Background

Since its first reported case in late December 2019 in Wuhan, China, and rapid global spread, Coronavirus Disease 2019 (COVID-19) has resulted in devastating consequences [1, 2] and was declared as a global public health emergency on March 11, 2020, by the World Health Organization (WHO) [3]. Due to the absence of confirmed treatment and vaccination for about 12 months after the onset of the outbreak, most of the research was focused on clinical, pharmaceutical, and epidemiological aspects of the pandemic [3, 4]. Nevertheless, this does not downplay the role of public health in ensuring equity in access to needed services [5].

Racial/ethnic and Socioeconomic Status (SES) disparities in health are prominent [6–8] and have been linked directly or indirectly to individual life expectancy and mortality [9]. Evidence shows a connection between lower SES and higher risk of infectious diseases’ incidence and severity [10–12]. Therefore, narrowing the disparity gaps or eradicating racial/ethnic and SES disparities is a public health priority [13]. This calls for research studies to identify and understand the status of disparities across different population groups [7, 14]. Because their specific SES can potentially influence the incidence and severity of the disease in various ways, populations of low-SES and racial/ethnic minority groups should be considered as high-risk populations during the epidemics [15].

Racial/ethnic and SES disparities are observed in the health care utilization and health outcomes of populations during the COVID-19 pandemic, especially in terms of morbidity and mortality [5, 16, 17]. Gross disparities in hospitalization rates and mortality between racial/ethnic groups in the context of COVID-19 highlight the shortcomings of public health strategies in achieving “optimal health for all” [5, 18]. For instance, several studies have shown disproportionate adverse effects of COVID-19 on African Americans [16, 19–21]. Disparities in COVID-19 is not merely a concern of developed countries. It may affect Low- and Middle-Income Countries (LMICs) even more severely [22–24]. Nevertheless, many LMICs do not have appropriate surveillance systems and responsive healthcare infrastructures [25] to address such issues reliably.

It is well-documented that individuals from low-SES or racial/ethnic minority groups are more vulnerable to COVID-19 [12, 26, 27]. It is crucial to explore the racial/ethnic and SES disparities to reach a precise understanding of the presence and extent of potential disparities in the context of COVID-19. Additionally, several commentaries and editorials have highlighted the essential need for exploring the disparities in COVID-19 to conduct early interventions [5, 28–30]. Although some studies suggest the existence of racial/ethnic and SES disparities in COVID-19, the evidence is inconsistent. For example, while some studies reported higher mortality rates in racial/ethnic minority groups [12, 31, 32], some other studies did not report such differences by race/ethnicity [33, 34]. Conducting a review study to systematically collect data and synthesize the findings from different studies may provide researchers and policymakers with insights into possible disparities in COVID-19.

Although a review study has rigorously examined the disparities in the era of COVID-19, this study has focused only on clinical outcomes with a narrower focus on race/ethnicity [30]. We have included a broader range of outcomes in the racial/ethnic minority groups and low SES populations. Two main objectives of this study include (1) to systematically review the evidence on the association of race/ethnicity with health outcomes and health resources in the context of COVID-19, and (2) to systematically review the evidence on the association of SES with health outcomes and healthcare resources in the context of COVID-19.

Methods

This systematic review adheres to the four-step flowchart suggested by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement [35].

Eligibility criteria for the inclusion of studies

Population

The target population includes the population of countries during the pandemic of COVID-19.

Exposure

The exposure is defined as belonging to racial/ethnic minority and/or low-SES populations. A framework showing the types and definitions of the exposures is presented in Fig. 1.

Fig. 1
figure 1

Characteristics and definition of the exposures and types of the outcome measures

Full size image

Outcomes

The main outcomes of interest include 1) death from COVID-19, 2) COVID-19 incidence/infection, 3) COVID-19 hospitalization, 4) ICU admission for COVID-19, 5) need for mechanical ventilation, 6) confirmed diagnosis for COVID-19, and 7) access to testing for COVID-19.

Types of studies

We included empirical English-language peer-reviewed observational studies in the review. We excluded the non-peer-reviewed papers, editorials, reviews, commentaries, mathematical modeling reports, and methodological and conceptual papers from the review.

Search strategy and information sources

Keywords were identified in three main domains—coronavirus infectious disease, disparity, and racial/ethnic minority groups—through a rapid literature review and using medical subject heading (MeSh) and Embase subject headings (EMTREE) (Table 1). An experienced medical librarian contributed to the development of the preliminary search strategy. In the first step, we searched each of the domains independently, followed by a comprehensive search using a composition of all domains to ensure that the final search strategy was appropriate. Complete search strategies are provided in the supplementary file (Additional file 1 in supplementary files)

Table 1 key terms for database search strategy
Full size table

We searched major electronic databases (from late 2019 onwards), including Medline (via PubMed), Web of Science, Cochrane Library, Scopus, CINAHL, and Embase, on November 20, 2020, and updated on March 1, 2021. Moreover, we searched the references and citation lists of relevant studies, Google Scholar search engine, preprint sources such as medRxiv, WHO website, and other relevant sources to identify the relevant studies.

Data management and extraction

After pooling all search results in the Endnote reference management software program and removing duplicated results, two reviewers screened the titles/abstracts to find all relevant studies (KG and SA). Then, two researchers carefully read the full text of the relevant studies based on the eligibility criteria (SI and AK). Three authors (SI, AK, and DN) extracted essential information and entered it into the data extraction table designed in Microsoft Word. The extraction form contained items such as bibliographic information, type of study, participants and sample size, race/ethnicity ascertainment source, study settings, data source, exposure and outcomes, and main results (See Additional file 2 in supplementary files). The investigators conducted a pilot test of three articles to ensure the validity of the data extraction form before starting the main data extraction.

Assessment of methodological quality

Using standardized critical appraisal instruments for observational studies from Joanna Briggs Institute (JBI) [36], two investigators (SA and DN) independently assessed the quality of eligible studies. This measure was taken to confirm the internal validity of the review’s findings and prevent problems of spurious precision due to confounded or biased statistics. The studies were divided into three categories: cross-sectional studies, cohort studies, and case-control studies, and relevant checklists were used to assess each study type. The evaluation checklists included 8, 11, and 10 questions for cross-sectional, cohort, and case-control studies, respectively. The answers to each item in these tools were “Yes”, “No”, “Unclear” or “Not applicable”. The answer “yes” was assigned the score “1” and answers “no”, “cannot be answered,” or “not applicable” were scored “0”. The quality score of each study was calculated and reported as a percentage. The final score for each study was given by agreement between the two evaluators. Any disagreement among the two investigators was addressed and solved by discussing it with the other authors.

Data synthesis

Due to the high level of heterogeneity in the study settings, participants, and methods, as well as different types of exposures and outcome measures, we did not conduct Meta-analysis. By aggregating the information extracted from each manuscript, we synthesized and organized the results from the included studies. We also explained the main results in narrative and tabular formats.

Results

Our search in major electronic databases yielded 13,558 results. After removing the repetitive results and screening for relevant title/abstracts, 77 studies were selected for full-text review. Finally, 52 studies were included in the review [22–24, 26, 27, 32–34, 37–80]. The study selection process and reasons for excluding the studies after full-text review are presented in Fig. 2.

Fig. 2
figure 2

Prisma diagram for study selection

Full size image

The quality assessment results of the included studies showed that the average score of cross-sectional studies was 86.25%, cohort studies was 90.9%, and case-control studies was 100%. Among the cross-sectional studies, one of the most problematic items was strategies to deal with confounding factors stated. Among cohort studies, strategies to address incomplete follow-up was the main problem (See Additional file 3 in the supplementary files).

The majority of the studies were from the United States (37 studies). Studies have targeted different race/ethnicity and SES characteristics to examine the disparities in COVID-19 outcomes. Almost all studies used self-reported data on race/ethnicity, either already documented in the database they used for data collection, or gathered and recorded in the patients’ medical documents. A total of 29 out of 52 studies examined the disparities using individual-level data.

Racial/ethnic disparities and COVID-19 outcomes

A total of 40 studies examined the association between race/ethnicity and COVID-19 outcomes, including the risk of incidence/infection, confirmed diagnosis, access to testing, hospitalization, ICU admission, death, and need for mechanical ventilation. According to the results of 11 studies [32, 33, 37, 43, 47, 55, 74, 75, 77, 79, 80] some disadvantaged racial/ethnic minority groups were at a higher risk of COVID-19 infection compared to their White counterparts or those from racial/ethnic majority groups in the region. In contrast, one study from the US did not report a significant association between a higher proportion of Black or Asian residents and a higher risk of COVID-19 incidence/infection [68].

Concerning COVID-19 deaths, nine studies [33, 42, 44, 48, 57, 61, 64, 65, 68] did not report significant associations between race/ethnicity and risk of COVID-19 death; 11 studies [23, 26, 32, 39, 46, 47, 55, 59, 69, 74, 79, 80] reported that racial/ethnic minority groups were at greater risk of dying from COVID-19. However, in contrast, two studies showed reverse association [49, 50]. There was also incongruity in the results of other studies [37, 41, 60, 62, 63, 66]. For example, a study in the US showed a statistically significantly increased hazard of in-hospital mortality for Hispanics and Asians. Nevertheless, there was no significant increased risk of in-hospital mortality for African Americans compared to Whites [60]. Another study from the US also concluded that Hispanics were two times more likely to die from COVID-19 than Whites, while there was a nonsignificant difference comparing Blacks and Whites [41]. According to another study from the US, compared with non-Hispanic White patients, Black patients and Hispanic patients were at a lower risk of death from COVID-19, whereas Asian patients had a higher risk of mortality [62]. Another study showed that although the US population of South Asians was at a higher risk of COVID-19 death compared with the rest of the population, there was no significantly higher risk of COVID-19 death for Blacks [63]. One study from the United Kingdom showed that whereas the risk of COVID-19 death was not statistically significantly different between Black or Mixed/Other racial/ethnic populations compared to Whites, Asians showed a higher risk of COVID-19 death than Whites [66]. A study from the US reported that while African American population was at a higher risk of death from COVID-19, this was not the case for Asians [37].

Concerning confirmed diagnosis, ten studies [27, 33, 41, 48, 50, 57, 58, 61, 67, 72] showed that racial/ethnic minority groups were more likely to have positive test results than the rest of the population tested for COVID-19. However, some studies reported inconsistent results for different racial/ethnic groups. For example, a study from the US showed a statistically significant association between the positive COVID-19 test results and belonging to Black, Hispanic, or Asian racial/ethnic populations. However, this association was reversed for the non-Hispanic Asian population living in the high-density populated areas [51]. On the other hand, one study showed that while the Latinx ethnic group had a higher likelihood of positive test results than non-Latinx patients, Black patients had fewer positive test results than non-Black patients [56]. See Table 2 for more information.

Table 2 Summary of the results of included studies
Full size table

A few studies examined the access to testing for COVID-19 among different racial/ethnic groups. One study showed that Black populations were more likely to get tested for COVID-19 than Hispanic and White populations [61]. One study did not report significant differences between race/ethnicity and testing for COVID-19 [68]. One study reported that racial/ethnic minority groups, including non-Hispanic Black, Hispanic, and Asian populations had lower population-weighted travel time to the testing sites than White populations [75]. In contrast, another study showed that areas with a higher proportion of the White population had a higher number of total tests [67]. One study showed that compared to other racial/ethnic groups (including Whites, Asians, and Hispanics), a smaller percentage of African Americans were tested for COVID-19 in an ambulatory setting. Most African Americans were tested in hospitals, either in the emergency department or during inpatient hospitalization [34].

Nine studies [33, 34, 41, 44, 47, 48, 54, 57, 66] showed that the racial/ethnic minority populations were more likely to be hospitalized for COVID-19 than majority racial/ethnic groups or the rest of the population. Nonetheless, one study from the US reported inconsistent results showing that while the Blacks were about twice as likely as Whites to be hospitalized for COVID-19, no statistically significant differences were found comparing Hispanics and non-Hispanic Whites [60]. Another study from the US found that although Asians were more likely to require hospitalization for COVID-19 than Whites, no statistically significant differences were found between Blacks or multiracial groups and Whites [64].

Three studies showed that racial/ethnic minority populations had a disproportionately higher ICU admissions rate than the rest of the populations [23, 42, 49]. One study from Ireland, for instance, reported that compared with White Irish patients, all other ethnic groups had an approximately fourfold increased risk of ICU admittance after adjusting for age [42]. Another study from Kuwait showed that South Asians had approximately six times higher odds of being admitted to the ICU when compared to Arabs [23]. One study from the US also reported that a smaller percent of White or non-Hispanic patients were admitted to ICU compared with non-White patients [49]. In contrast, two studies did not report a statistically significant association between race/ethnicity differences and risk of ICU admission for COVID-19 [44, 57]. Moreover, whereas one study from the US showed that Black patients were more likely than Whites to receive mechanical ventilation [48], another study from the US showed that there were no significant racial/ethnic differences in receiving mechanical ventilation [57]. A summary of the results is presented in Table 2.

Socioeconomic disparities and COVID-19 outcomes

Overall, 28 studies examined the association between COVID-19 outcomes and SES characteristics. In this regard, studies showed that poor housing conditions [38], living in poverty [24, 45, 75] or deprivation [76, 78], employment in the healthcare and social assistance and transportation industries [45], lack of insurance [45, 74], household overcrowding [24, 75], lower household income [24, 37, 43, 45, 70], no or lower level of education [74, 79], and urban residency [76] were associated with a higher risk of COVID-19 incidence/infection. Some studies showed contradicting results, reporting that areas with higher median income [37, 74] or lower poverty [74] had a higher infection rate. One study from the US, on the other hand, showed that areas with a higher proportion of uninsured individuals were associated with a lower rate of infection [68]. Moreover, a few studies demonstrated that households’ income differences [75], average household size [79], and lack of insurance [37, 79] were not significantly associated with the risk of COVID-19 incidence.

Regarding COVID-19 deaths, studies showed that a low level of education [40, 74, 79], poverty [24, 37], poor housing conditions [38], low family income [40], deprivation [42, 71, 78, 79], speaking in a language other than the national language in a country [71, 73], household overcrowding [24], being an immigrant [40], and unemployment [59, 71] were associated with a higher risk of death from COVID-19. Nevertheless, there were some contradicting results as well. For example, higher risks of death from COVID-19 was associated with lower rates of unemployment [74], a lower proportion of uninsured individuals in a region [68, 73], and higher household income [74, 79]. A few studies reported that low SES [52], poverty [57], large household size [79], a higher proportion of uninsured individuals in a region [79], and deprivation [52] were not associated with higher risks of COVID-19 death. However, one study reported that living in poverty was associated with fewer reported deaths in some counties and more reported deaths in some other counties [73].

Several studies corroborate that people living in urban areas [27], areas with higher levels of poverty [45], deprivation [27, 58], rental housing [45], lack of insurance [45], lower household income [45, 72], higher employment in the transportation and healthcare industries [45], and overcrowded households [53, 72] were more likely to have positive COVID-19 test results. However, there was some inconsistent evidence as well. For example, one study from the US did not report a statistically significant association between poverty status and confirmed diagnosis [57]. Another study from the US reported that higher median household income was associated with a higher likelihood of positive COVID-19 tests [74].

One study showed that the likelihood of getting a COVID-19 test was lower for non-English speakers than English speakers [53]. In one study, the number of total tests was not associated with the SES index, which was a composite score made of SES variables including household income, gross rent, poverty, education, working class, unemployment, and household density [67]. Another study showed that testing frequency was higher in areas with fewer uninsured individuals [68]. Three studies from the US showed that people living in densely populated areas [44] or areas with a higher poverty level [57] were at a higher risk of hospitalization for COVID-19. Two studies showed that people with Medicaid/Medicare or no reported insurance had a higher risk of hospitalization [34] and mortality [65] than those with commercial insurance. One study did not find an association between living in deprivation and hospitalization [52].

Regarding ICU admission, two studies did not report an association with deprivation [42, 52], and one study did not show an association with the SES index [52]. However, one study showed that individuals living in higher poverty levels were at higher risk of ICU admission for COVID-19 [57]. Regarding mechanical ventilation, one study did not report a statistically significant association between the need for mechanical ventilation and living in poverty among patients with COVID-19 [57].

Discussion

This systematic review studied the racial/ethnic and SES disparities in health outcomes and access to health resources during the COVID-19 pandemic. Due to the inconsistencies among studies, especially regarding different definitions for racial/ethnic minority groups and low-SES, it was challenging to summarize the results. Generally, evidence showed that racial/ethnic minority populations were at a higher risk of infection, having positive test results, and hospital admissions for COVID-19 [32, 33, 37, 43, 47]. Studies on the risk of death from COVID-19 among different racial/ethnic populations and those with different SES showed different results. Regarding access to testing, the inconsistency was much more evident. Two studies showed greater access to testing for racial/ethnic minority populations [61, 75], two studies yielded reverse results [34, 67], and one study showed no significant differences [68]. We also found similar results for ICU admission following the COVID-19 infection.

Nineteen out of 28 studies showed that people from low SES, including those with poor housing conditions, poverty, household overcrowding, and lower level of education, were at higher risk of infection [24, 38, 43, 45, 68, 70, 75, 76, 78, 80], death [24, 37, 40, 42, 71, 74, 78–80] and confirmed diagnosis [27, 45, 58, 81]. However, this was not always the case [37, 52, 57, 73–75, 80, 82]. For instance, according to the results of four studies, areas with a higher median income had a higher rate of infection, death, or confirmed diagnosis [37, 73, 74, 80].

Lack of insurance and unemployment affected the outcomes in different ways. Out of 28 studies, four studies reported a positive association between lack of insurance and COVID-19 incidence [45, 74] and death [68, 73]. Moreover, two studies did not report a notable association between lack of insurance and risk of COVID-19 death and COVID-19 incidence/infection [37, 79]. Just one study reported lack of insurance as a predictor of decreased risk of COVID-19 incidence [68]. Out of 28 studies on SES, two studies showed that a higher rate of unemployment was associated with a higher risk of death from COVID-19 [59, 71]. In contrast, one study reported that a higher risk of death from COVID-19 was associated with lower rates of unemployment [74]. One possible explanation for the inconsistent results regarding unemployment and lack of insurance is that information regarding the health outcomes of unemployed and uninsured individuals might not be wholly documented due to the underutilization of services. More studies are needed to explore the impacts of unemployment and lack of insurance on patients with COVID-19.

Though racial/ethnic minority groups were frequently identified as the most vulnerable populations during the epidemics, they are exceptionally vulnerable in the COVID-19 pandemic because the transmission of the infection is strongly associated with the background and socioeconomic characteristics of individuals. There are some potential reasons for the higher incidence and severity of COVID-19 in racial/ethnic minority groups and individuals from low SES. When speaking about SES, we generally focus on people’s occupation, income, and education level [83]. The risk of the transmission of COVID-19 in professions with constant in-person interactions is higher than in other professions. As a result, the incidence of the disease among service industry workers is higher [84]. Additionally, people with low SES are more likely to experience work stress, which increases the risk of cardiovascular disease [85] and disrupts the function of the immune system [86], consequently resulting in lower resistance to COVID-19. Low household income can influence the housing conditions of the individuals to increase the risk of the spread of infectious diseases among those living in small and overcrowded housing units [6]. Lower education levels may increase the COVID-19 severity indirectly by behavioral pathways, poor diet, smoking and other risk factors, and problems with effectively navigating healthcare systems [87]. Racial/ethnic minorities are usually impacted by higher poverty rates, are economically disadvantaged, and are more likely to work in jobs unsuitable for remote working [58, 88–90].

This systematic review can inform policymakers, practitioners, and researchers of the potential inequalities in health outcomes and access to services, helping them adopt effective strategies to manage COVID-19 as a public health emergency. There is no doubt that most of the racial/ethnic minorities and those from low SES are more vulnerable to COVID-19; therefore, the information provided by this review study can provide authorities with insight into the inequalities that COVID-19 poses to these vulnerable populations.

Limitations

We acknowledge the limitations of this study. First, race/ethnicity and SES data might have been incomplete (or inaccurate in some cases) in medical records and datasets, and their accuracy evaluation were not reported [6]. This could have made it difficult for researchers to measure health disparities in COVID-19 and have resulted in major inconsistencies among the studies, making it challenging to congregate the results. Additionally, many LMICs do not have appropriate information systems to provide quality data [25], especially regarding case ascertainment and SES information, and in some cases, the assignment of cause-of-death. As a result, there are fewer studies from LMICs. Second, the diagnostic tests and degree of accuracy (sensitivity and specificity) might vary from study to study, resulting in the misclassification of infected and healthy individuals. Finally, we only included the English-language studies in the review. As a result, we might have missed several studies from non-English language publications. Heterogenous populations and different levels of study design (individual patient-level or spatial analysis) can also contribute to contradicting findings.

In order to mitigate the limitation to some degree, we extracted detailed information from each study and summarized them in the data extraction sheet. This would help readers to review the summary results of each study in line with their potential limitations. Despite the aforementioned limitations, we believe that this systematic review can provide insight into the status of the disparity in the COVID-19 and can contribute to future research in the field.

Conclusion

Although we observed the inconsistencies among the included studies, most of these studies showed that racial/ethnic minority populations in a region are at greater risk of COVID-19 infection, hospitalization, confirmed diagnosis, and mortality. Additionally, most of the studies cited factors such as low level of education, poverty, poor housing conditions, low family income, speaking in a language other than the national language in a country, and household overcrowding as risk factors of COVID-19 incidence, death, and confirmed diagnosis. The potential impact of lack of insurance and unemployment on the outcome measures such as the need for mechanical ventilation, ICU admission, and access to testing for COVID-19 were limited and inconsistent. Further studies are needed to fill the gaps. This systematic review also revealed a major incongruity in the definition of minority ethnic/race groups and SES among the studies.