Data suggest that impacts of COVID-19 differ among U.S. racial/ethnic groups. This systematic review evaluates racial/ethnic disparities in SARS-CoV-2 infection rates and COVID-19 outcomes, factors contributing to disparities, and interventions to reduce them.
Abstract Background:Data suggest that the effects of coronavirus disease 2019 (COVID-19) differ among U.S. racial/ethnic groups.
Purpose:To evaluate racial/ethnic disparities in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection rates and COVID-19 outcomes, factors contributing to disparities, and interventions to reduce them. (PROSPERO: CRD42020187078)
Data Sources:English-language articles in MEDLINE, PsycINFO, CINAHL, Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews, and Scopus, searched from inception through 31 August 2020. Gray literature sources were searched through 2 November 2020.
Study Selection:Observational studies examining SARS-CoV-2 infections, hospitalizations, or deaths by race/ethnicity in U.S. settings.
Data Extraction:Single-reviewer abstraction confirmed by a second reviewer; independent dual-reviewer assessment of quality and strength of evidence.
Data Synthesis:37 mostly fair-quality cohort and cross-sectional studies, 15 mostly good-quality ecological studies, and data from the Centers for Disease Control and Prevention and APM Research Lab were included. African American/Black and Hispanic populations experience disproportionately higher rates of SARS-CoV-2 infection, hospitalization, and COVID-19–related mortality compared with non-Hispanic White populations, but not higher case-fatality rates (mostly reported as in-hospital mortality) (moderate- to high-strength evidence). Asian populations experience similar outcomes to non-Hispanic White populations (low-strength evidence). Outcomes for other racial/ethnic groups have been insufficiently studied. Health care access and exposure factors may underlie the observed disparities more than susceptibility due to comorbid conditions (low-strength evidence).
Limitations:Selection bias, missing race/ethnicity data, and incomplete outcome assessments in cohort and cross-sectional studies must be considered. In addition, adjustment for key demographic covariates was lacking in ecological studies.
Conclusion:African American/Black and Hispanic populations experience disproportionately higher rates of SARS-CoV-2 infection and COVID-19–related mortality but similar rates of case fatality. Differences in health care access and exposure risk may be driving higher infection and mortality rates.
Primary Funding Source:Department of Veterans Affairs, Veterans Health Administration, Health Services Research & Development.
The health effects of coronavirus disease 2019 (COVID-19) due to spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) have been unevenly distributed in the United States. Infections, hospitalizations, and deaths have varied among and within regions and communities, prompting questions about which populations are at higher risk and why. Health disparities, defined as a higher burden of illness, injury, disability, or mortality among one group relative to another, are well documented in the United States (1, 2). In recent history, the 2009 H1N1 pandemic resulted in higher rates of hospitalizations and mortality among African American/Black and Hispanic populations (3). Early reports on disparities in COVID-19 have shown similar patterns (4, 5).
According to a framework that describes the progress of health disparities research, once health disparities have been detected, the following phases of research are focused on identifying the factors driving disparities and designing and testing interventions to mitigate them (6–8). In contrast to past public health crises that were better understood after the fact, we are detecting disparities in COVID-19 in nearly real-time owing to online access to data sets, sharing of scientific work before the journal peer-review process (online posting of preprints), and use of social media (9). At this pace, it is possible to progress quickly to the intervention phase of health disparity research while the pandemic is still ongoing. The aim of this systematic review is to synthesize evidence on racial and ethnic disparities related to COVID-19, factors underlying them, and effectiveness of interventions to reduce them.
MethodsThis article is based on a systematic review conducted by the Department of Veterans Affairs (VA) Evidence Synthesis Program (ESP) which examines health disparities in COVID-19 and past public health crises (PROSPERO registration: CRD42020187078). Key questions were developed by 3 of the authors (D.K., K.K., S.S.) and were revised with input from the Veteran Health Administration's (VHA) Office of Health Equity. We followed standard methods and reporting guidelines for systematic reviews (10, 11).
Data Sources and SearchesWe searched the literature in 2 phases: from database inception through 3 June 2020 and then through 31 August 2020. Using terms related to COVID-19 and disparities, we searched the following sources for English-language articles: MEDLINE ALL (Ovid), PsycINFO, CINAHL, the Cochrane Central Register of Controlled Trials and the Cochrane Database of Systematic Reviews (Ovid EBM Reviews), and Scopus (conference proceedings only). Initially, we also searched the preprint database medRxiv.org for relevant articles. We searched gray literature sources for data on our outcomes of interest through 2 November 2020 (the Supplement shows the complete search strategy).
Study SelectionWe included observational studies that examined SARS-CoV-2 infections, hospitalizations, and deaths stratified by race or ethnicity in U.S. settings. We included 2 types of studies examining SARS-CoV-2 infection: those based on polymerase chain reaction (PCR) testing and seroprevalence studies based on antibody testing (immunoassays for IgG, IgM, or total antibodies). We included all studies with unadjusted results for infection rates, reasoning that infection rates are minimally confounded by such factors as age and comorbid conditions. However, we excluded cohort and cross-sectional studies examining disparities in hospitalizations and deaths if results were not at least age-adjusted. We would have included observational studies or trials of interventions to mitigate disparities, but we found none.
We modified our study selection criteria for our second search, given an increase in the quality and number of studies we identified. We initially included preprints but excluded them from our second search. We also initially included ecological studies (studies of population-level outcome data, such as census tract, ZIP code, and state) but excluded them from our second search for 2 reasons. First, we found that results of individual-level and population-level studies from our first round of searching were largely consistent. Second, given the larger volume of relevant literature identified in our second search, we prioritized studies of individual-level data, on the rationale that they would more reliably address our key questions. Associations between outcomes and racial/ethnic composition in given region may not correspond to differences among racial/ethnic groups within regions (ecological fallacy). Two authors (D.K., H.S., K.K., K.M., M.R., or S.Y.) examined titles and abstracts for potential relevance, and 2 authors (C.A., D.K., K.K., K.M., S.A., or S.Y.) independently reviewed full-text articles for inclusion.
Data Extraction and Quality AssessmentOne author (C.A., H.S., K.M., M.R., or S.Y.) abstracted details of study setting, population, exposures, and outcomes of interest, and a second author (C.A., D.K., K.K., K.M., M.R., S.A., S.S., or S.Y.) verified accuracy. Two authors (C.A., D.K., J.A., K.K., K.M., S.A., S.V., or S.Y.) independently assessed study methodological quality by using adapted versions of the Newcastle-Ottawa Quality Assessment Scale (12).
Data Synthesis and AnalysisWe synthesized evidence qualitatively and did not perform meta-analyses owing to limited or differently reported data and study heterogeneity (for example, different covariates used in adjusted models). To synthesize evidence on factors mediating disparities, we used the framework developed by Quinn and Kumar and their colleagues to categorize factors into 3 groups related to exposure, susceptibility, and health care access (13, 14).
Two reviewers (K.M. and D.K.) independently rated the strength of the body of evidence using criteria that assessed study limitations, directness of the population studied and the outcomes measured, consistency of results across studies, and precision of effect estimates (15). All authors discussed strength of evidence assessments to achieve consensus. For this review, we applied the following general algorithm: Evidence from multiple large methodologically sound studies with consistent results received a rating of “high”; evidence from fewer studies or studies with smaller sample sizes but consistent results received a rating of “moderate”; and this same type of evidence with inconsistent results received a rating of “low.”
Role of the Funding SourceThe authors did not receive funding for this study outside of salary support from their respective institutions.
ResultsThe study flow diagram (Figure) shows the search and study selection processes (10). We included 52 studies plus data retrieved on 2 November 2020 from the Centers for Disease Control and Prevention (CDC) and APM Research Lab (16–18). We identified 37 cohort or cross-sectional studies of individuals tested for or diagnosed with SARS-CoV-2 and 15 ecological studies (Appendix Table 1) (16–47-48–70). Among studies of individuals tested for SARS-CoV-2, sample sizes ranged from 121 to more than 62 000, and most studies were conducted within a single hospital or health care system. Two cohort studies and 9 ecological studies examined disparities across the United States, and the remainder focused on specific cities or states. Studies most frequently evaluated SARS-CoV-2 infections, followed by COVID-19 deaths and hospitalizations, and most frequently evaluated these outcomes for African American/Black and Hispanic populations. Data regarding Asian, American Indian/Alaska Native, Native Hawaiian, Pacific Islander, and other populations were much less commonly reported. Studies adjusted for a wide variety of covariates, which are categorized according to the Quinn–Kumar framework in Supplement Table 1.
Figure. Study flow diagram.CCRCT = Cochrane Central Register of Controlled Trials; CDSR = Cochrane Database of Systematic Reviews; EBM = Evidence-Based Medicine;
* Exclusions applied to July–August search results (second literature search) owing to the high volume of new studies, including those based on individual-level (rather than population-level) data.
† Includes 12 preprints from April–June publications (first literature search) as well as 6 published studies that were first identified as preprints.
Appendix Table 1. Characteristics of Included Studies. Appendix Table 1. Characteristics of Included Studies—Continued. Appendix Table 1. Characteristics of Included Studies—Continued. Appendix Table 1. Characteristics of Included Studies—Continued.Four (11%) cohort and cross-sectional studies were good methodological quality and 25 (68%) were fair quality. Reasons that studies were assessed to be fair (rather than good) quality included lack of reporting of missing data; lack of adjustment for potential confounders; and unclear assessment of outcomes, including whether outcomes were assessed for all participants and whether follow-up was adequate. Eight (22%) cohort and cross-sectional studies were poor methodological quality due to possible selection bias and/or high levels of missing data (defined as ≥20%). Most (80%) ecological studies were good-quality. Supplement Tables 2 and 3 present quality assessments.
SARS-CoV-2 InfectionsEvidence suggests that African-American/Black and Hispanic populations experience higher rates of SARS-CoV-2 infection compared with non-Hispanic White populations (Appendix Table 2). Among 15 cohort and cross-sectional studies (13 fair-quality, 2 poor-quality) comparing the risk for a positive SARS-CoV-2 PCR test between African American/Black and White populations, including 10 large studies of more than 1000 individuals each, all but 2 studies detected a disparity (19, 21, 23-25, 29, 32, 34–36, 40, 45, 47, 51, 53). The studies detecting a disparity estimate that African American/Black populations have a 1.5 to 3.5 times higher risk for infection than White populations. Our confidence in these findings is high, meaning that future studies are very likely to show similar estimates (both in direction of findings and magnitude). These results are supported by evidence from 6 ecological studies (4 good-quality, 2 fair-quality) (Supplement Table 4), which also consistently found that a higher percentage of African American/Black individuals in a given population was associated with a higher rate of diagnosed COVID-19 (55, 58, 60, 62–64). Two seroprevalence studies also identified disproportionate infection rates among African American/Black populations (30, 48). Of the 3 studies that did not find a difference, 2 were small studies of unique participants (homeless patients and pregnant or postpartum patients) (23, 29). The third study was a large seroprevalence study of health care personnel in the New York City region (42). It is unclear why findings from this study are inconsistent with others.
Appendix Table 2. Cohort and Cross-sectional Studies of SARS-CoV-2 Infection and Seroprevalence Rates, by Race/Ethnicity.Similarly, evidence from 13 of 19 cohort and cross-sectional studies (1 good-quality, 15 fair-quality, 3 poor-quality), including 8 large studies of more than 1000 individuals each, suggests that Hispanic populations have higher rates of SARS-CoV-2 infection (according to SARS-CoV-2 PCR or serologic tests) compared with non-Hispanic White populations (Appendix Table 2) (19–21, 23, 24, 26, 27, 29, 32, 35, 36, 40, 42, 47, 48, 51, 53). We have moderate confidence in these findings (rather than high) given that results were less consistent overall (compared with findings for African American populations) and estimates of the magnitude of increased risk varied. For example, 8 studies conducted in a hospital or health care setting found that Hispanic populations had a 1.3 to 7.7 times higher risk for a positive SARS-CoV-2 PCR result compared with non-Hispanic White populations, but a community-based study of residents and workers within a San Francisco census tract found that the risk was more than 28 times higher (20, 24, 26, 29, 32, 40, 47, 51, 53). Results of the community-based study, in which more than 52% of individuals with a positive PCR result were asymptomatic at the time of testing, may signal that disparities in infection rates are more pronounced than is being detected through testing in health care settings (26). Eight ecological studies detected a disparity in infection rates among Hispanic populations compared with non-Hispanic White populations or generally among minority populations compared with non-Hispanic White populations (55, 56, 59–61, 65, 67, 68). The studies that did not detect a disparity in SARS-CoV-2 infections included the large seroprevalence study of health care personnel discussed above that also did not identify a disparity for African American/Black populations (42). Other studies not finding a disparity in infection rates among Hispanic populations had low representation of Hispanic persons (ranging from less than 1% to 9%) and may not have been adequately powered to detect a difference or were small studies of unique participants (19, 23, 31, 36, 45).
Among 7 cohort and cross-sectional studies that evaluated SARS-CoV-2 infection risk (according to SARS-CoV-2 PCR or serologic tests) in Asian populations, 6 found no difference and 1 identified a higher risk for infection compared with White populations (23, 24, 30, 36, 42, 45, 53). The only ecological study to evaluate infections among Asian populations specifically identified a higher risk (55). Given less data overall and inconsistency, we have low confidence in these findings. It is likely that more nuanced studies of specific subpopulations and settings will have different results.
American Indian/Alaska Native, Pacific Islanders, and other racial/ethnic groups were not studied sufficiently to draw any conclusions regarding disparities in infection risk.
COVID-19 HospitalizationsEvidence suggests that African-American/Black populations have a higher risk for hospitalization due to SARS-CoV-2 compared with White populations (Table 1). Among 11 cohort and cross-sectional studies (1 good-quality, 10 fair-quality) evaluating hospitalization rates, results of 7 studies suggest that African-American/Black populations are 1.5 to 3 times more likely to be hospitalized compared with White populations (19, 22, 28, 32, 34, 37, 39, 43, 44, 46, 52). These findings are supported by data from the CDC's COVID-19-Associated Hospitalization Surveillance Network (COVID-NET), which show that African-American/Black populations have a 4 times higher risk for hospitalization compared with White populations (16). We have moderate confidence in these findings (rather than high) because not all study findings are consistent. Results of 4 studies, including a large nationally representative retrospective cohort study conducted in the VHA setting, did not identify a statistically significant difference in hospitalization rates (28, 43, 46, 52). Reasons for this inconsistency are unclear. Although future studies are likely to find that African-Americans have disproportionately higher rates of hospitalization due to SARS-CoV-2, the size of this disparity may vary by geographic region, health care access, and other factors not yet identified.
Table 1. Cohort and Cross-sectional Studies of COVID-19 Hospitalizations, by Race/Ethnicity.Five cohort studies (1 good-quality, 3 fair-quality, 1 poor-quality) found that Hispanic populations have a higher risk for hospitalization compared with White and non-Hispanic populations, although this finding was only statistically significant in 2 studies, which both found that the risk is 1.5 times higher (Table 1) (19, 22, 28, 32, 43). We have moderate confidence in this finding, given that the direction of results (evidence of a disparity) is consistent. Moreover, results of cohort studies are supported by data from COVID-NET showing that Hispanic persons have more than 4 times the risk for hospitalization compared with non-Hispanic White persons (16).
Hospitalization rates appear to be equal for Asian populations compared with non-Hispanic White populations on the basis of 2 retrospective cohort studies and COVID-NET data (16, 22, 43). However, our confidence in this finding is low, given that hospitalization rates among Asian persons have been infrequently studied and results to date may not be generalizable to diverse Asian subpopulations.
Hospitalization risk for other racial/ethnic groups has not been sufficiently studied to draw conclusions. No studies of individual-level data have focused on the American Indian/Alaskan Native population, although COVID-NET data suggest that this group has a 4 times higher risk for hospitalization compared with non-Hispanic White persons (16).
COVID-19 MortalityEvidence suggests that African-American/Black populations disproportionately account for COVID-19 deaths compared with non-Hispanic White populations (Table 2). Our confidence in this finding is high, meaning that future studies are very likely to show the same estimate of effect. Analysis of data from the CDC's National Center for Health Statistics (NCHS) shows that African-American/Black populations experience approximately 15% excess deaths (defined as the percentage of COVID-19 deaths for a racial/ethnic group compared with the percentage of that racial/ethnic group in the population), and data from APM Research Lab show that African American/Black populations have 3.2 times the risk for mortality compared with White populations (17, 18). In addition, 5 ecological studies (all good-quality) consistently found higher mortality among African-American/Black populations, although not all results are statistically significant (57, 60, 63, 64, 66).
Table 2. Cohort and Cross-sectional Studies of COVID-19 Deaths, by Race/Ethnicity.Evidence also suggests that Hispanic populations disproportionately account for COVID-19 deaths (Table 2). We have moderate (rather than high) confidence in this finding, given that fewer studies examined this outcome and results are less consistent. Data from NCHS shows that Hispanic populations have approximately 21% excess deaths, and data from APM Research Lab shows that Hispanic populations have 3.2 times the mortality risk compared with non-Hispanic White persons (17, 18). While results of studies evaluating mortality specifically for Hispanic populations are inconsistent, 2 ecological studies found that a greater proportion of non–English-speaking individuals and those with minority status in each county or state are associated with higher rates of COVID-19 deaths (57, 60, 61, 65). Conversely, a study of 1624 counties found that a higher percent minority population was associated with fewer COVID-19 deaths (69). Reasons for these different findings are unclear, and a limitation of this study was that county data on COVID-19 deaths were often not disaggregated by race or ethnicity.
Evidence suggests that Asian populations have similar COVID-19 mortality compared with White populations, whereas American Indian/Alaska Native and Pacific Islander populations have higher mortality rates. Data from NCHS shows that Asian populations experience 2.2% fewer deaths than would be expected on the basis of population (17), whereas data from APM Research Lab finds that Asian populations have 1.2 times the mortality risk compared with White persons (18). Data from NCHS shows that American Indian/Alaska Native populations experience 1.9% excess deaths (17), whereas APM Research Lab finds that indigenous populations and Pacific Islanders have 3.1 and 2.4 times the mortality risk compared with White persons, respectively (18). We have low confidence in these findings, given less data overall. More research is needed to draw stronger conclusions regarding disparities in COVID-19 mortality for these groups.
COVID-19 Case FatalityIn contrast to overall mortality, which reflects deaths due to diagnosed and undiagnosed COVID-19, case fatality reflects deaths among those with confirmed COVID-19. Evidence from 15 cohort and cross-sectional studies (2 good-quality, 10 fair-quality, and 3 poor-quality) suggests that there is not disparity in case fatality by race/ethnicity (Table 2) (19, 32, 34, 38, 39, 41, 43, 44, 47, 49, 50, 52–54, 70). We have moderate confidence in this finding for African-American/Black and Hispanic populations. Despite the consistency of findings, our confidence is not higher because studies were predominantly conducted within a single health care system, allowing only a determination of case fatality within that system but not across systems. Exceptions were a small study of women with gynecologic cancer and COVID-19 treated in 6 different hospital systems in New York City and a larger study of 154 acute care hospitals in 13 states (data from COVID-NET) (38, 39). These studies also did not identify a racial/ethnic disparity in case fatality rates, but future studies explicitly designed to evaluate case fatality rates by race/ethnicity including different hospitals and health systems may have different findings.
Similarly, no disparity in case fatality was found in 2 studies including Asian populations, although our confidence in these findings is low, given less data overall (43, 53). We identified no studies or data reporting case-fatality for other racial/ethnic groups.
Factors Underlying Racial/Ethnic Disparities in COVID-19A subset of 20 studies (10 individual-level and 10 ecological) used 1 or more statistical models to control for variables that could influence SARS-CoV-2 infection rates, hospitalizations, and/or deaths. These variables may be categorized into factors affecting exposure, susceptibility, and health care access by using the Quinn–Kumar framework (Supplement Table 1) (12, 13). A smaller subset of studies (5 cohort and 2 ecological) used a series of statistical models sequentially incorporating susceptibility and/or exposure and health care access variables to determine which factors explained the observed disparities (Supplement Table 5) (22, 32, 34, 44, 47, 63, 65). Overall, results of these models suggest that exposure and health care access variables underlie COVID-19-related disparities more than susceptibility (that is, comorbid conditions), although our confidence in this finding is low. When models controlled for susceptibility factors only, observed disparities persisted in 3 studies (22, 47, 65). However, in full statistical models incorporating susceptibility, exposure, and health care access, the magnitude of the disparity decreased or the disparity was no longer observed in five studies (22, 34, 44, 47, 65). Further supporting the concept that exposure factors may be largely driving COVID-19 racial/ethnic disparities, results of a study of more than 20 000 adults tested for SARS-CoV-2 in Houston found that that population density explained disparities in infection rates for non-Hispanic Black populations (odds ratio [OR], 1.03 [95% CI, 1.01 to 1.05]) and population density and income explained infection-rate disparities for Hispanic populations (OR, 1.02 [CI 1.01 to 1.02] and OR, 1.04 [CI, 1.02 to 1.06], respectively) (51). Moreover, the lack of difference in COVID-19 case-fatality rates by race/ethnicity also suggests that exposure-related factors are contributing more to disparities than susceptibility. If susceptibility was driving disparities, we would expect to find a disproportionate rate of in-hospital deaths among minority groups, which so far has not been the case.
Interventions to Mitigate Racial/Ethnic Disparities in COVID-19We did not identify any studies addressing interventions to mitigate racial/ethnic disparities in COVID-19.
DiscussionWe conducted this systematic review to synthesize evidence on racial and ethnic disparities during the COVID-19 pandemic in the United States, specifically disparities in SARS-CoV-2 infections, hospitalizations, and deaths. Identifying health disparities is the first step in disparities research, followed by identifying which factors are driving disparities and designing and testing interventions to mitigate them. This review helps advance this research agenda with 6 main findings. First, African-American/Black populations experience disproportionately higher SARS-CoV-2 infection rates and excess mortality due to COVID-19 (high strength of evidence) but not higher case-fatality rates (moderate strength of evidence). Second, Hispanic populations experience disproportionately higher infection rates and excess mortality due to COVID-19, but not higher case-fatality rates (moderate strength of evidence). Third, African American/Black and Hispanic populations have an increased risk for hospitalization due to COVID-19 (moderate strength of evidence). Fourth Asian populations appear to have similar rates of infections, hospitalizations, and deaths as White populations (low strength of evidence). Fifth, American Indian, Alaska Native, and Pacific Islander populations experience excess mortality due to COVID-19 (low strength of evidence). Finally, observed disparities are more likely to be due to exposure-related factors than susceptibility (that is, comorbid conditions) (low strength of evidence). To our knowledge, this is the first systematic review to comprehensively characterize racial/ethnic health disparities in COVID-19 in the United States and the factors underlying observed disparities. We only identified one other relevant review in our search of a COVID-19 review repository (www.covid19reviews.org/) (search date 2 November 2020), which was not limited to U.S. settings (71).
Given that SARS-CoV-2 testing to date has largely occurred in the health care context (rather than in community-based settings), findings are most applicable to populations seeking or engaged in health care. Findings are also largely derived from regions most affected by the first six months of the COVID-19 pandemic (mostly large cities), although CDC and APM Research Lab data are updated regularly and we reported their findings through early autumn 2020. It is unclear how the disparities discussed in this review will change as the COVID-19 pandemic evolves, particularly considering how closely linked access to testing is with SARS-CoV-2 infection rates, hospitalization rates, and case fatality rates. For example, if testing rates increase or testing expands to more community-based settings, we may find that observed disparities in test positivity become smaller if more testing leads to more negative test results. Alternatively, more testing may uncover more asymptomatic infections and widen the observed disparities. Similarly, changes in testing may have downstream effects on the observed disparities in hospitalization rates if the percentage of those tested requiring hospitalization increases or decreases. Population-based mortality is less affected by testing rates and will ultimately be the most informative outcome to monitor as the pandemic evolves.
This evidence base has limitations. First, 23% of studies (12 of 52) are preprints and have not gone through a formal journal peer-review process. Although we performed quality assessment of each study, the overall vetting process for these studies has not been as rigorous as for studies published in peer-reviewed journals, and we may have missed errors in reporting or data analysis. Second, most cohort and cross-sectional studies had varying levels of incomplete data on race and ethnicity and handled these missing data in different ways (either by excluding participants from analysis or grouping them into “unknown” race). The implications of these missing data in terms of our synthesis and strength of evidence ratings is unknown. Third, disparities for African-American/Black and Hispanic populations were more frequently described than disparities for other racial/ethnic groups. More research is needed to understand the magnitude and nuances of disparities for other populations and subpopulations, especially for Asian subpopulations and American Indian/Alaska Native and Pacific Islander populations overall. Fourth, owing to the nature of available data, cohort and cross-sectional studies infrequently evaluated exposure and health care access covariates, whereas ecologic studies infrequently evaluated susceptibility covariates. The result is that no study provides a complete picture of the relative importance of exposure, susceptibility, and health care access on racial/ethnic disparities for a given population. The number, type, and definition of these covariates also varied considerably across studies, making it difficult to quantitatively synthesize results. Finally, findings of ecological studies could be confounded by differences in local public health policies or programs.
Limitations of our methods include our focus on racial/ethnic health disparities. Several other types of disparities likely exist, including those related to socioeconomic status, disability status, and urban or rural location. A second limitation of our methods was our search process. We may have missed studies that stratified SARS-CoV-2 infection rates and other outcomes by race/ethnicity if the information was in an online appendix or otherwise not prominently featured in the text.
In conclusion, moderate- to high-strength evidence from this systematic review of 52 studies and analysis of data from the CDC and APM Research Lab finds that African-American/Black and Hispanic populations experience a disproportionate burden of SARS-CoV-2 infections and COVID-19-related mortality, but not higher case-fatality rates (mostly defined as in-hospital morality). Evidence is insufficient to draw strong conclusions regarding disparities in COVID-19 for other racial/ethnic groups. Increased susceptibility to COVID-19 does not seem to explain the observed disparities, but more evidence is needed to confirm this finding and evaluate the effects of health care access and exposure-related factors, such as population density. Most urgently, interventions to mitigate these disparities need to be designed and tested.
Supplementary Material FootnotesThis article was published at Annals.org on 1 December 2020
ReferencesThis section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary MaterialsRetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.3