoral student (2b) and postdoc/asst prof (2c) subsamples. Participants were asked to rate the degree to which their workload changed on a Likert scale from 1 = Greatly decreased to 5 = Greatly increased . Error bars represent the 95% CI. Numeric values in each bar are the estimated marginal mean and standard error. * indicates p < .05, ** indicates p < .01, and *** indicates p < .001. See S4 Table in S1 File for results of each model.

https://doi.org/10.1371/journal.pone.0274278.g002

When using the full sample or the doctoral subsample, we found that women and non-binary individuals and those with a disability reported a significantly greater increase in workload compared to men and those without a disability, respectively, but these differences did not emerge in the postdoc/asst prof subsample. When using the full sample, we found that URM and White individuals reported a significantly greater increase in their workload compared to Asian respondents. In the doctoral student subsample, URM individuals’ workloads only differed from Asian individuals’ workloads. No differences by race emerged in the postdoc/asst prof subsample. Finally, only in the postdoc/asst prof subsample, first-generation scholars reported a significantly greater increase in workload compared to non-first-generation scholars.

Change in concern about career advancement ( Fig 3 ).

All groups reported increased concern about their career advancement due to the COVID-19 pandemic. Across the three samples, those with a disability reported a significantly greater increase in concern about career advancement compared to those without a disability. Further, postdoctoral scholars reported a greater increase in these concerns compared to assistant professors and doctoral students. However, across samples, differences in COVID-19 related concerns about career advancement did not emerge based on gender, parental caregiving, sexual identity, or first-generation status.

Estimated marginal means of reported change in concern about career advancement due to the COVID-19 pandemic for the full sample (3a) and doctoral student (3b) and postdoc/asst prof (3c) subsamples. Participants were asked to rate the degree to which their concern about career advancement changed on a Likert scale from 1 = Greatly decreased to 5 = Greatly increased . Error bars represent the 95% CI. Numeric values in each bar are the estimated marginal mean and standard error. * indicates p < .05, ** indicates p < .01, and *** indicates p < .001. See S5 Table in S1 File for results of each model.

https://doi.org/10.1371/journal.pone.0274278.g003

We found subsample differences in concern about career advancement due to the pandemic by race. In the full sample, URM participants reported a greater increase in concern about their career advancement than White participants; in the doctoral student subsample, Asian participants reported greater increase in concern than White individuals; and for the postdoc/asst prof subsample, URM participants were significantly more concerned about career advancement than both White and Asian participants.

Change in support from mentors ( Fig 4 ).

Nearly all groups reported decreased mentor support, although means remained close to the scale midpoint. Across the three samples, those with a disability reported a significantly greater decrease in support from their mentors due to the pandemic compared to those without a disability. Further, assistant professors reported a significantly greater decrease in support from their mentors due to the pandemic compared to doctoral students and postdoctoral scholars. However, we did not find COVID-19 related differences in support from mentors by gender, sexual identity, or first-generation status.

Estimated marginal means of reported change in mentor support due to the COVID-19 pandemic for the full sample (4a) and the doctoral student (4b) and postdoc/asst prof (4c) subsamples. Participants were asked to rate the degree to which their support from mentor(s) changed on a Likert scale from 1 = Greatly decreased to 5 = Greatly increased . Error bars represent the 95% CI. Numeric values in each bar are the estimated marginal mean and standard error. * indicates p < .05, ** indicates p < .01, and *** indicates p < .001. See S6 Table in S1 File for results of each model.

https://doi.org/10.1371/journal.pone.0274278.g004

We found subsample differences in support from mentors by parental caregiving and race. For the full sample and the postdoc/asst prof subsample, primary caregivers reported significantly greater decrease in mentor support than non-primary caregivers and non-parents; no differences were found among doctoral students. For the full sample and the doctoral subsample, we found that White participants reported a significantly greater decrease in support from their mentors due to the pandemic compared to both URM and Asian participants; no differences in mentoring support by race were found among the postdoc/asst prof subsample.

Work disruptions due to COVID-19 life challenges

Work disruptions due to physical health problems ( fig 5 )..

There were differences in the extent to which work was disrupted because of COVID-19 related physical health problems for many socio-demographic statuses. Across the three samples, women and non-binary individuals and those with a disability reported significantly greater work disruptions due to physical health problems compared to men and those without a disability, respectively. Further, doctoral students reported greater work disruptions due to physical health problems compared to postdoctoral scholars. Across samples, we did not observe differences in work disruptions due to physical health problems by sexual identity.

Estimated marginal means of disruptions to work due to physical health problems during the COVID-19 pandemic for the full sample (5a) and the doctoral student (5b) and postdoc/asst prof (5c) subsamples. Participants were asked to rate the degree to which physical health problems due to the COVID-19 pandemic disrupted their work with a Likert scale from 1 = Did not disrupt my work to 5 = Greatly disrupted my work . Error bars represent the 95% CI. Numeric values in each bar are the estimated marginal mean and standard error. * indicates p < .05, ** indicates p < .01, and *** indicates p < .001. See S7 Table in S1 File for results of each model.

https://doi.org/10.1371/journal.pone.0274278.g005

When examining the subsamples, we found several differences in work disruptions due to physical health problems. For the postdoc/asst prof subsample, we found that primary caregivers reported greater work disruptions due to physical health problems than non-primary caregivers and non-parents. For the full sample and the doctoral student subsample, we observed differences in physical health problems by race such that URM participants reported greater work disruptions than both White and Asian participants. Finally, for the full sample and doctoral student subsample, there were greater work disruptions due to physical health problems for first-generation scholars than non-first-generation scholars; these differences did not emerge in the postdoc/asst prof subsample.

Work disruptions due to mental health problems ( Fig 6 ).

Across the three samples, those with a disability and sexual minority participants reported a greater disruption to their work due to mental health problems compared to those without a disability and heterosexual individuals, respectively. Doctoral students reported a greater disruption to their work due to mental health problems compared to both postdoctoral scholars and assistant professors. Across samples, no differences in work disruptions due to mental health problems were found by first-generation status.

Estimated marginal means of disruptions to work due to mental health symptoms during the COVID-19 pandemic for the full sample (6a) and the doctoral student (6b) and postdoc/asst prof (6c) subsamples. Participants were asked to rate the degree to which mental health problems due to the COVID-19 pandemic disrupted their work with a Likert scale from 1 = Did not disrupt my work to 5 = Greatly disrupted my work . Error bars represent the 95% CI. Numeric values in each bar are the estimated marginal mean and standard error. * indicates p < .05, ** indicates p < .01, and *** indicates p < .001. See S8 Table in S1 File for results of each model.

https://doi.org/10.1371/journal.pone.0274278.g006

We found differences in work disruptions caused by mental health problems by gender, parental caregiving, and race. For the full sample and the postdoc/asst prof subsample, primary caregivers reported greater disruption to their work due to mental health problems compared to non-primary caregivers. In the full sample and doctoral student subsample, URM participants and women and non-binary participants reported more work disruptions due to mental health problems than Asian participants and men, respectively; these differences did not emerge in the postdoc/asst prof subsample.

Work disruptions due to increased caregiving responsibilities ( Fig 7 ).

Across the three samples, parents with primary caregiving responsibilities reported more work disruptions due to increased caregiving responsibilities than non-primary caregivers and non-parents; non-primary caregivers reported more work disruptions due to increased caregiving responsibilities than non-parents. Participants with a disability reported greater work disruptions due to additional caretaking responsibilities compared to those without a disability. Doctoral students and assistant professors also reported more work disruptions due to additional caretaking responsibilities than postdoctoral scholars. However, no differences in work disruptions due to increased caregiving were found by gender or sexual identity.

Estimated marginal means of disruptions to work due to additional caretaking responsibilities during the COVID-19 pandemic for the full sample (7a) and doctoral student (7b) and postdoc/asst prof (7c) subsamples. Participants were asked to rate the degree to which additional caretaking responsibilities due to the COVID-19 pandemic disrupted their work with a Likert scale from 1 = Did not disrupt my work to 5 = Greatly disrupted my work . Error bars represent the 95% CI. Numeric values in each bar are the estimated marginal mean and standard error. * indicates p < .05, ** indicates p < .01, and *** indicates p < .001. See supplemental Table S9 for results of each model.

https://doi.org/10.1371/journal.pone.0274278.g007

We found subsample differences in work disruptions due to additional caretaking by race and first-generation status. For both the full sample and the doctoral subsample, URM and Asian participants reported greater work disruptions due to additional caretaking responsibilities compared to White participants; however, for the postdoc/asst prof subsample, URM participants reported greater work disruptions due to additional caretaking responsibilities compared to White and Asian participants. When using the full sample and the doctoral student subsample, we found that first-generation scholars reported more work disruptions due to additional caretaking than non-first-generation participants.

Our results (summarized in Tables 1 and 2 ) indicate a wide array of consequences of living and working through the COVID-19 pandemic for early-career scientists. These consequences are not evenly shouldered across academic scholars. Consistent with our hypothesis, individuals from marginalized groups reported more work disruptions due to COVID-19 than those from more privileged groups. Further, important differences were observed among scholars with different socio-demographic statuses. Changes in work outcomes (e.g., research progress) most often differed by parental caregiving, race, disability status, and career stage, whereas work disruptions due to COVID-19 related life challenges (e.g., physical health problems) were seen across most socio-demographic groups. We describe these patterns below and offer recommendations for support mechanisms for marginalized early-career scientists.

https://doi.org/10.1371/journal.pone.0274278.t001

https://doi.org/10.1371/journal.pone.0274278.t002

The pandemic’s impact was particularly notable for early-career scientists with disabilities. Specifically, those who reported a physical or mental disability were significantly more likely to experience all seven types of work disruptions examined in the study. There has been limited systematic research on the experiences of scientists with at least one disability [ 22 ]. However, our results highlight the importance of understanding their experiences. Many scholars with disabilities were not provided with appropriate accommodations prior to the pandemic, and have often been left out of the narrative about shifting to work-from-home and remote teaching throughout the COVID-19 pandemic [ 22 ]. Our findings further support existing discussions of the need for mental health resources, mentorship, and material support to increase the inclusion of academics with mental and physical disabilities [ 23 ].

Generally, URM participants reported more negative changes in work outcomes than other racial groups, extending prior findings of increased COVID-related stress and burnout among faculty of color [ 3 , 16 – 19 ]. URM scholars in our sample, compared to other racial groups, reported a greater increase in workload, growing concern about career advancement, and more work disruptions from COVID-related physical health problems, mental health problems, and increased caretaking responsibilities. However, in the full sample White participants reported the greatest decrease in mentor support and, similar to URM scholars, White participants reported a greater increase in their workload compared to Asians. We postulate that mentors with more limited capacity to provide support because of the pandemic may have directed that support towards mentees of color (URM and Asian) rather than White mentees.

Our findings about the effects of COVID-19 on Asian participants’ work outcomes were sometimes more similar to URM participants and sometimes more similar to White participants. For example, similar to URM doctoral students, Asian doctoral students reported more work disruptions due to additional caretaking responsibilities than White students, and less of a decrease in support from mentors. However, similar to White doctoral students, Asian students reported fewer work disruptions due to physical health problems than URM scholars. The complexity of the findings for Asian scholars may reflect that in STEM, this group can experience marginalization even when they are not underrepresented and suggests that more research focused on their COVID-related experiences is warranted. More nuanced work on Asian scholars’ COVID experiences in academia is particularly important given the increase in anti-Asian racism and violence that grew in the United States in 2020 and 2021 in part due to the COVID-19 pandemic [ 24 , 25 ].

Furthermore, we found some subsample differences in COVID-19 outcomes related to race. Most often, we observed similar findings in the overall sample and the doctoral student subsample that did not emerge in the postdoc/asst prof subsample; this was the case for workload, mentor support, and work disruptions from both physical health and mental health symptoms. These differences may be due to a lack of power to detect differences due to the smaller numbers of Asian and URM participants. However, in some cases, the pattern of means differed across subsamples, suggesting that the effect of racial marginalization may differ at the doctoral stage as compared to later career stages, perhaps due to career-stage changes in demographic representation.

A few differences in work outcomes emerged by sexual identity and first-generation status. In particular, first-generation college students who were postdocs or assistant professors reported a greater increase in their workload compared to non-first-generation scholars at the same career stage. Further, although no other differences emerged in work impacts by first-generation status (e.g., research progress), first-generation doctoral students experienced more work disruptions due to physical health problems and additional caregiving responsibilities. These findings suggest that more research on first-generation scholars is warranted in order to better understand the impact of COVID-19 on their careers over time. Sexual minority participants reported more disruptions due to mental health problems. This last result is consistent with a recent study of LGBT+ college students that found nearly 60% of the students sampled experienced increased anxiety and distress during the COVID-19 pandemic, notably associated with lower levels of family support [ 26 ].

Notably, our findings that women and gender non-binary participants reported greater work disruptions due to physical and mental health problems than men are consistent with prior research findings on the greater stress and burnout reported by women during the pandemic [ 27 ]. However, although early research found strong gender differences in productivity and other COVID-19 work impacts [ 3 , 9 – 12 ], our results showed less consistent patterns in these areas. We found that compared to men, women and gender non-binary doctoral students reported a greater increase in their workload due to COVID-19, but no differences emerged in research progress, concern about career advancement, or support from mentors. We may have found fewer gender differences in work disruptions than other studies in the literature because we examined differences in subjective perceptions of COVID-19 impacts (e.g., research progress) as opposed to objective measures (e.g., number of publications). In making their subjective assessments of impacts like research progress, participants may have compared themselves to others in their group (e.g., women compared themselves to other women and men to other men) facing similar types of COVID-19-related challenges. This comparison might have led women and gender non-binary individuals to minimize the effect of COVID-19 on their work.

We found that participants with more parental caregiving demands reported significantly greater decreases in research progress and greater work disruptions that they attributed to additional caregiving responsibilities. For the postdoc/asst prof subsample, these greater caregiving demands were also associated with less support from mentors, and greater work disruptions due to physical and mental health problems. These results underscore findings from other research indicating that it has been challenging for parents to navigate workplace responsibilities during the pandemic [ 21 ]. Differences by parental caregiving were more likely to emerge for the postdoc/asst prof subsample than the doctoral subsample, likely because our sample contained a greater percentage of assistant professors who were parents (54%) than postdocs (21%) or doctoral students (6%) who were parents.

Our research extends previous work documenting the challenges faced by assistant professors during the pandemic by comparing this group to other early-career scholars (i.e., doctoral students and postdoctoral scholars), finding differences across all seven work outcomes examined. Compared to doctoral students and/or postdocs, assistant professors reported the greatest decrease in their research progress and support from mentors and the largest increases in workload and work disruptions due to additional caretaking responsibilities. These results extend previous research that demonstrated disproportionately high impacts of COVID-19 on early-career faculty compared to mid- and later-career faculty members which were theorized to reflect a “perfect storm” of demands on tenure-track assistant professors–pressure for productivity to achieve tenure, the need for mentoring to support them in a new career stage, and the likelihood of having young children [ 27 ].

Work disruptions due to physical and mental health problems during the pandemic were particularly acute for doctoral students, likely aggravating the mental health crisis among graduate students that existed before the pandemic [ 28 ]. Although postdoctoral scholars reported lesser increases in workload and work disruptions due to caretaking responsibilities than other groups, postdocs were the most concerned about their career advancement, perhaps due to pandemic-related hiring freezes and postponements. Taken together, these results point to the likely long-term negative effects of COVID-19 on academic science, particularly if these work disruptions are related to career intentions.

Limitations

The data presented in this paper were part of a study about early career experiences in academia. Although an important dataset, we recognize some specific limitations. First, given the time of data collection (April-May of 2021), validated measures of COVID-19 impacts had not been developed. Therefore, we included only single-item measures of COVID-19 impacts on research and academic work. Second, we measured respondents’ subjective experiences rather than objective measures such as the number of manuscripts submitted to journals [ 1 – 3 ]. However, measures of subjective experiences at work and school are critical, as prior literature indicates that they strongly predict organizational satisfaction and intentions to remain on a science career path [ 29 , 30 ].

Third, there are limitations related to the relatively small number of participants within particular career stages or socio-demographic groups. Although response rates were robust, our small sample of postdoctoral scholars and assistant professors required us to pool these two career stages in the subsample analyses rather than analyze them separately. Even combined, we may not have had enough power to detect some small but important effects in this subsample. Further, although we capture differences by socio-demographic status such as race and gender, our findings do not speak to the unique intersectional impacts of COVID-19 for those occupying multiple marginalized and minoritized statuses, such as Black women or sexual minority caregivers, for whom multiple systems of inequality may create additional and unique demands and challenges [ 31 ].

Finally, our project examines scholars’ experiences in four fields that span the spectrum of life, physical, and social sciences (biology, economics, physics, and psychology). Although our data do not speak directly to the outcomes of the COVID-19 pandemic beyond these fields, other research has identified that scholars in the arts and humanities have faced significant, long-term barriers to their scholarly work due to the pandemic [ 3 ].

Implications and recommendations

An understanding of the breadth of COVID-19 impacts and their differential consequences for scholars across socio-demographically marginalized communities is critical for university administrators and others to determine the best way to respond equitably to pandemic-related disruptions. Our results improve understanding of how inequalities are reproduced in the academy and point to career stages and socio-demographic groups that may require additional support to recover from the pandemic professionally. Our findings suggest the need for longitudinal studies that track the long-term impacts of COVID-19 on career outcomes of early-career scholars, particularly for scholars from marginalized groups. Our supplemental analyses (see S12-S16 Tables in S1 File ) indicated that COVID-19 disruptions were associated with negative work outcomes; thus, the socio-demographic differences in COVID-19 disruptions that we observed may result in reduced job satisfaction and professional role confidence while increasing turnover intentions and burnout for minoritized and marginalized early-career academics. Without data-driven policies and structural supports, academic science risks a significant and long-term loss of diversity.

To limit the loss of early-career scholars from marginalized and minoritized groups, it is crucial that policies be adopted that provide institutional support to those affected most significantly by the pandemic. Although the acute challenges of the pandemic will likely diminish over the coming years, COVID-19 impacts may be a sort of “canary in the coal mine,” highlighting areas where social crises (e.g., COVID-19 pandemic, racial injustice, #MeTooPhD) amplify already existing inequalities for academics from marginalized groups. For example, the greater work disruption due to physical health problems we found for women and non-binary individuals, primary caregiver postdoctoral scholars and assistant professors, URM individuals, and those with disabilities, may indicate that the academic workload was already disproportionally distributed, with more of the burden–likely including invisible service and other unrecognized labor–being carried by individuals in these groups [ 32 ]. Thus, the following recommendations are both specific to addressing the negative impact of COVID-19 and broadly aimed at creating a more inclusive academic science context.

The pandemic has illustrated that certain practices, such as remote options for meetings, conferences, and professional development opportunities, increase accessibility for individuals with disabilities and/or caregiving responsibilities. Similarly, recording talks or webinars creates flexibility for those balancing many competing responsibilities. A study of disabled workers prior to the pandemic found that only 5% of respondents were able to work from home, despite remote work often being preferable for individuals with disabilities [ 33 ]. The positive impact of certain remote working practices challenges academia to reconsider a “return-to-normal,” and instead create a more inclusive working environment by maintaining multi-modal forms of participation in conferences, talks, and networking events for early-career researchers, as well as opportunities for remote and hybrid work.

Some of the findings of our research pointed to the perceived loss of mentoring support concurrent with an increase in mental and physical health problems for early-career scientists. Therefore, providing additional mentor support during the next few years, especially for assistant professors who have lost much of this support during the pandemic, could be an important step for ensuring their success in academia. At the same time, it is also important to recognize the range of other service commitments placed on mentors, especially those from underrepresented groups, and provide them with the time and resources needed to support early-career scholars. Additionally, institutions need to dramatically increase their investment in mental and physical health resources, particularly for the support of doctoral students, gender and sexual minority scholars, primary caregivers, and persons with disabilities.

Our findings of reduced research progress and accompanying career-related concerns suggest the need for institutions to account for pandemic-related work challenges in institutional rewards systems. Universities can provide resources, including research funds, along with opportunities for faculty, postdoctoral scholars, and doctoral students to document the impacts of the pandemic on their work so that steps can be taken to alleviate those impacts [ 34 ]. Moreover, institutions will need to thoughtfully consider how to adjust their evaluations (e.g., annual reviews, promotion and tenure, and hiring processes) to account for differences in academic records due to COVID-19’s disparate impact on marginalized communities [ 6 , 8 , 35 ]. These include holistic review practices that weigh quality over quantity and review criteria that account for the visible and invisible service and emotional labor that is often performed by scholars from marginalized groups. These steps are urgently needed to maintain the future strength and diversity of the STEM workforce.

Supporting information

https://doi.org/10.1371/journal.pone.0274278.s001

Acknowledgments

The authors would like to thank Dr. Petal Grower for her feedback and revisions on this manuscript as well as her contributions to the CLIMBS UP team. We would also like to thank Alexander D. Garcia-Settles for his assistance preparing tables for the manuscript.

• 1. Flaherty C. COVID-19: A Moment for Women in STEM? Inside Higher Ed; 2021. Available from: https://www.insidehighered.com/news/2021/03/10/covid-19-moment-women-stem
• 2. National Academies of Sciences, Engineering, and Medicine. The Impact of COVID-19 on the Careers of Women in Academic Sciences, Engineering, and Medicine. The National Academies Press; 2021. Available from: https://nap.nationalacademies.org/catalog/26061/the-impact-of-covid-19-on-the-careers-of-women-in-academic-sciences-engineering-and-medicine https://doi.org/10.17226/26061
• View Article
• Google Scholar
• PubMed/NCBI
• 7. Goodwin SA, Mitchneck B. STEM Equity and Inclusion (Un)Interrupted? Inside Higher Ed. 2020. Available from: https://www.insidehighered.com/views/2020/05/13/ensuring-pandemic-doesnt-negatively-impact-women-stem-especially-those-color
• 14. Mickey EL, Clark D, Misra J. Measures to Support Faculty During COVID-19. Inside Higher Ed. 2020. Available from: https://www.insidehighered.com/advice/2020/09/04/advice-academic-administrators-how-best-support-faculty-during-pandemic-opinion
• 17. Pettit E. Faculty Members Are Suffering Burnout. These Strategies Could Help. The Chronicle of Higher Education. 2021. Available from: https://www.chronicle.com/article/faculty-members-are-suffering-burnout-so-some-colleges-have-used-these-strategies-to-help
• 18. Tugend A. “On the Verge of Burnout”: Covid19’s impact on faculty well-being and career plans. The Chronicle of Higher Education. 2020. Available from: https://connect.chronicle.com/rs/931-EKA-218/images/Covid%26FacultyCareerPaths_Fidelity_ResearchBrief_v3%20%281%29.pdf
• 20. Chirikov I, Soria KM, Horgos B, Jones-White D. Undergraduate and Graduate Students’ Mental Health During the COVID-19 Pandemic. UC Berkeley: Center for Studies in Higher Education. 2020. Available from: https://escholarship.org/uc/item/80k5d5hw
• 33. Give disabled people the right to work from home after Covid-19, says UNISON. UNISON. 2020. Available from: https://www.unison.org.uk/news/press-release/2020/08/give-disabled-people-right-work-home-covid-19-says-unison/

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 23 September 2023

Unraveling the controversial effect of Covid-19 on college students’ performance

• Luca Bonacini 1 ,
• Giovanni Gallo 2 &
• Fabrizio Patriarca 2  

Scientific Reports volume  13 , Article number:  15912 ( 2023 ) Cite this article

4147 Accesses

4 Citations

3 Altmetric

Metrics details

• Epidemiology
• Health care economics
• Social evolution

We disentangle the channels through which Covid-19 has affected the performance of university students by setting up an econometric strategy to identify separately changes in both teaching and evaluation modes, and the short and long term effects of mobility restrictions. We exploit full and detailed information from the administrative archives of one among the first universities to be shut down since the virus spread from Wuhan. The results help solving the inconsistencies in the literature by providing evidence of a composite picture where negative effects such as those caused by the sudden shift to remote learning and by the exposure to mobility restrictions, overlap to opposite effects due to a change in evaluation methods and home confinement during the exam’s preparation. Such overlap of conflicting effects, weakening the signaling role of tertiary education, would add to the learning loss by further exacerbating future consequences on the “Covid” generation.

Similar content being viewed by others

Schooling and Covid-19: lessons from recent research on EdTech

Heterogeneous peer effects of college roommates on academic performance

Grade inflation effects of capacity expansion in higher education: a longitudinal study in undergraduate teacher education programs from 2003 to 2022

Introduction.

There is a wide and varied literature stressing how the pandemic crisis has harmed the accumulation of human capital. In this article we focus on tertiary education and in particular on students’ performance. While in the case of primary and secondary education the literature converges on the emergence of a consistent learning deficit 1 , in the case of tertiary education the picture is much more controversial. To untangle the knot it is important to consider the variety of channels through which the pandemic might have affected students’ outcomes. Indeed, together with channels that affected all the population, both directly on health and indirectly through containment and lockdown measures, in the case of college education there are specific channels related to the shutdown of in presence activities as the sudden shift to remote learning, the temporary return of students to their places of origin and the change in student assessment methods that also shifted to online mode. Each channel has had impacts on different aspects of students’ careers, with different intensities and even in opposite directions. In our opinion, this composite picture helps explaining the lack of uniqueness of the evidence provided by the related literature developed so far.

Separating the overall effects between different channels requires very detailed data as to implement satisfactory econometric strategies to go beyond the identification of the overall effect based on simple comparison of pre- and post-pandemic values. For this purpose, we use the administrative data of one among the first Universities directly involved in the spread of the virus outside China: the University of Modena and Reggio Emilia. We can track between 2018 and 2021 a total of about 38,000 students, who have taken about 400,000 exams, with high-level details on the characteristics of examinations, study paths, background of students and teachers fixed effects.

By exploiting this rich dataset, we build an econometric strategy based on difference-in-differences estimations 2 to analyze the exams marks by distinguishing between the contrasting effects of the change in teaching and in assessment modes, and then consider separately the effects of exposure to lockdown measures.

On the one hand, while the transition to distance learning may have had a negative impact on learning, as confirmed by the literature on lower levels of education 3 , 4 , the need to change the assessment method may have had an opposite effect on measured performance. Indeed, since the shift to online exams made more difficult to avoid plagiarism or other misconduct 5 , this might have incentivized students to cheat. Furthermore, the exams mode itself (e.g. alone or in the classroom, with interviews or quizzes) may have affected students’ performance during the exams, and finally also teachers evaluation attitudes could have become less stringent. To solve the possible overlap of contrasting effects and correct for the possible divergent dynamics of actual and measured students’ performance, we exploit the pre-existence of courses where classes were already given, though partially, in remote mode even before the pandemic, although exams mode were the same as for the other courses. In this way we can build a difference-in-differences identification strategy exploiting the heterogeneity related to the fact that the extent shift of teaching mode has been different though the change in assessment mode has been the same.

On the other one hand, we use the information on the exam date as to take into account the effect of lockdown measures. This information allows us to build a proxy of exposure to restriction which is both time and space varying, by matching the data on the pattern of restrictions in Italian regions. Indeed, the prolonged closure of a university with a supra-regional students pool, located in an area with a relatively high cost of living, has led the majority of students to return to their homes. This led to (exogenous) different exposures to containment measures, since they had a predominantly regional character.

The results also give us a composite evidence that help us explain the puzzled results found in the literature about the effects of lockdown measures: while the overall exposure to containment measures appears to have a significant negative impact on students, being confined at home during the preparation of the exams turns out instead to have had a positive effect.

In the next sections, after a review of the related literature, we lay out a description of the case study and of the data used. Next, we present the econometric strategy and then discuss the results. Before concluding, in the final session we also perform some robustness checks.

Tertiary education and the pandemic

While the socio-economic consequences of the Covid-19 have been already studied in deep from many points of views, papers focused on the impact of the pandemic on higher education are still few and provide contradictory results. We can split this branch of literature into two groups of studies: those using pupils’ surveys 6 , 7 , 8 and those considering data on students’ actual outcomes 9 , 10 , 11 , 12 . Overall, the first ones find negative effects of the pandemic, while the second ones mostly agree on the contrary.

A pioneering contribution is provided by 6 , which surveyed approximately 1500 students at one of the largest public institutions in the United States. To our knowledge, their analysis is the first trying to get the impact of the pandemic on students’ outcomes. Results show large negative effects. Due to Covid-19, 13% of students have delayed graduation, 40% have lost a job, internship, or job offer, and 29% expect to earn less at age 35. Moreover, these effects have been highly heterogeneous: one quarter of students increased their study time by more than 4 weekly hours due to Covid-19, while another quarter decreased their study time by more than 5 h per week. This heterogeneity often followed existing socioeconomic divides. Lower-income students are 55% more likely than their higher-income peers to have delayed graduation due to Covid-19.

In the same spirit 8 , conducted an online survey on 3163 Queens College students during the summer 2020. She analyses the effect of the Covid-19 outbreak on current and expected outcomes through an estimation of individual-level subjective treatment effects. She finds that due to the pandemic, between 14 and 34% of students considered to drop-out, as they think to lose their financial assistance, or to postpone their graduation. The pandemic also deprived 39% of students of their jobs and reduced their earnings by 35%. Finally, her analysis also reveals that the effect of the pandemic on social insecurity has been different on the basis of the students’ well-being as it has been deeper for students with a federal Pell grant than their peers.

Hu et al. 7 make a contribution to the analysis on students’ self-perception as they differentiate their analysis to the previous ones asking about students’ conditions two years later since the outbreak of the pandemic, in the period between January 17 to February 25, 2022. They surveyed 151 college students in Northern Michigan asking how much their learning quality is influenced by the Covid-19 and they conclude that respondents’ education was severely affected by the pandemic, averaging a score of 7.58 on a scale of 10. These results suggest that the negative impact of Covid-19 on students’ self-perception is not limited to the short run.

Contrasting results are provided instead by the second stream of literature as in 10 , 11 , 12 . Gonzalez et al. 10 analyze the effects of Covid-19 confinement on the autonomous learning performance of students in higher education through a sort of randomized control experiment. Their study relies on a field experiment with 458 students at Universidad Autonoma de Madrid. The control group corresponds to academic years 2017/2018 and 2018/2019. The experimental group comprehends students from 2019/2020. The results show a significant positive effect of the Covid-19 confinement on students’ performance as they changed their learning strategies to a more continuous habit. Similar results hold in 11 . They estimate the effects of online education during the Covid-19 lockdown on student performance through a difference-in-differences approach using administrative data from Chinese Middle Schools. They consider three schools in the same county in Baise City before and after the Covid-19 onset. School A is the control group, as it did not provide any online educational support to its students. School B and C (treatment group) used an online platform. They point out a positive effect of online education by 0.22 of a standard deviation on student academic results. They also found that the results are homogeneous between rural and urban students.

Other contributions mainly focus on the heterogeneity of the effect across groups, but even in none of these we can find an evidence of a decrease in overall performance. Rodríguez-Planas 8 uses an event study approach to compare the gap between low-income students and their peers in the same University. She concludes that lower-income students with a lower performance during the pre-pandemic period outperformed their higher-income peers thanks to the different use of the flexible grading policy based on their financial and academic needs. In contrast, in the absence of the flexible grading policy, lower-income top-performing students would have underperformed relative to their higher-income counterparts. Engelhardt et al. 13 compare university students’ performance in the first semester affected by Covid-19 to that of the previous three ones. They do not find significant differences in performance across periods. These results are confirmed also for low-income, first-generation, and minority students. Castellanos-Serrano et al. 14 focus on the academic consequences of the Covid-19 in gender inequalities by several education performances. They consider 7477 students enrolled in just one faculty from the 2016/2017 to 2020/2021 academic years. Using a basic pre-post identification strategy, they find heterogeneous effects of the pandemic by sex since women’s results worsened in comparison to those of the pre-covid-19 period to a greater extent than for men. Besides, all sex slightly improved their results over the pandemic period. Maldonado and De Witte 15 consider the last year of primary schools in the Dutch-speaking Flemish region of Belgium. Using a 6-year panel, they perform a linear regression model with a pre-post Covid variable and find that, on average, students of the 2020 cohort experienced significant learning losses. Moreover, inequality within and across schools increased as a result of the Covid-19 crisis. Altindag et al. 16 leverage data from 15,000 students enrolled in a U.S. public university to investigate the performance of students in in-person compared to online courses during the pandemic. Using a student fixed effects model, the authors find that students in in-person courses fared better than online students with respect to their grades, the propensity to withdraw from the course, and the likelihood of receiving a passing grade. Agostinelli et al. 17 decompose the potential channels operating through the online learning, peers interactions, and the time spent with the parents. They conclude that each of these channels contribute to higher educational inequality during the pandemic.

All these studies target at the overall impact of Covid-19. Differently, Bird et al. 9 focus on the specific impact of the pandemic-triggered shift to online education. To do that they use data on students attending Virginia’s community colleges and set up an econometric strategy partially similar to that of part of our analysis: they use a difference-in-differences strategy in which the treatment groups is composed by the students enrolled in an in-person course and the control group is composed by the students which the course was provided online also before the Covid-19 widespread. Differently to the present contribution, their primary outcome of interest is the course completion, namely a binary variable equal to one whether the student received any grade sufficient (A, B, C, D, P + , or P), zero otherwise. The authors find that the shift to the online modality led to a modest decrease in course completion between 3 and 6%. This reduction in course completion is primarily driven by a large increase in course withdrawals (37% or + 2.7 percentage points in absolute terms) and, more narrowly, by an increase in course failure (10.8% or + 1.3 percentage points).

It is thus worth to notice that by focusing on a specific channel of the impact of the Covid-19 period, results shows a different picture than the one offered by the aggregate evidence. Delving deeper in this direction, in this paper we will try to solve the apparent puzzle. Our basic hypothesis is that the coexistence of negative effects reported subjectively or detected in the analysis of specific channels, together with positive effects resulting from the analysis of the overall outcomes is mainly due to the coexistence of positive effects on reported performance due to a change in evaluation standards, and negative effects on actual performance.

The case study

The case study is the University of Modena and Reggio Emilia. Unimore is a medium-sized Italian university, with a wide range of fields organized in 12 departments, ranked in the middle among Italian high education institutions, with a predominantly regional and national enrollment pool. As we will see in the econometric strategy session, this last characteristic together with the peculiarities of the relationship with pandemic events will be valuable for the purpose of the identification strategy we will use in this study. A final feature of the case study, that we will exploit in “ Econometric strategy ” section, is that a significant share of Unimore’s departments, before the pandemic, already offered degree programs where each single course provides mixed in-presence and remote classes.

At the same time, the university has recently undertaken a process of integrating all micro-data from administrative sources or interviews into a single database, Unimoredata, which enable us to analyze with a very high level of detail the performance trends of its students along the period of interest.

The pandemic at Unimore

On 21 February 2020 the Coronavirus had just begun to spread outside China and the first outbreaks of the virus were detected in the North-East of Italy. Two days later, on February 23, due to the dynamics of the virus in the neighborhood, the Emilia-Romagna Region imposed a four days closure of the activities to all the universities in its territory, thus including Unimore. This has been the first restrictive measure involving educational institutions, which will anticipate all other restrictive measures, including the first large-scale red zone, the one that the following week was imposed to the territory of the Modena province (i.e. the Italian name for the NUTS-3 region level). Indeed, since the virus spread over, the next week lessons did not turn back to in presence and the closing measures were instead extended to all Italian Universities as early as March 4, according to restrictive measures that will last until the summer.

In the Italian university system, the yearly activity is divided into two semesters, with lessons taking place from late September to December for the first semester and from late February up to the end of May in the second one. Consequently, the closure of the in-presence activities at Unimore, coincides exactly with the beginning of the second semester of the academic year 2019/2020. As a result, the shift toward remote learning at Unimore, unlike in the case of the other universities, has completely covered the semester affected by the first stage of the pandemic.

After the first wave of the virus, most Italian universities opted for solutions allowing at least a partial resumption of in-presence activities for the following semester. Unimore, instead, adopted a very restrictive policy announcing already in May 2020 that the activities would have remained in remote for all the first semester of the following academic year (i.e. 2020/2021) and that it would have been possible to attend the lessons remotely in the second semester of the following academic year independently from the evolution of the pandemic. The lessons turned back in presence only at the end of the second semester of the academic year 2020/2021and only for the first-year students. The latter decision, taken in December 2020, was driven by the fact that a second wave of Covid-19 contagions was in place during that period and a third wave was largely expected for the successive months. In Italy, to be noted, the first wave of Covid-19 contagions took roughly place from February to May 2020, the second wave from October to December 2020, and the third wave from February to April 2021.

Following the timing of the main waves of coronavirus contagions, the pandemic period can be split in three different sub-periods in the Unimore context. The first one arrives up to September 2020 and corresponds to the first wave of contagions, the complete shift of the University activities to remote mode, and to the national restrictive measures. The second period, from October 2020 to March 2021, was characterized by the fact that Unimore was still closed and lockdown measures took a regional level dimension using a four colors classification. According to this new mechanism, the tightening of restrictive measures was based on a set of indicators at the regional level—mostly related to pressures of Covid-19 contagions on the healthcare system—which distinguished white, yellow, orange and red zones.. The third period, from April 2021 onwards, was instead characterized by a partial return to in-presence activities at Unimore thanks to a progressive loosening of social distancing measures and the massive vaccination campaign.

As for the scheduling of exams, whose grades are the outcome variable we are going to consider, in line with the other Italian universities, Unimore provides three regular sessions of exams: the winter session, from the beginning of January up to the end of February; the summer session, spanning from the half of May to the end of July; and the fall session, from the end of August to the end of September. According to the specific course, there are also a number of cases where exams are held in extra-ordinary sessions (April to May and October to December). The first exams in the Covid-19 period are thus the ones in April 2020, the last exams of the first sub-period ends with the exams of the fall 2020 regular session, the second sub-period starts with the extra-ordinary sessions of October and December 2020, includes the 2021 winter session end finishes with the exams of the extraordinary session in spring 2021, the last period covers the regular sessions of summer and fall 2022.

The Unimore dataset

This study relies on Unimoredata, a database created with a specific Unimore project integrating all students’ individual information from administrative records and many large scale surveys (e.g. the Almalaurea post-degree surveys on early access to the labour market) since 2001.

Specifically, for the purpose of the presented analysis, we refer to a dataset merging together detailed information from the following administrative archives: (1) the register containing demographic characteristics of each student; (2) the archive reporting yearly information on each Unimore course attended by each student; and (3) the archive collecting all exams passed by each student attending Unimore. The latter dataset is particularly important for our analysis, as it contains full information about students’ passed exams, like the obtained mark, the date of notification, the subject, the teaching period, and teachers’ characteristics. According to the administrative data collection policies in Italian public Universities, failed exams are instead not recorded. Further investigation, however, have shown that during the pandemic the dynamics of passed exams had very a similar path to those of average exams marks which, as we will see below, have slightly increased. At same time, drop out rates increased by 2.1 percentage points, showing thus very similar patterns as those record elsewhere as in 9 .

The analysis focuses on the grades of passed exams held in the period ranging from January 2018 to September 2021, thus our reference period starts from more than two years before the pandemic and then covers all the period characterized by the first three and major waves of Covid-19. We decide to restrict the sample of analysis considering only students aged 18–36 years old. Despite students being 37 years old or more represent a clear minority group (about 2% of the sample), we choose to exclude them from the analysis because their peculiar characteristics makes overall unclear their condition during the pandemic (e.g. they may be employed in remote working or in layoff/furlough period). Due to similar reasons, we also drop from the sample those students who still haven’t held any exam one year after the standard end of the course (about 5% of the sample). We also drop the exams for which we miss information about the teacher since they correspond to courses taught by teachers who are recruited on annual contracts and thus normally change from year to year (about 9.5% of the sample). In conclusion, our analysis relies on a sample of 371.000 exams held and passed by about 38,000 students. A detailed description of all variables used in the analyses and main descriptive statistics on the sample of students are presented in the Supplementary Material (Supplementary Table S1 and Table S2 respectively).

In the second part of the analysis, we build a difference-in-differences (DID henceforth) identification strategy exploiting also the information about the courses held with mix modality of teaching. However, as the provision of such kind of courses is not common to all departments, we exclude from the sample of analysis all observations referring to departments where these course are not supplied. With this last sample restriction the second part of the analysis relies on about 230 thousand exams. Also the main descriptive statistics on this reduced sample of students are presented in the Supplementary Material (Supplementary Table S3 ).

Econometric strategy

The performance of students exams is analyzed by looking at the mark of each single exam as resulting from the administrative archives.

The benchmark model uses the following linear specification:

where \({y}_{j,i,t}\) is the mark obtained at the j exam of the student i at time t (if the student attends and passes the exam); \({X}_{i,t}\) and \({Z}_{j,t}\) are two vectors respectively of student level and exam level controls (some of them are time varying); \({m}_{t}\) is the month of the exam; \({C}_{t}\) is the dummy variable for the Covid-19 period, that is set alternatively as a single dummy or a set of dummies distinct by the 3 sub-periods outlined above, and \({\varepsilon }_{j,i,t}\) is the error term. The equation is estimated with linear OLS and errors are clustered at student level. The set of controls at student level \({X}_{i,t}\) includes: students’ demographic characteristics as gender, age, NUTS-3 region level region of birth and region of residence; the kind of upper secondary school attended before university (11 different categories); a dummy variable for being a sophomore or junior student and the number of exams already passed by the student at each exam date (i.e. proxies of students’ tenure and quality). The set of controls at the exam level \({Z}_{j,t}\) includes: the specific department of the degree program; a dummy for master degree courses (vs bachelor ones); the number of university credits (CFU) related to the exam; the exam month. To be clear, in the Italian system each exam correspond to an amount of credits varying from 3 to 12, and usually equal to 6 and 9; the greater is the number of credits the higher is the complexity and somewhat the difficulty of the exam. Formally, a CFU represents about 25 studying hours (in general assuming 7/8 h of lessons attendance and 17/18 h of ‘study at home’). A bachelor degree is generally reached after the completion of 180 CFU, while master degree courses count 120 CFU. Furthermore, we include in \({Z}_{j,t}\) also teachers individual fixed effects to account for this important source of heterogeneity, corresponding to 1160 dummy variables in the benchmark case.

In this benchmark model we thus focus on the coefficient \(\theta\) representing the overall impact of the pandemic, similarly to what most of the literature outlined above does. As we discussed above, this approach would catch the effect on measured performance rather than to actual one. Thus, once set up this base model, we move to assess an identification strategy aimed to disentangle the effect of the changing teaching (and thus learning) methodologies first, and then the effects of the exposure to restrictive measures.

Identifying the impact of (suddenly) changing teaching models

In this section we set up an econometric strategy to identify the impact of the shift from in-presence to remote learning brought about by social distancing measures. Thus, we are not going to evaluate the effectiveness of different teaching methodologies in normal times, we are instead analyzing the impact of a forced sudden shift that has also often caught unprepared teachers and technical staff.

As we have anticipated in previous sessions, possible negative effects on students’ actual performance could be overshadowed by opposing changes in measured performance related to changing examination modes. To avoid student misconduct, and in compliance with the general directives of the Italian Ministry of Education, Unimore adopted a set of arrangements to the remote examination modes that included student room control systems, software to control the activities of the personal computers used for examination tests (Safe Exam Browser), and limits to the ratio of examining students to teachers assigned to video surveillance. Such arrangements have reduced possible misbehavior however surely not eliminated it. At same time, the same shifts of exams modes with this related arrangements might have impacted on students’ performance during the exam. An analysis of the impacts on actual student performance, therefore, cannot disregard all this performance measurement problems. To this end, we will set up a DID identification strategy relying on the fact that while the shift in exams mode, with the related performance measuring biases, has equally concerned all courses, the change in teaching modalities has not been equal for all. Indeed, many Departments at Unimore, before the pandemic, already included in their supply degree programs with an hybrid online and in presence learning. In these programs all courses have only a share of teaching using traditional face to face methodology, and this share corresponds on average to the half of the course teaching activities with very little variation among courses. At same time, this courses have all same in-presence evaluation modes independently from the teaching modality. Thus, the shut off of in-presence activities had different consequences in terms of intensity in changing teaching modes among hybrid and standard courses but same consequences in terms of changes in evaluation modes and then also in performance measuring standards. In particular, we can argue without loss of generality that the impact in terms of changing teaching methodologies was double in the case of standard in-presence programs respect to hybrid ones.

We exploit this option in a DID approach by adding to the base specification in Eq. ( 1 ) the course modality variable and its interaction term with the Covid-19 variable:

where the variable \({D}_{j}\) is a dummy representing the course modality in normal times: in-presence ( \({D}_{j}=1\) ) or hybrid one. A negative sign of the interaction coefficient \(\pi\) would evidence a relatively worst performance for exams in traditional programs respect to those in hybrid ones and thus, according to the DID strategy interpretation of causal inference, supporting for a negative impact of the shift to distance teaching. Moreover, since the teaching modality shift is double for courses in standard programs respect to those in hybrid programs, in terms of magnitude we can state that the impact estimated is a lower bound estimation that should correspond to half of the actual impact.

In the sample of analysis, these hybrid courses represent around 19% of the students and the online teaching usually represent half of the classes for each exam. As students and exams could have different features in the two kind of programs, we correct for possible composition biases by using an Inverse Probability Weighting (IPW) strategy with the hybrid mode variable D as treatment variable. The IPW estimate relies on the following set of covariates: students’ demographic characteristics (i.e. gender, age, and NUTS-3 level region of birth and residence); the kind of upper secondary school; the year of enrollment; the specific department; and a dummy for master degree courses (vs bachelor ones). Finally, to properly isolate the effect of changing in teaching modes, we restrict the sample to the exams corresponding to classes taught in the immediately preceding teaching period (i.e. about 161 out of 223 thousand of exams). In fact, exams can be attended either in the months immediately following the end of classes but also in next semesters, several months after. We limit our analysis to the former case of ‘on-schedule’ exams. With this sample restriction we narrow the analysis on exams prepared by students attending courses taught according the modalities corresponding to the same specific period (before and after Covid-19 and also, in case, to the specific sub-periods). Moreover, by doing so, we can focus on exams whose preparation is more strictly related to the classes attendance rather than to the use of supplementary materials, such as handbooks or slides.

Identifying the impact of the exposure to restrictions

To identify the effect of exposure to restrictive and lockdown measures, we exploit the consequences of the very prudential policy implemented by Unimore about the recovery of normal activity described above.

As elsewhere, the closure of universities led to the return to their origin places of a large part of students being not resident in the neighborhood of the University. Suggestive is the case of Milan, where the news of the regional lockdown for the following day, circulated in advance because of a communication mistake, caused an exodus of students from North to South Italy so massive as to strongly impact on the spread of the virus in the southern regions of the Country while it was still concentrated only in the Northern regions.

During summer 2020, while Covid-19 related restrictions had been loosened by the national government, the universities were allowed to decide autonomously whether to re-start in presence activities for the next year. The decision in most campus or university cities contexts to reactivate in-presence activities, with the need to bear the cost of new infrastructure needed to respect legal prescriptions for social distancing, have also been driven by the economic interests of the neighborhood, for which the closure of the university leads to significant losses, like in the case of the owners of rental properties, commercial activities, and so on.

This was not actually the case for Unimore. In this area, the university has indeed a significant impact on its territory, but the economic vocation is another, ranging from automotive (Ferrari, Maserati, etc.) to food processing (e.g. Parmigiano Reggiano, Modena’s Balsamico), via robotics and ceramics. Moreover, the Modena city hospital, which was among those most put under pressure since the first waves of the pandemic, is part of the same university and has significant political weight even in the managerial offices (the same chancellor was a professor of the department of Medicine). As a consequence, the subjective experiences of professors and other civil servants grounded in departments operating within the Modena hospital understandably had a weight on their attitudes on the level of precautions to take.

As a result, Unimore adopted different decision respect to most universities, as the neighboring University of Bologna, which guaranteed a reopening of activities also through ad hoc investments for mixed teaching and the intervention of public institutions providing housing supports for students. Just before the end of the second semester of 2020, Unimore finally announced that the activities of the first semester of the following academic year—starting in September 2020—would have kept the distance mode. This exacerbates the emptying of the cities of Modena and Reggio Emilia, as evidenced by the attention given by the local press. Indeed, since then, also for the contribution of the very high living costs characterizing the cities of Modena and Reggio Emilia, most students returned to their homes and freshmen did not come in Modena and Reggio Emilia to find a new accommodation. This depletion is also confirmed by the fact that at the end of the second semester of the 2021/2022 academic year, when in-person attendance was reopened for a number of courses, despite the announcement made well in advance, only a minority share of students actually returned physically to the classroom while the rest continued to attend remotely. This decision did not turn out to be so wrong if one considers that the arrival of the second and third waves of the virus also induced the other universities to close down again.

At the same time, restrictive measures took a regional articulation from October 2020, following the four-color classification mentioned above. This induced a strong heterogeneity in students exposure to restrictions. The restrictions adopted in the case of red classification are similar to the lockdown implemented nationwide from March to May 2020, thus an overall home confinement. Accordingly, the time-varying restrictions in place at the residence of each student are a reliable proxy of the restrictions to which she has been subjected having a relevant time and space varying dimension. Figure  1 gives evidence of the regional heterogeneity of the cumulated restrictions from the beginning of the pandemic to September 2021, but the time-varying dimension of restrictions is relevant as well. To be noted, for the sake of the analysis, the national level lockdown imposed during the first wave of the virus, which lasted 70 days, is considered as a red zone and included on each regions’ records.

Cumulated number of days in red classified regional conditions.

We exploit this peculiarity to analyze two different aspects of the exposure to the restrictive measures. First, we consider the impact of cumulated exposition to restrictive measures since the start of pandemic. Second, we consider the effect of exposure to restrictive measures during the exam preparation period. To do all this, we add to the benchmark model in Eq. ( 1 ) one variable in two different cases. For each date of exam, in the first case we compute the cumulated number of days that the region of residence has passed under red zone restrictions while in the second one we compute the share of days in red zone over the 14 days before the exam. As we count among the days spent in a red zone also those related to the national level lockdown, when these variables still have a time-varying dimension and then allow for some heterogeneity, we can use all the data period from May 2020 onwards.

In detailing our findings, we start by providing an overall picture of students’ performance after and before the Covid-19 pandemic in Table 1 . In the first column we report estimates of the model specification presented in Eq. ( 1 ) and, in particular, the coefficient of the Covid-19 dummy variable being 1 for the whole period ranging from April 2020 to September 2021. The coefficient is positive and significant at 1% level.

In terms of magnitude, considering that exam marks at the Italian universities are expressed over 30 points with 18 being the minimum of passed exams and the standard deviation in the sample is 3.7 points, the value of 0.186 reported in the first column of Table 1 is not negligible although low. When we look at the three sub-periods of pandemic discussed in “ The pandemic at Unimore ” section separately, the coefficient is still positive and significant for each sub-period (second column of Table 1 ). The positive effect is concentrated in the first two periods of the pandemic, where the coefficient is a bit greater than 0.2. In the third period (i.e. April-September 2021), the coefficient becomes much lower but it remains still significant. The lower magnitude of the coefficient in the last period is consistent with the partial reopening of in-presence activities, which could blur the pandemic influence on the students’ performances. To account for this possible confounding factor, in the third column of Table 1 , we report the estimate of the overall impact limited to the first year of the pandemic only, thus limiting the reference period to April 2021 rather than September 2021. In this case the coefficient of the Covid-19 dummy variable has a value close to those reported in the first two sub-periods of pandemic.

In conclusion, the evidence provided in Table 1 would suggest that in relation to the sample of passed exams, students’ performance has slightly benefited from the pandemic, consistently with other studies of the literature surveyed above which use the same Covid-19 period dummy variable approach or else rely on some descriptive evidence. Our explanatory hypothesis, that we attempt to confirm in what follows, is that this unexpected outcome is mainly driven by a misalignment between the reported performance and the actual one. Indeed, because of the shut off of all in-presence activities, not only classes but also the exam evaluation shift to remote, becoming more slack. (To be clear, we are not able to assess whether these changes in the evaluation standards are due to a change in the kind of exams made—which also shifted from in-presence to remote—or to the adoption of magnanimous criteria by teachers.) This hypothesis could also fit with the partial different behavior of the last period, when time elapsed and experience cumulated could have impact on the effectiveness of assessment modes.

In what follows, we go beyond the analysis of the overall effect on reported performance to explore the two main different channels through which the pandemic may have negatively impacted actual performances: the sudden shift to remote teaching and the home confinement.

The impact of (suddenly) changing teaching models

Table 2 shows the estimation results of the model specification presented in Eq. ( 2 ) and corresponding to the identification strategy outlined in “ Identifying the impact of (suddenly) changing teaching models ” section. This strategy is aimed at disentangling the effect of the sudden shift to remote teaching on students’ performances. To do that, as anticipated in “ Identifying the impact of (suddenly) changing teaching models ” section, we first restrict the sample to the departments having both in-presence and hybrid courses (see “ The Unimore dataset ” section), then consider only the exams corresponding to classes taught in the immediately preceding teaching period, and then estimate the IPW weights using the course modality as treatment variable.

Column 1, 3 and 5 of Table 2 presents the same base model shown in the previous section restricted to departments providing at least one hybrid course and with the addition of a control variable for the course modality (1 if in-person and 0 otherwise) and using the IPW correction (see Supplementary Table S4 for the first stage estimations). To be noted, Supplementary Table S5 , which is the equivalent of Table 1 in the subsample used in this IPW case, highlights that the pandemic-related coefficient does not change much with respect to the one presented in Table 1 . This evidence confirms that the sample restrictions here adopted, as well as the bias on the coefficient of variables not related to the IPW treatment variable due the application of the IPW correction, does not affect significantly our results. In column 2, 4 and 6 of the same table we use the DID specification presented in “ Identifying the impact of (suddenly) changing teaching models ” section.

In the baseline case, exam marks of students attending in-presence courses are lower if compared to those reported by students attending hybrid courses. When we consider the DID model which adds the interaction term, however, the effect of attending in-presence courses is not significant anymore while the coefficient of the interaction term is negative and strongly significant. Columns 3 and 4 of Table 2 present the same analysis shown in Columns 1 and 2 limiting the reference period to the first year of pandemic (i.e. up to May 2021), thus focusing on the period during which all classes were attended remotely. Clearly, in this case, the magnitude of the interaction term is much larger than before (1.0 vs 0.6 points), as well as the one of the Covid-19 dummy (1.2 vs 0.3 points).

Summing up, the hypothesis according to which the sudden shift to remote teaching had negatively affected students’ performance finds evidence in our results. As hybrid courses generally have half of lessons in presence, we can estimate the total impact of the change in teaching modality by doubling the coefficient of the interaction term, and thus obtaining a value of about 2 points out of 30. To better understand the extent of the estimated effect related to the pandemic, it should be considered that this value represents more than half of the standard deviation of exam marks and 6.6% of the overall marks range. Our estimated value of the losses related to the shift to remote teaching is close to the upper threshold of the results provided by 9 although obtained with different econometric set-up, unit level analysis, performance outcome and in a case study of another country (US vs Italy).

At same time we also confirm the hypothesis that changes in assessment modes are prominent drivers of the increase in student reported outcomes evidenced in the literature. This effect has offset the negative impact of the pandemic period misaligning the effective performance of students from the measured one. Indeed, when we shift from the base to the DID specifications the coefficient of the Covid-19 variable increases substantially and to an extent close to the absolute value of the coefficients of the interactions included.

The impact of the exposure to restrictions

We move now to the analysis of the impact of the exposure to mobility restrictions on students’ performances described in “ Identifying the impact of the exposure to restrictions ” section. To do this purpose, we slightly restricts the sample of the benchmark case (see Table 1 ) as we exclude the exams held by students who are not resident in Italy (they represent less than 2% of the full sample of exams). The second column of Table 3 adds to the base model—whose results are reported in column 1—the overall number of days each student spent under red zone restrictions, while the third column adds the variable reporting the share of days spent under red zone restrictions over the 14 days before the exam. As explained above, in the former case we focus on the cumulated impact of restrictions, while in the latter we assess the impact of being confined at home in the days just before the exam’s session, corresponding to the period of exams’ preparation.

Table 3 highlights that the number of days spent under red zone restrictions decreases the exam marks. One day more spent under lockdown restrictions corresponds to a reduction of 0.003 points. Considering that at the end of the reference period the average value of this variable is 105 days, we can estimate the average effect on students’ exam marks at the end of the pandemic to be about one third of point. At same time, as also in the previous section, when we take into account this negatively impacting channel, the estimated coefficient of the Covid-19 dummy increases.

While the results shows a negative long run effect of home confinement, that can be ascribed to mental stress issues, when we look at the effect in the short run, things substantially change. Our results show that a greater number of days spent under red zone restrictions during the two weeks preceding the exam (i.e. probably those on which the preparation to the exam is mainly concentrated) engenders an increase of students’ exam marks. In this case, the Covid-19 dummy coefficient does not report any relevant variation, confirming the change of examination modality to represent the main explanation of the positive impact on measured performance.

In conclusion, the results of our analysis suggest a composite effect of mobility restrictions. On the one hand, consistently with the results provided by 6 , they might have increased the amount of time allocated to study for exams thus improving performances in the short run. On the one other hand, however, in the long run the protracted exposure to the restrictions clearly reduced the students’ outcomes.

Robustness checks

In this session we present two different robustness checks, one for each of the two channels we considered in main analysis: the change in teaching modalities and the exposure to mobility restrictions.

As for the change in teaching methods, we perform a placebo test analysis. Instead of restricting the sample to on-schedule exams only, we consider the other exams: those made during the pandemic but related to courses attended in the pre-Covid-19 semesters. This test should therefore be considered as valid if two conditions hold. The first one is that the coefficients of the baseline model of the Covid variables are still positive and significant. This would confirm the increase in exams grade is due to the change in exams modality and not to change in teaching modes. The second condition requires that in the DID specification the coefficients of the interaction term between the Covid-19 dummy and the in-presence course one are found to be insignificant or to have very small magnitude. Results of the placebo test, presented in Table 4 , confirm the robustness of our results. In fact, while in the baseline model the coefficients of the covid variables confirm the baseline specification results, the coefficients of the interaction terms in the DID specification are always insignificant and their magnitude is strongly reduced if not even with opposite sign if compared to those reported in Table 2 .

As for the effect of restrictions on students’ performances, one possible weakness of our strategy is the fact that some students may not have returned back to their households and thus the restrictions in place in the region of origin may not correspond to the actual restrictions to which these students where subject to. This would affect our estimates but only partially since in the first stage of pandemic, the variables of interest have only a time variation, not spatial, because restrictions had national dimension. As to the following period, the option of not coming back home does not apply to freshmen students since the decision to keep university activities in remote mode for all the first semester, and to allow in any case to attend classes in remote for all the rest of the year, was communicated well before the opening of course registration. Besides, for the same reasons the case of not coming back home even after the first Covid-19 wave is less likely to have occurred for non-freshmen students because of the rent costs that could be saved. To be noted, house rent costs in Modena and Reggio Emilia are indeed particularly high if compared to other university cities as recorded by the yearly official statistics on living costs performed by the Italian Institute of Statistics, which places the two cities among the highest in Italy for living costs. Finally, it is likely that the climate of fear and concern that had spread in the early stages of the pandemic pushed most of people returning to their household of origin just before the end of the first national lockdown in May 2020 independently from the high economic incentives.

Anyway, to account for this possible source of bias we perform a sensitivity analysis by restricting the sample to students resident out of the Modena and Reggio Emilia provinces. We consider only students who faced the same decisions about where to spend the periods of suspension of university in-person activities, thus the bias would affect randomly all kind of students. Table 5 highlights that the coefficients of variables regarding the effect of restrictive measures do not change substantially with respect to those reported in Table 3 , overall confirming the robustness of our main results.

In the Supplementary Material we also report a heterogeneity analysis of our main results (i.e. those in Table 1 and Table 2 ) to assess whether they present any relevant change when distinguishing departments by ERC sector or teachers by age group (aged 59 or younger vs aged 60 or older). Specifically, Supplementary Table S6 and Table S8 show the heterogeneity of the Covid-19 impact on students’ exam marks by ERC sectors, while Supplementary Table S7 and Table S9 do the same by teachers’ age group.

Supplementary Table S6 points out that coefficients in the first column and the last column always have the same statistically significance and direction. As for the magnitude, departments in the Life Sciences sector (e.g. Medicine and Nursing) seem the most affected by Covid-19, while coefficients of Social Sciences and Humanities and STEM sectors are very similar each other. These results are overall confirmed in the analysis by pandemic period with the exception of Social Sciences and Humanities departments, where the coefficient for the Covid III period is positive but insignificant (in line with results in Table 2 though). Moreover, while the in-presence students appear to have different performances by ERC sector, Supplementary Table S8 highlights that the Covid-19 effect related to the change of teaching modality is negative and significant in all departments except for those in Life Sciences. As for the heterogeneous effects by teachers’ age group, Supplementary Table S7 shows that coefficients are very similar, then suggesting that older teachers have not behaved differently from others. Nonetheless, the DID analysis in Supplementary Table S9 points out a heterogeneous causal effect of Covid-19, which is significant only for the subgroup of older teachers when considering the full period (column 2). This evidence seems to suggest that teachers’ reaction to pandemic-related changes was similar during the first year of pandemic, but the effect has lasted longer among older teachers. All in all, they were more vulnerable to the COVID disease and probably have had a harder time to adapt to the online modality.

Finally, Supplementary Table S10 presents a robustness check on the overall effect of the time spent in a regional red zone during the two weeks preceding the exam (see Table 3 ). First, we provide an estimation where Covid-19 period dummies are included. Second, we provide an estimation focusing on the first year of pandemic only, to assess whether the effect estimated for the full period is stable or not over time. Supplementary Table S10 clearly shows that the effect of the variable of interest here is slightly lower than the one presented in Table 3 (differences are not significant at 10 percent level though), but still strongly positive and significant. This evidence confirms the effect of being forced at home during the two weeks before the exam is actually quite stable over the analyzed period.

Conclusions

In this study, we have focused on the effect of the pandemic on the performance of university students. By exploiting the opportunities provided by an administrative dataset containing very detailed information on the University of Modena and Reggio Emilia (Unimore), one of the forerunners of the restrictions imposed worldwide to universities during Covid-19 early stages, we have tried to solve some inconsistencies in the literature and to unbundle the two main channels through which the pandemic changed university students’ pathways: the shift to remote lessons and the exposure to lockdown measures.

On the one hand, the results of the DID estimations based on the distinction between full in-presence programs and hybrid ones suggests a mismatch between actual performance and measured performance related to the change in assessment methods and/or parameters. In the standard design that uses the Covid-19 period as treatment, the evidence is that of an overall albeit slight improvement in average marks: in the context of a grading system with marks expressed in thirtieths, with 18 as the minimum grade of passed exams and a variance of 3.6, during the pandemic the score of passed the exams increased by one sixth of point, a result substantially in line with that of the literature which also shows slightly positive overall effects in a number of different outcomes of students’ performance. Besides, to a more detailed insight, the pandemic still appears to have had negative effects on student performance. The evidence gathered allows us to estimate the impact of the sudden change in lecture modes in nearly two thirtieths. This result is in line with the literature focusing on specific aspects of the Covid-19 impact and also coherent with the studies on students’ subjective evaluations. Despite this channel seems to have been more relevant, also the psychological effects due to exposure to lockdown measures result as significant: at the end of the period considered, the cumulative impact of exposure to home confinement amounting to about one third of point. At the same time, being confined at home in the two weeks prior to the examination date appears to have had a positive impact: being forced to stay at home during all the two weeks before the exams increase the average grade by nearly one sixth of point. Nonetheless, the driver of the overall positive effect on students’ grades seems to be the change in evaluation standard, that result in having increased student grades by a value in the range of 2–2.5 thirtieths.

As a result, if we look at the effect on student’s actual performance, and thus on their process of human capital accumulation, we can support the evidence of an appreciable negative impact that has been, however, offset on the surface by an average more slack assessment systems. This gives rise to two different kind of problems. The first concerns the most well-known and direct aspect: the loss in terms of human capital accumulation, a significant loss that might have long-term effects. There is however also a further aspect. This generation of students will turn out to be less prepared compared to the others, regardless their similar average marks. This, over time, could produce a stigma effect by fostering a widespread perception that those who studied in the pandemic years are less capable if compared with other ones with same degree or marks. While this may be true for some, in particular for those who have benefited most from the different assessment modes, it is not true for all. Anyway, the signaling role of their degree on job applicants would be weakened. This could result in a process of statistical discrimination: an efficient practice for those who implement it, the employers, but as unfair for an already hard-hit generation of students.

Data availability

The datasets generated and analyzed during the current study uses the information coming from the administrative archives of the University of Modena and Reggio Emilia. They are not publicly available due restrictions related to data ownership but they are available together with all do files from the corresponding author on reasonable request by remote connection to a dedicated server. The research did not rely on any kind of experiments on humans and/or the use of human tissue samples. The whole research was performed in accordance with relevant guidelines/regulations, in particular with all requirements imposed by the Italian Data Protection Authority (GDPR) in its November 27, 2008 Requirements (Gazzetta Ufficiale No. 300, December 24, 2008) and subsequent and possible adjustments and amendments. In compliance with Regulation (EU) 2016/679 of the European Parliament and of the Council of April 27, 2016, Legislative Decrees August 10, 2018 No. 101 and May 18, 2018 No. 51 of the Italian Government, the study did not required ethics approval and/or individual consent of the involved persons (the students of Unimore), who, in any case, at the time of matriculation at Unimore were informed about the processing of personal data also for purposes that respond to and are aimed at implementing the exercise of institutional powers vested in the university, including research.

Betthäuser, B. A., Bach-Mortensen, A. M. & Engzell, P. A systematic review and meta-analysis of the evidence on learning during the COVID-19 pandemic. Nat. Hum. Behav. 7 , 375–385 (2023).

Article   PubMed   Google Scholar  

Angrist, J. D. & Krueger, A. B. Empirical Strategies in Labor Economics. Handbook of Labor Economics. 3(A) 1277–1366. Springer (1999).

Grewenig, E., Lergetporer, P., Werner, K., Woessmann, L. & Zierow, L. COVID-19 and educational inequality: How school closures affect low-and high-achieving students. Eur. Econ. Rev. 140 , 103920 (2021).

Article   PubMed   PubMed Central   Google Scholar  

Fuchs-Schündeln, N. Covid-induced school closures in the United States and Germany: Long-term distributional effects. Econ. Policy 37 (112), 609–639 (2022).

Article   Google Scholar  

Ives, B. & Cazan, A. M. Did the COVID-19 pandemic lead to an increase in academic misconduct in higher education?. Higher Education, on-line first (2023).

Aucejo, E. M., French, J., Araya, M. P. U. & Zafar, B. The impact of COVID-19 on student experiences and expectations: Evidence from a survey. J. Public Econ. 191 , 104271 (2020).

Hu, K. et al. The impact of the COVID-19 pandemic on college students in USA: Two years later. Psychiatry Res. 315 , 114685 (2022).

Rodríguez-Planas, N. Hitting where it hurts most: COVID-19 and low-income urban college students. Econ. Educ. Rev. 87 , 102233 (2022).

Bird, K. A., Castleman, B. L. & Lohner, G. Negative impacts from the shift to online learning during the COVID-19 crisis: Evidence from a statewide community college system. AERA Open 8 , 23328584221081220 (2022).

Gonzalez, T. et al. Influence of COVID-19 confinement on students’ performance in higher education. PLoS ONE 15 (10), e0239490 (2020).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Clark, A. E., Nong, H., Zhu, H. & Zhu, R. Compensating for academic loss: Online learning and student performance during the COVID-19 pandemic. China Econ. Rev. 68 , 101629 (2021).

Rodríguez-Planas, N. COVID-19, college academic performance, and the flexible grading policy: A longitudinal analysis. J. Public Econ. 207 , 104606 (2022).

Engelhardt, B., Johnson, M. & Meder, M. E. Learning in the time of Covid-19: Some preliminary findings. Int. Rev. Econ. Educ. 37 , 100215 (2021).

Castellanos-Serrano, C., Escribano, G., Paredes-Gázquez, J. & San-Martín González, E. What is behind the gender gap in economics distance education: Age, work-life balance and COVID-19. PLoS ONE 17 (8), e0272341 (2022).

Maldonado, J. E. & De Witte, K. The effect of school closures on standardised student test outcomes. Br. Edu. Res. J. 48 (1), 49–94 (2022).

Altindag, D. T., Filiz, E. S. & Tekin, E. Is online education working? National Bureau of Economic Research, No. w29113 (2021).

Agostinelli, F., Doepke, M., Sorrenti, G. & Zilibotti, F. When the great equalizer shuts down: Schools, peers, and parents in pandemic times. J. Public Econ. 206 , 104574 (2022).

Download references

Author information

Authors and affiliations.

Department of Economics, University of Bologna, 40126, Bologna, Italy

Luca Bonacini

Department of Economics Marco Biagi, University of Modena and Reggio Emilia, 41121, Modena, Italy

Giovanni Gallo & Fabrizio Patriarca

You can also search for this author in PubMed   Google Scholar

Contributions

All the authors contributed equally to the paper.

Corresponding author

Correspondence to Fabrizio Patriarca .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Supplementary tables., rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Bonacini, L., Gallo, G. & Patriarca, F. Unraveling the controversial effect of Covid-19 on college students’ performance. Sci Rep 13 , 15912 (2023). https://doi.org/10.1038/s41598-023-42814-7

Download citation

Received : 22 February 2023

Accepted : 14 September 2023

Published : 23 September 2023

DOI : https://doi.org/10.1038/s41598-023-42814-7

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

When perceived similarity overrides demographic similarity: examining influences on stem students’ developmental mentor networks.

• Rachelle Pedersen
• Anna Woodcock
• Paul R. Hernandez

International Journal of STEM Education (2024)

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

• Introduction
• Conclusions
• Article Information

The solid line represents the association between the 2 variables. The dashed lines represent the 95% CI. The circles represent the different studies this particular graph is accounting for, while the sizes of the circles represent the weight of each of the studies.

eTable 1. Detailed Study Summary Characteristics of All 68 Included Studies

eTable 2. Study Summary Characteristics for Comorbidities

eTable 3. Adjustment of Relative Risk Ratios (RRs) for Additional Variables

eTable 4. Adjustment of Odds Ratios (ORs) for Additional Variables

eTable 5. Combined Prevalence of Cohort and Cross-sectional Studies

eTable 6. Summary of Q and I 2  Statistics for Study Variables

eFigure 1. PRISMA Workflow for Studies Included in Analysis

eFigure 2. Funnel Plots for Deceased Individuals in Cohort and Cross-sectional Studies

eFigure 3. Funnel Plots for Patients Admitted to ICU or Hospitalized in Cohort Studies

eFigure 4. Funnel Plots for COVID-19 Positive Patients in Cohort and Cross-sectional Studies

eFigure 5. Forest Plot for COVID-19 Positivity in Cohort and Cross-sectional Studies

eFigure 6. Forest Plot for Patients Admitted to ICU or Hospitalized in Cohort Studies

eFigure 7. Forest Plot for Deceased Individuals in Cohort and Cross-sectional Studies

eFigure 8. Metaregression for County Median Income

eFigure 9. Spearman Correlations for Measures of Clinical Care Quality

eFigure 10. Metaregression for Clinical Care Measures

eFigure 11. Leave-One-Out Sensitivity Analysis for Deceased Individuals in Cohort and Cross-sectional Studies

eFigure 12. Leave-One-Out Sensitivity Analysis for Patients Admitted to ICU or Hospitalized in Cohort Studies

eFigure 13. Leave-One-Out Sensitivity Analysis for COVID-19 Positive Patients in Cohort and Cross-sectional Studies

eFigure 14. Forest Plots for Deceased Patients After Removing Dominating Studies

eFigure 15. Forest Plots for Positive Individuals After Removing Dominating Studies

eMethods 1. Methods Pertaining to Search Criteria and Data Collection

eMethods 2. Citations of Articles that Appeared to Meet Inclusion Criteria but Were Excluded

eMethods 3. Description of Statistical Methods Used in Analyses

• Error in Figures 1 and 2 JAMA Network Open Correction December 21, 2021
• Error in Interpretation of Mortality JAMA Network Open Correction February 24, 2022

See More About

Sign up for emails based on your interests, select your interests.

Customize your JAMA Network experience by selecting one or more topics from the list below.

• Academic Medicine
• Acid Base, Electrolytes, Fluids
• Allergy and Clinical Immunology
• American Indian or Alaska Natives
• Anesthesiology
• Anticoagulation
• Art and Images in Psychiatry
• Artificial Intelligence
• Assisted Reproduction
• Bleeding and Transfusion
• Caring for the Critically Ill Patient
• Challenges in Clinical Electrocardiography
• Climate and Health
• Climate Change
• Clinical Challenge
• Clinical Decision Support
• Clinical Implications of Basic Neuroscience
• Clinical Pharmacy and Pharmacology
• Complementary and Alternative Medicine
• Consensus Statements
• Coronavirus (COVID-19)
• Critical Care Medicine
• Cultural Competency
• Dental Medicine
• Dermatology
• Diabetes and Endocrinology
• Diagnostic Test Interpretation
• Drug Development
• Electronic Health Records
• Emergency Medicine
• End of Life, Hospice, Palliative Care
• Environmental Health
• Equity, Diversity, and Inclusion
• Facial Plastic Surgery
• Gastroenterology and Hepatology
• Genetics and Genomics
• Genomics and Precision Health
• Global Health
• Guide to Statistics and Methods
• Hair Disorders
• Health Care Delivery Models
• Health Care Economics, Insurance, Payment
• Health Care Quality
• Health Care Reform
• Health Care Safety
• Health Care Workforce
• Health Disparities
• Health Inequities
• Health Policy
• Health Systems Science
• History of Medicine
• Hypertension
• Images in Neurology
• Implementation Science
• Infectious Diseases
• Innovations in Health Care Delivery
• JAMA Infographic
• Law and Medicine
• Leading Change
• Less is More
• LGBTQIA Medicine
• Lifestyle Behaviors
• Medical Coding
• Medical Devices and Equipment
• Medical Education
• Medical Education and Training
• Medical Journals and Publishing
• Mobile Health and Telemedicine
• Narrative Medicine
• Neuroscience and Psychiatry
• Notable Notes
• Nutrition, Obesity, Exercise
• Obstetrics and Gynecology
• Occupational Health
• Ophthalmology
• Orthopedics
• Otolaryngology
• Pain Medicine
• Palliative Care
• Pathology and Laboratory Medicine
• Patient Care
• Patient Information
• Performance Improvement
• Performance Measures
• Perioperative Care and Consultation
• Pharmacoeconomics
• Pharmacoepidemiology
• Pharmacogenetics
• Pharmacy and Clinical Pharmacology
• Physical Medicine and Rehabilitation
• Physical Therapy
• Physician Leadership
• Population Health
• Primary Care
• Professional Well-being
• Professionalism
• Psychiatry and Behavioral Health
• Public Health
• Pulmonary Medicine
• Regulatory Agencies
• Reproductive Health
• Research, Methods, Statistics
• Resuscitation
• Rheumatology
• Risk Management
• Scientific Discovery and the Future of Medicine
• Shared Decision Making and Communication
• Sleep Medicine
• Sports Medicine
• Stem Cell Transplantation
• Substance Use and Addiction Medicine
• Surgical Innovation
• Surgical Pearls
• Teachable Moment
• Technology and Finance
• The Art of JAMA
• The Arts and Medicine
• The Rational Clinical Examination
• Tobacco and e-Cigarettes
• Translational Medicine
• Trauma and Injury
• Treatment Adherence
• Ultrasonography
• Users' Guide to the Medical Literature
• Vaccination
• Venous Thromboembolism
• Veterans Health
• Women's Health
• Workflow and Process
• Wound Care, Infection, Healing

Get the latest research based on your areas of interest.

Others also liked.

• Download PDF
• X Facebook More LinkedIn
• CME & MOC

Magesh S , John D , Li WT, et al. Disparities in COVID-19 Outcomes by Race, Ethnicity, and Socioeconomic Status : A Systematic Review and Meta-analysis . JAMA Netw Open. 2021;4(11):e2134147. doi:10.1001/jamanetworkopen.2021.34147

Manage citations:

© 2024

• Permissions

Disparities in COVID-19 Outcomes by Race, Ethnicity, and Socioeconomic Status : A Systematic Review and Meta-analysis

• 1 Division of Otolaryngology–Head and Neck Surgery, Department of Surgery, University of California, San Diego
• 2 Research Service, VA San Diego Healthcare System, San Diego, California
• 3 The University of California Davis School of Medicine, Sacramento
• 4 Department of Radiology, University of California, San Diego
• 5 Radiology Service, VA San Diego Healthcare System, San Diego, California
• Correction Error in Figures 1 and 2 JAMA Network Open
• Correction Error in Interpretation of Mortality JAMA Network Open

Question   Are race and ethnicity–based COVID-19 outcome disparities in the United States associated with socioeconomic characteristics?

Findings   In this systematic review and meta-analysis of 4.3 million patients from 68 studies, African American, Hispanic, and Asian American individuals had a higher risk of COVID-19 positivity and ICU admission than White individuals. Socioeconomic disparity and clinical care quality were associated with COVID-19 mortality and incidence in racial and ethnic minority groups.

Meaning   In this study, members of racial and ethnic minority groups had higher rates of COVID-19 positivity and disease severity than White populations; these findings are important for informing public health decisions, particularly for individuals living in socioeconomically deprived communities.

Importance   COVID-19 has disproportionately affected racial and ethnic minority groups, and race and ethnicity have been associated with disease severity. However, the association of socioeconomic determinants with racial disparities in COVID-19 outcomes remains unclear.

Objective   To evaluate the association of race and ethnicity with COVID-19 outcomes and to examine the association between race, ethnicity, COVID-19 outcomes, and socioeconomic determinants.

Data Sources   A systematic search of PubMed, medRxiv, bioRxiv, Embase, and the World Health Organization COVID-19 databases was performed for studies published from January 1, 2020, to January 6, 2021.

Study Selection   Studies that reported data on associations between race and ethnicity and COVID-19 positivity, disease severity, and socioeconomic status were included and screened by 2 independent reviewers. Studies that did not have a satisfactory quality score were excluded. Overall, less than 1% (0.47%) of initially identified studies met selection criteria.

Data Extraction and Synthesis   Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed. Associations were assessed using adjusted and unadjusted risk ratios (RRs) and odds ratios (ORs), combined prevalence, and metaregression. Data were pooled using a random-effects model.

Main Outcomes and Measures   The main measures were RRs, ORs, and combined prevalence values.

Results   A total of 4 318 929 patients from 68 studies were included in this meta-analysis. Overall, 370 933 patients (8.6%) were African American, 9082 (0.2%) were American Indian or Alaska Native, 101 793 (2.4%) were Asian American, 851 392 identified as Hispanic/Latino (19.7%), 7417 (0.2%) were Pacific Islander, 1 037 996 (24.0%) were White, and 269 040 (6.2%) identified as multiracial and another race or ethnicity. In age- and sex-adjusted analyses, African American individuals (RR, 3.54; 95% CI, 1.38-9.07; P  = .008) and Hispanic individuals (RR, 4.68; 95% CI, 1.28-17.20; P  = .02) were the most likely to test positive for COVID-19. Asian American individuals had the highest risk of intensive care unit admission (RR, 1.93; 95% CI, 1.60-2.34, P  < .001). The area deprivation index was associated with mortality rates in Asian American and Hispanic individuals ( P  < .001). Decreased access to clinical care was positively correlated with COVID-19 positivity in Hispanic individuals ( P  < .001) and African American individuals ( P  < .001).

Conclusions and Relevance   In this study, members of racial and ethnic minority groups had higher risks of COVID-19 positivity and disease severity. Furthermore, socioeconomic determinants were strongly associated with COVID-19 outcomes in racial and ethnic minority populations.

As of August 19, 2021, more than 209 million people across the world had been infected by COVID-19, with the United States accounting for more than 36 million cases and 618 000 deaths. 1 COVID-19 disproportionately affects racial and ethnic minority groups. 2 To reduce exposure and mortality rates, it is critical to identify the disparities associated with greater occurrences of COVID-19 among different populations. 3

In a meta-analysis of 50 articles, 4 it was shown that African American and Asian American patients were at a higher risk of intensive care unit (ICU) admission because of COVID-19 than White patients. A separate meta-analysis examining 45 articles 5 indicated that race may be associated with worse COVID-19 outcomes because of the increased occurrence of comorbidities in racial and ethnic minority groups. However, these studies did not examine the role of socioeconomic determinants, which disproportionately affect racial and ethnic minority populations. Another study 6 explored underlying factors for COVID-19 outcomes in racial and ethnic minority groups but did not integrate data from external sources, such as county median income. As such, current meta-analyses lack investigations assessing how socioeconomic determinants may be associated with COVID-19 disease severity in minority populations.

Individual cross-sectional and cohort studies have found that COVID-19 infection rates in racial and ethnic minority groups are associated with low socioeconomic status and income. 7 , 8 Specifically, studies have found that there is a positive association between COVID-19 risk and area deprivation index (ADI). 9 Past studies have also demonstrated that 11.7% of African Americans individuals are uninsured, compared with 7.5% of White individuals, thus potentially leading to more severe disease outcomes because of lack of access to medical care. 10 Geographic variation may also play a role in COVID-19 disease severity, as rural hospitals and communities often lack resources. 11 Therefore, it is plausible that these social determinants might be associated with COVID-19 disease severity in racial and ethnic minority populations.

In this study, we examine the associations of race and ethnicity with COVID-19 positivity rates, mortality, hospitalization, and ICU admission in the United States. We then associate these outcomes with various social determinants through adjusted and unadjusted relative risk ratio (RR) and odds ratio (OR) calculations and metaregression analysis. To our knowledge, we are the first to examine social determinants of health in racial disparities of COVID-19 outcomes through a systematic review and meta-analysis, which provides a more accurate understanding than results published in single-site studies.

We conducted a systematic search of studies published from January 1, 2020, to January 6, 2021, in the PubMed, medRxiv, bioRxiv, Embase, and the World Health Organization COVID-19 databases. We used search terms pertaining to COVID-19 and disparities (eMethods 1 in the Supplement ) and only included studies that reported data on race and ethnicity as well as the following variables: socioeconomic status, COVID-19 positivity, hospitalization, ICU admission, mortality, and location/geography. All included studies were conducted in the United States.

Two independent reviewers (S.M. and D.J.) screened the titles, abstracts, and full text of each eligible study from the selected databases. Disagreements were resolved through discussion with a third reviewer (Y.L.). We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses ( PRISMA ) guideline for selection of papers in this meta-analysis (eFigure 1 and eMethods 2 in the Supplement ). The Joanna Briggs Institute critical appraisal tools were used to assess the quality of evidence from all studies in respect to study design. Studies were not included in our analysis if they scored lower than a 6 of 8 (75%) for cohort studies and 9 of 11 (82%) for cross-sectional studies. The complete search and inclusion strategy can be found in the eMethods 1 and 2 in the Supplement .

Data were extracted from the 68 studies screened with the PRISMA guidelines. We collected details from studies regarding study setting and type and patient demographic characteristics, comorbidities, and outcomes (eMethods 1 in the Supplement ), using the same independent reviewer design as during study selection. Following the initial data review, socioeconomic variables quantifying disparities in health, income, and geography were extracted from external sources using zip code and congressional district location (eMethods 1 in the Supplement ). 12 - 15 External measures of socioeconomic disparities were not extracted for studies that occurred at a statewide level or that included data from patients across the United States, as the specific tools we used to determine these values were limited by units of analysis at the county, congressional district, and/or geographic address level.

COVID-19 has been strongly associated with lower socioeconomic status in racial/ethnic minorities. 16 Accordingly, ADI was used as a quantitative measure of socioeconomic disadvantage, and it accounts for several factors, such as income, education, employment, and housing quality. The Urban Core Opportunity Index (UOI) measures the urbanicity of geographic location, through the characterization of factors such as the amount of renters and households without vehicles. 15

We also examined the association of clinical care quality with COVID-19 positivity, mortality, ICU admission, and hospitalization through metaregression analysis. Specifically, we investigated the following measures of clinical care quality: preventable hospital stays, ratio of the population to primary care physicians, and percentage of uninsured individuals. A higher rate of preventable hospital stays represents a lower quality of available medical care, and a higher ratio of the population to physicians refers to a larger population with access to only 1 primary care physician. 14

All data analysis was conducted using R Studio version 4.1.1 (R Project for Statistical Computing). Analyses were conducted separately for each racial and ethnic group in the following cohorts: COVID-19 positivity, ICU admission, hospitalization, and mortality. Studies with missing data for a particular cohort or variable were excluded from the respective analysis. The following analyses were conducted to investigate the association of race and ethnicity with COVID-19 outcomes. Combined prevalence refers to the incidence of COVID-19 outcomes in a certain population per 1000 patients. Metaregression analysis was conducted to assess associations between study effect size and socioeconomic variables extracted by study location. Relative risk ratios (RRs) and odds ratios (ORs) were also used to assess the associations of race and ethnicity with COVID-19 outcomes, with White individuals as the reference group. Both RR and OR values were adjusted for several key confounders using a linear mixed-effect model (eMethods 3 in the Supplement ). Statistical significance was set at P  < .05, and all tests were 2-tailed. The Egger test was used to assess publication bias, with P  < .05 as the level of statistical significance (eFigures 2-4 in the Supplement ). Information for all the studies is reported in detail in eTables 1 and 2 in the Supplement .

A total of 4 318 929 patients from 68 studies 17 - 84 were included in this meta-analysis ( Table ). Overall, 370 933 patients (8.6%) were African American, 9082 (0.2%) were American Indian or Alaska Native, 101 793 (2.4%) were Asian American, 851 392 identified as Hispanic/Latino (19.7%), 7417 (0.2%) were Pacific Islander, 1 037 996 (24.0%) were White, and 269 040 (6.2%) identified as multiracial or of other racial or ethnic group. The studies were separated into cohort and cross-sectional studies for data analysis. All unadjusted and adjusted RR and OR values are reported in Figure 1 and Figure 2 and eTables 3 and 4 in the Supplement .

In age- and sex-adjusted analyses, we found that African American and Hispanic individuals were significantly more likely to test positive for COVID-19 than White individuals (African American: RR, 3.54; 95% CI, 1.38-9.07; P  = .008; Hispanic: RR, 4.68; 95% CI, 1.28-17.20, P  = .02) ( Figure 1 ). There was a lack of data to calculate age- and sex-adjusted RR and OR values for Asian American individuals. Following adjustment for ADI, African American and Hispanic individuals were almost 2 times as likely to test positive for COVID-19 as White individuals (African American: RR, 2.01; 95% CI, 1.04-3.88; P  = .04; Hispanic: RR, 2.09; 95% CI, 1.13-3.88; P  = .02), followed by Asian American individuals (RR, 1.12; 95% CI, 1.04-1.21; P  = .003) ( Figure 1 ). After adjustment for clinical care quality, we found that African American individuals were still the most likely to test positive for COVID-19 (RR, 1.79; 95% CI, 1.11-3.17; P  = .03), followed by Asian American individuals (RR, 1.16; 95% CI, 1.03-1.31; P  = .02) ( Figure 1 ). Hispanic individuals did not exhibit significant results following adjustment for clinical care quality. Interestingly, adjustment for the UOI demonstrated that Asian American individuals face the highest risk of COVID-19 positivity (RR, 1.13; 95% CI, 1.07-1.19, P  < .001) (eTable 3 in the Supplement ). We did not observe significant results in African American and Hispanic individuals following adjustment for UOI. Combined prevalence values demonstrated similar trends, with African American individuals having the highest prevalence of COVID-19 positivity (eFigure 5 in the Supplement ). In summary, with some exceptions, adjusting for ADI and clinical care quality significantly decreased the risk of COVID-19 infection in African American and Hispanic individuals when compared with White individuals. However, the risk still remained high in these populations following adjustment.

COVID-19 disease severity was assessed through ICU admission and hospitalization rates among various racial and ethnic groups (eFigure 6 in the Supplement ). Following adjustment for sex, Asian American individuals had a significant RR of 1.93 (95% CI, 1.60-2.34; P  < .001) compared with White individuals ( Figure 1 ).

The combined prevalence of COVID-19 mortality rates in cohort studies was highest among White individuals (161.12 per 1000 patients), followed by African American individuals (143.99 per 1000 patients), Hispanic/Latino individuals (130.51 per 1000 patients), and Asian American individuals (42.99 per 1000 patients) (eTable 5 in the Supplement ). In cross-sectional studies, the combined prevalence of mortality rates were highest among African American individuals (277.15 per 1000 patients), followed by Hispanic individuals (213.34 per 1000 patients), White individuals (173.38 per 1000 patients), and Asian individuals (80.4 per 1000 patients) (eTable 5 and eFigure 7 in the Supplement ).

In cross-sectional studies, the ADI-adjusted RR for mortality was 0.44 (95% CI, 0.31-0.61; P  < .001) for Hispanic individuals, and the county median income–adjusted RRs for mortality were 0.43 (95% CI, 0.41-0.46; P  < .001) for Hispanic individuals and 0.44 (95% CI, 0.36-0.54; P  = .001) for Asian American individuals.

We further investigated the association of ADI with COVID-19 positivity and disease severity by race and ethnicity through metaregression analysis. A higher ADI corresponds to worse socioeconomic status. Accordingly, we found that an increase in ADI was positively associated with the mortality rates of Asian American and Hispanic individuals in cross-sectional studies ( P  < .001) ( Figure 3 ). Interestingly, an increase in ADI was negatively associated with mortality rates of Hispanic individuals in cohort studies ( P  = .03) ( Figure 3 ).

We conducted metaregression analysis to assess the association of county median income with COVID-19 outcomes by race and ethnicity. Although ADI is a more comprehensive measure of socioeconomic deprivation, we also analyzed county median income because it provided more significant results for RR/OR adjustment in comparison with ADI. Therefore, we determined that we should further examine any associations with income, as it may have been more strongly associated with COVID-19 outcomes than other socioeconomic measures included in ADI. In cohort studies, we found that county median income was negatively associated with mortality rates in Asian American populations ( P  < .001). In cross-sectional studies, higher county median income was associated with lower mortality rates in Hispanic and African American individuals ( P  < .001). County median income was also negatively associated with the proportion of White individuals admitted to the ICU ( P  = .02) ( Figure 4 ; eFigure 8 in the Supplement ).

Through further metaregression analysis, we determined that Hispanic individuals had a positive association of increasing income and positivity rates ( P  = .03). However, African American individuals displayed a negative association between income and positivity rates ( P  = .02).

We additionally conducted Spearman correlations to assess the degree of association between these studied determinants. We observed a strong, positive correlation between county median income and area deprivation index ( R  = 0.61; P  < .001), as county median income was among the measures included in developing the ADI (eFigure 9 in the Supplement ). We found there was a lesser degree of association between county median income and measures of clinical care quality (eFigure 9 in the Supplement ).

In cohort studies, we found that an increase in number of preventable hospital stays ( P  = .04) and the population served by 1 primary care physician ( P  = .009) were associated with a decrease in positivity among Asian American individuals (eFigure 10 in the Supplement ). Conversely, the population served by 1 primary care physician was positively associated with COVID-19 positivity among Hispanic individuals ( P  < .001). In cross-sectional studies, we found that the ratio of the population served to primary care physicians was positively correlated with mortality among White individuals ( P  < .001). The percentage of uninsured individuals was positively associated with positivity among African American ( P  < .001) and White ( P  = .01) individuals in cohort and cross-sectional studies.

We found that cohort studies detailing the proportion of Asian American and Hispanic individuals who tested positive for COVID-19, cohort studies detailing the proportion of Asian American individuals admitted to the ICU, cross-sectional studies detailing the proportion of Asian American individuals who died, and cohort studies detailing the proportion of African American individuals who died exhibited publication bias. To evaluate the association of study heterogeneity with summary proportions, we conducted leave-one-out sensitivity analysis to measure the effects of outliers (eTable 6 and eFigures 11-13 in the Supplement ). We found that summary proportions were not significantly altered by the removal of these outliers (eFigures 14 and 15 in the Supplement ). However, we observed that following the removal of the outlier in the COVID-19 mortality group (ie, cohort studies), African American individuals had the highest rate of mortality followed by Asian American individuals. Prior to removal of the outlier, we found that mortality rates were highest among White individuals. We additionally observed high heterogeneity statistics in our results, indicating that there may be variability in the studies included.

In our meta-analysis, we found that COVID-19 positivity and ICU admission rates were higher in African American, Hispanic, and Asian American individuals compared with White individuals, with some exceptions. Our results are consistent with previous findings that suggest that racial and ethnic minority groups face a higher risk of ICU admission and COVID-19 positivity than White populations. 65 , 85 - 89

However, current meta-analyses do not provide associations with socioeconomic variables, which are highly implicated in COVID-19 outcomes. Therefore, in this study, we aimed to investigate both racial and ethnic disparities in COVID-19 outcomes as well as their associations with socioeconomic variables.

Following adjustment for ADI and clinical care quality, we found that risk of COVID-19 positivity in African-American and Hispanic individuals substantially decreased. However, the risk for COVID-19 positivity following adjustment remained higher in these minoritized populations when compared with White individuals. As such, this occurrence may be because of the overrepresentation of members of racial and ethnic minority groups in essential jobs, which increase exposure to COVID-19. Furthermore, comorbidities, such as hypertension or obesity, are prevalent among minority populations, thus contributing to worsened disease outcomes. 8 , 90 , 91 To our knowledge, we are the first to adjust RRs and ORs of race-associated COVID-19 outcomes using health care quality and access.

We further examined the association of socioeconomic determinants with COVID-19 positivity rates, mortality rates, hospitalization, and ICU admission in racial and ethnic minority groups through metaregression analysis. Increased deprivation was found to be associated with increased mortality in Asian American individuals. Paradoxically, an increase in county median income was associated with increased mortality rates in Asian American individuals. This result suggests that factors other than income that contribute to ADI, such as education, housing equality, and employment, could affect Asian American populations. One hypothesis is that a large number of Asian American individuals work in health care settings, which can lead to increased mortality rates that do not reflect the general population of the surrounding community. 92

An increase in deprivation was also found to be associated with decreased mortality rates in Hispanic individuals in cohort studies, although the opposite result was seen in cross-sectional studies. This inconsistency suggests that further research is needed to establish conclusively the association between mortality rates and deprivation in Hispanics.

We additionally assessed associations between measures of clinical care quality and COVID-19 outcomes. Curiously, we found that an increase in preventable hospital stays and the population served by 1 primary physician were associated with a decrease in the percentage of Asian American individuals who tested positive for COVID-19, suggesting again that other variables may be affecting COVID-19 positivity rates in this population.

Conversely, we observed a positive association between lack of primary care physician access (ie, increased ratio of population to physician) and COVID-19 positivity among Hispanic individuals. Past studies have demonstrated that Hispanic individuals are less likely to delay care if the primary care physician to patient ratio is improved. 93

An increase in the number of uninsured individuals was also positively associated with COVID-19 positivity among African American individuals. African American individuals are less likely to have health insurance coverage compared with White individuals. 94 Members of racial and ethnic minority groups who are uninsured may also not have access to COVID-19 tests. 10

Collectively, our findings demonstrate that racial and ethnic minority groups have faced higher risk of COVID-19 positivity and ICU admission. Public health policies should address socioeconomic and racial disparities to reduce exposure to and fatality from COVID-19 in underrepresented populations. Increasing equitable access to health care and improving resources for underserved populations may reduce exposure to COVID-19 in racial/ethnic minorities.

Our study has several limitations. First, we found high heterogeneity statistics, indicating that there may be variation in the effect sizes of the studies. Second, a number of publications that were included had incomplete or missing data on mortality, positivity, ICU admission, and hospitalization rates. Moreover, there were limited data on several racial and ethnic groups. There was also a lack of information on comorbidities and socioeconomic status variables (ADI, UOI, clinical care quality, income), which limited our ability to adjust for all variables in the multivariable models, although we were able to adjust for some variables individually in separate models (eTables 3 and 4 in the Supplement). Imputation methods were considered but not pursued as it would be difficult to accurately impute measures from patient studies in distinct health systems.

Additionally, several study cohorts exhibited publication bias. As publication bias reduces the accuracy of results, the validity of results in these particular study cohorts may be limited.

In this study, African American, Hispanic, and Asian American individuals were at considerably higher risk of COVID-19 positivity and ICU admission compared with White individuals. Adjustment for social determinants of health and socioeconomic factors decreased risks of COVID-19 positivity in racial and ethnic minority groups; however, several factors were not accounted for by these variables. We also observed that decreased access to clinical care was positively associated with COVID-19 positivity in Hispanic and African American individuals. In conclusion, we found that racial and ethnic disparities in COVID-19 outcomes could be accounted for by socioeconomic determinants in some populations, such as African American, Hispanic, and Asian American individuals.

Accepted for Publication: September 12, 2021.

Published: November 11, 2021. doi:10.1001/jamanetworkopen.2021.34147

Correction: This article was corrected on December 21, 2021, to fix errors in Figures 1 and 2, and on February 24, 2022, to remove the interpretation of mortality risk estimates from unadjusted regression models.

Open Access: This is an open access article distributed under the terms of the CC-BY License . © 2021 Magesh S et al. JAMA Network Open .

Corresponding Author: Weg M. Ongkeko, MD, PhD, Division of Otolaryngology–Head and Neck Surgery, Department of Surgery, University of California, San Diego, 9500 Gilman Dr, La Jolla, CA 92093 ( [email protected] ).

Author Contributions: Dr Ongkeko had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Ms Magesh, Mr John, and Mr W. Li contributed equally and are co–first authors.

Concept and design: Magesh, John, W. Li, Mattingly-app, Ongkeko.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Magesh, John, Mattingly-app, Ongkeko.

Critical revision of the manuscript for important intellectual content: Magesh, John, W. Li, Y. Li, Jain, Chang, Ongkeko.

Statistical analysis: Magesh, John, W. Li, Y. Li, Mattingly-app.

Obtained funding: Chang, Ongkeko.

Administrative, technical, or material support: W. Li, Y. Li, Jain, Ongkeko.

Supervision: W. Li, Ongkeko.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by grant R00RG2369 from the University of California, Office of the President/Tobacco-Related Disease Research Program Emergency COVID-19 Research Seed Funding to Dr Ongkeko.

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

• Register for email alerts with links to free full-text articles
• Access PDFs of free articles
• Manage your interests
• Save searches and receive search alerts

The pandemic has had devastating impacts on learning. What will it take to help students catch up?

Subscribe to the brown center on education policy newsletter, megan kuhfeld , megan kuhfeld director of growth modeling and data analytics - nwea jim soland , jim soland assistant professor, school of education and human development - university of virginia, affiliated research fellow - nwea karyn lewis , and karyn lewis vice president of research and policy partnerships - nwea emily morton emily morton research scientist - nwea.

March 3, 2022

As we reach the two-year mark of the initial wave of pandemic-induced school shutdowns, academic normalcy remains out of reach for many students, educators, and parents. In addition to surging COVID-19 cases at the end of 2021, schools have faced severe staff shortages , high rates of absenteeism and quarantines , and rolling school closures . Furthermore, students and educators continue to struggle with mental health challenges , higher rates of violence and misbehavior , and concerns about lost instructional time .

As we outline in our new research study released in January, the cumulative impact of the COVID-19 pandemic on students’ academic achievement has been large. We tracked changes in math and reading test scores across the first two years of the pandemic using data from 5.4 million U.S. students in grades 3-8. We focused on test scores from immediately before the pandemic (fall 2019), following the initial onset (fall 2020), and more than one year into pandemic disruptions (fall 2021).

Average fall 2021 math test scores in grades 3-8 were 0.20-0.27 standard deviations (SDs) lower relative to same-grade peers in fall 2019, while reading test scores were 0.09-0.18 SDs lower. This is a sizable drop. For context, the math drops are significantly larger than estimated impacts from other large-scale school disruptions, such as after Hurricane Katrina—math scores dropped 0.17 SDs in one year for New Orleans evacuees .

Even more concerning, test-score gaps between students in low-poverty and high-poverty elementary schools grew by approximately 20% in math (corresponding to 0.20 SDs) and 15% in reading (0.13 SDs), primarily during the 2020-21 school year. Further, achievement tended to drop more between fall 2020 and 2021 than between fall 2019 and 2020 (both overall and differentially by school poverty), indicating that disruptions to learning have continued to negatively impact students well past the initial hits following the spring 2020 school closures.

These numbers are alarming and potentially demoralizing, especially given the heroic efforts of students to learn and educators to teach in incredibly trying times. From our perspective, these test-score drops in no way indicate that these students represent a “ lost generation ” or that we should give up hope. Most of us have never lived through a pandemic, and there is so much we don’t know about students’ capacity for resiliency in these circumstances and what a timeline for recovery will look like. Nor are we suggesting that teachers are somehow at fault given the achievement drops that occurred between 2020 and 2021; rather, educators had difficult jobs before the pandemic, and now are contending with huge new challenges, many outside their control.

Clearly, however, there’s work to do. School districts and states are currently making important decisions about which interventions and strategies to implement to mitigate the learning declines during the last two years. Elementary and Secondary School Emergency Relief (ESSER) investments from the American Rescue Plan provided nearly $200 billion to public schools to spend on COVID-19-related needs. Of that sum, $22 billion is dedicated specifically to addressing learning loss using “evidence-based interventions” focused on the “ disproportionate impact of COVID-19 on underrepresented student subgroups. ” Reviews of district and state spending plans (see Future Ed , EduRecoveryHub , and RAND’s American School District Panel for more details) indicate that districts are spending their ESSER dollars designated for academic recovery on a wide variety of strategies, with summer learning, tutoring, after-school programs, and extended school-day and school-year initiatives rising to the top.

Comparing the negative impacts from learning disruptions to the positive impacts from interventions

To help contextualize the magnitude of the impacts of COVID-19, we situate test-score drops during the pandemic relative to the test-score gains associated with common interventions being employed by districts as part of pandemic recovery efforts. If we assume that such interventions will continue to be as successful in a COVID-19 school environment, can we expect that these strategies will be effective enough to help students catch up? To answer this question, we draw from recent reviews of research on high-dosage tutoring , summer learning programs , reductions in class size , and extending the school day (specifically for literacy instruction) . We report effect sizes for each intervention specific to a grade span and subject wherever possible (e.g., tutoring has been found to have larger effects in elementary math than in reading).

Figure 1 shows the standardized drops in math test scores between students testing in fall 2019 and fall 2021 (separately by elementary and middle school grades) relative to the average effect size of various educational interventions. The average effect size for math tutoring matches or exceeds the average COVID-19 score drop in math. Research on tutoring indicates that it often works best in younger grades, and when provided by a teacher rather than, say, a parent. Further, some of the tutoring programs that produce the biggest effects can be quite intensive (and likely expensive), including having full-time tutors supporting all students (not just those needing remediation) in one-on-one settings during the school day. Meanwhile, the average effect of reducing class size is negative but not significant, with high variability in the impact across different studies. Summer programs in math have been found to be effective (average effect size of .10 SDs), though these programs in isolation likely would not eliminate the COVID-19 test-score drops.

Figure 1: Math COVID-19 test-score drops compared to the effect sizes of various educational interventions

Source: COVID-19 score drops are pulled from Kuhfeld et al. (2022) Table 5; reduction-in-class-size results are from pg. 10 of Figles et al. (2018) Table 2; summer program results are pulled from Lynch et al (2021) Table 2; and tutoring estimates are pulled from Nictow et al (2020) Table 3B. Ninety-five percent confidence intervals are shown with vertical lines on each bar.

Notes: Kuhfeld et al. and Nictow et al. reported effect sizes separately by grade span; Figles et al. and Lynch et al. report an overall effect size across elementary and middle grades. We were unable to find a rigorous study that reported effect sizes for extending the school day/year on math performance. Nictow et al. and Kraft & Falken (2021) also note large variations in tutoring effects depending on the type of tutor, with larger effects for teacher and paraprofessional tutoring programs than for nonprofessional and parent tutoring. Class-size reductions included in the Figles meta-analysis ranged from a minimum of one to minimum of eight students per class.

Figure 2 displays a similar comparison using effect sizes from reading interventions. The average effect of tutoring programs on reading achievement is larger than the effects found for the other interventions, though summer reading programs and class size reduction both produced average effect sizes in the ballpark of the COVID-19 reading score drops.

Figure 2: Reading COVID-19 test-score drops compared to the effect sizes of various educational interventions

Source: COVID-19 score drops are pulled from Kuhfeld et al. (2022) Table 5; extended-school-day results are from Figlio et al. (2018) Table 2; reduction-in-class-size results are from pg. 10 of Figles et al. (2018) ; summer program results are pulled from Kim & Quinn (2013) Table 3; and tutoring estimates are pulled from Nictow et al (2020) Table 3B. Ninety-five percent confidence intervals are shown with vertical lines on each bar.

Notes: While Kuhfeld et al. and Nictow et al. reported effect sizes separately by grade span, Figlio et al. and Kim & Quinn report an overall effect size across elementary and middle grades. Class-size reductions included in the Figles meta-analysis ranged from a minimum of one to minimum of eight students per class.

There are some limitations of drawing on research conducted prior to the pandemic to understand our ability to address the COVID-19 test-score drops. First, these studies were conducted under conditions that are very different from what schools currently face, and it is an open question whether the effectiveness of these interventions during the pandemic will be as consistent as they were before the pandemic. Second, we have little evidence and guidance about the efficacy of these interventions at the unprecedented scale that they are now being considered. For example, many school districts are expanding summer learning programs, but school districts have struggled to find staff interested in teaching summer school to meet the increased demand. Finally, given the widening test-score gaps between low- and high-poverty schools, it’s uncertain whether these interventions can actually combat the range of new challenges educators are facing in order to narrow these gaps. That is, students could catch up overall, yet the pandemic might still have lasting, negative effects on educational equality in this country.

Given that the current initiatives are unlikely to be implemented consistently across (and sometimes within) districts, timely feedback on the effects of initiatives and any needed adjustments will be crucial to districts’ success. The Road to COVID Recovery project and the National Student Support Accelerator are two such large-scale evaluation studies that aim to produce this type of evidence while providing resources for districts to track and evaluate their own programming. Additionally, a growing number of resources have been produced with recommendations on how to best implement recovery programs, including scaling up tutoring , summer learning programs , and expanded learning time .

Ultimately, there is much work to be done, and the challenges for students, educators, and parents are considerable. But this may be a moment when decades of educational reform, intervention, and research pay off. Relying on what we have learned could show the way forward.

Related Content

Megan Kuhfeld, Jim Soland, Beth Tarasawa, Angela Johnson, Erik Ruzek, Karyn Lewis

December 3, 2020

Lindsay Dworkin, Karyn Lewis

October 13, 2021

Early Childhood Education Education Access & Equity Education Policy K-12 Education

Governance Studies

U.S. States and Territories

Brown Center on Education Policy

September 24, 2024

Michael Trucano, Sopiko Beriashvili

September 20, 2024

Megan Kuhfeld, Karyn Lewis

Europe PMC requires Javascript to function effectively.

Either your web browser doesn't support Javascript or it is currently turned off. In the latter case, please turn on Javascript support in your web browser and reload this page.

• DOI: 10.48206/kceba.2023.7.6.125
• Corpus ID: 266264893

The impact of college students’ major satisfaction on employment stress: Focusing on the control variables of part-time work experience

• Duksoon Yim
• Published in The Korean Career… 30 November 2023
• Psychology, Education
• The Korean Career, Entrepreneurship &amp; Business Association

Related Papers

Showing 1 through 3 of 0 Related Papers