In December 2019, the novel, severe acute respiratory syndrome coronavirus 2 caused an outbreak of severe respiratory infections in Wuhan, China. This infection is currently identified as coronavirus disease 2019 (COVID-19).
In the United States, the total number of cases as of August 27, 2020, exceeded 5.7 million people (1); of these, approximately 99,200 were in Michigan (2). Of all the counties in Michigan, Wayne County (which contains Detroit) has the largest absolute number of cases.
The literature has identified certain risk factors associated with a poorer prognosis in patients infected with COVID-19 (3–6). Although the medical literature is mixed with regard to racial disparities in other disease types (7–12), recent data suggest that there are racial differences in patients with COVID-19 (13,14). However, it remains unclear whether these findings are applicable to an ICU population of patients. We sought to characterize the cohort of COVID-19-positive patients admitted to the ICU in our hospital, with attention to whether there are racial differences in ICU length of stay (LOS) and other outcomes.
MATERIALS AND METHODS
Henry Ford Hospital, a tertiary-care hospital serving the racially diverse, metropolitan Detroit region, as of August 31, 2020, has admitted 1,832 patients diagnosed with COVID-19 (15). Henry Ford Hospital consists of a 156-bed ICU and 68 dedicated medical ICU beds with additional specialized surgical, cardiac, and neuroscience ICUs that occasionally serve as overflow for medical patients. As the surge progressed, these ICUs were employed systematically to accommodate all COVID-19 patients. ICU admissions represent approximately one-third of admissions at our hospital and is the site of most hospital deaths from COVID-19 (4,5). We performed a single-center retrospective cohort analysis among adult patients (age > 18 yr old) admitted to the ICU at a large urban teaching hospital.
Consecutive patients with laboratory-confirmed COVID-19 diagnosed using quantitative reverse transcriptase-polymerase chain reaction and managed in the ICU between March 13, 2020, and July 31, 2020, were included in the study. Patients who were clinically suspected of having COVID-19 but who ultimately tested negative were excluded. For patients admitted to the ICU greater than one time, the initial admission was exclusively evaluated in our study. The primary outcomes were ICU LOS and need for mechanical ventilation during the ICU stay and inhospital and 28-day mortality.
Race and ethnicity in the electronic health record are self-reported. We binarized to “people of color” (POC; Black, Arab, Hispanic/Latino, and Asian) or “White” (Caucasian).
Acute respiratory distress syndrome (ARDS) was defined according to the Berlin Criteria (16); mild ARDS is defined as a Pao2/Fio2 ratio of 200–300, moderate ARDS as a Pao2/Fio2 ratio of 100–200, and severe ARDS as a Pao2/Fio2 ratio of less than 100.
All analyses were performed with the STATA software, Version 16.1 (Stata Corp, College Station, TX).
Continuous variables were compared using the Student t test or the Kruskal-Wallis rank-sum test in the cases of nonnormally distributed variables and expressed, respectively, as means ± sd or median and interquartile range. Categorical variables were expressed as percentages and analyzed using a chi-square test.
We used univariate logistic regression to identify independent risk factors for need for mechanical ventilation and mortality. Cox proportional hazards model was used for estimating survival time. The variables used in all these analyses included race/ethnicity, sex, median income by zip code, body mass index, chronic obstructive pulmonary disease, asthma, obstructive sleep apnea/obesity hypoventilation syndrome, home oxygen use, diabetes mellitus, coronary artery disease, congestive heart failure, chronic kidney disease, end-stage renal disease, smoking status, venous thromboembolic disease, alcohol abuse, drug abuse, HIV status, outpatient medications (immunosuppressives, aspirin, systemic and inhaled corticosteroid, 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors, angiotensin converting enzyme inhibitor, angiotensin receptor blocker, nonsteroidal anti-inflammatory drug, and anticoagulant), and inpatient administration of steroids and hydroxychloroquine. Variables with a p value of less than or equal to 0.1 were included in the multivariate regression models.
For survival analysis, time to intubation was measured from the onset of symptoms prior to hospital presentation. For 28-day mortality, time to death was measured from day of admission. For patients never intubated, censored dates included time of ICU discharge or death. The validity of the proportionality assumption was verified by the Schoenfeld test. Median ICU LOS was calculated using Kaplan-Meier estimates and death was considered as a censored event.
This study was approved with waiver of informed consent by the Institutional Review Board of Henry Ford Hospital, Detroit, MI (13739).
Of the 365 patients admitted to the ICU during the study period, 219 (60.0%) were Black, 129 (35.3%) were White, two (0.6%) were Asian, eight (2.2%) were Hispanic/Latino, and seven (1.9%) were Arab. Because of the low numbers in several ethnicities, this study focused on two groups: POC (Black, Asian, Hispanic/Latino, and Arab) and White.
Sex was equally distributed between the groups. POC were younger than Whites (62.8 vs 67.1, respectively; p = 0.007). POC had less coronary artery disease than Whites (34 [14.4%] vs 35 [27.1%]; p =0.003) and less self-reported use of regular alcohol consumption (50 [21.2%] vs 12 [9.3%]; p = 0.004). There was no difference in median income by zip code of residence. Overall, 279 patients (76.4%) received steroids during the hospitalization (Table 1).
TABLE 1. -
Clinical Baseline Characteristics of Patients
||Persons of Color (n = 236)
||White (n = 129)
||62.8 ± 15.3
||67.1 ± 13.6
| Chronic obstructive pulmonary disease
| Obstructive sleep apnea/obesity hypoventilation syndrome
| Coronary artery disease
| Congestive heart failure
| Chronic kidney disease
| Deep vein thrombosis history
| Angiotensin-converting enzyme inhibitor
| Angiotensin-receptor blocker
| Nonsteroidal anti-inflammatory drugs
| Steroids, per oral
| Steroids, inhaled
|Median income, by zip code (95% CI)
Values are mean ± sd or n (%), unless otherwise indicated.
Of the 365 patients, 255 (69.9%) had ARDS. Of these patients, 23 (9.0%) had mild disease, 96 (37.6%) had moderate disease, and 136 (53.3%) had severe disease. The number of patients with ARDS and severity of disease was similar between the two groups (Fig. 1).
The number of days of symptoms prior to presentation to the hospital was similar in POC and Whites (5 [CI, 3–7] vs 4 [CI, 2–7]; p = 0.140). The presenting symptoms (constitutional, respiratory, and GI) were also similar between the groups (data not shown).
Overall, in our cohort, the median ICU LOS was 18 days (CI, 7–47 d). After adjusting for confounders and censoring deaths in the ICU, there was no difference in ICU LOS between POC (18 d [CI, 7–47] d) and Whites (18 d [CI, 6–48 d]; p = 0.979).
During the study period, 263 of 365 patients required invasive mechanical ventilation (72.1%) at some time during their ICU stay. The frequency of mechanical ventilation was similar in POC (171 [72.5%]) and Whites (92 [71.3%]) (p = 0.817) (Table 2). The median time from symptom onset to intubation was similar for POC (9 d [CI, 6–15 d]) and Whites (10 d [CI, 5–16 d]; adjusted hazard ratio 0.954 [CI, 0.726–1.25]; p = 0.733) (Fig. 2A).
TABLE 2. -
Predictors for Invasive Mechanical Ventilation
||Adjusted OR (95% CI)
|Persons of color
|Age ≥ 65 yr old
|Systemic steroid use upon presentation
|Chronic kidney disease
Overall, 180 patients (49.3%) died within 28 days of admission and 188 patients (51.5 %) died in the hospital. The 28-day mortality was lower in POC (107/236; 45.3%) than in Whites (73/129; 56.6%) (odds ratio 0.636 [CI, 0.412–0.98]; p = 0.040). This difference remained significant after adjusting for confounders. There was a similar trend with inhospital mortality, but the difference was not statistically significant after adjusting for confounders (Table 3). Of the nine patients who died in the hospital after day 28, seven were POC and two were White. Survival time was longer in POC than Whites (Fig. 2B; hazard ratio 0.627 [p = 0.010]). After adjusting for confounders, this difference remained significant (Fig. 2C; Supplemental Table 1, Supplemental Digital Content 1, https://links.lww.com/CCM/F990). All deaths in the ICU occurred during the index ICU admission. Seven patients (2%) were readmitted to the ICU during their hospitalization, all of whom survived to hospital discharge.
TABLE 3. -
Predictors of Mortality
|Adjusted OR (95% CI)
||Adjusted OR (95% CI)
|Persons of color
|Age ≥ 65 yr old
|Chronic kidney injury
|Congestive heart failure
|Coronary artery disease
|Chronic obstructive pulmonary disease
|Outpatient systemic steroids
OR = odds ratio, NA = not applicable in final model.
Detroit’s racial/ethnic profile includes 79% African-American, 10% Caucasian, 8% Hispanic/Latino, and 2% Asian (17). In our study, approximately 63% of patients were POC. This is consistent with the community representation of Detroit but does differ significantly from U.S. data (18).
Our study sought to identify whether race led to important differences in outcomes. We found a 28-day mortality benefit for POC versus Whites; however, there were no differences in hospital mortality, ICU LOS, or need for mechanical ventilation in a cohort of COVID-19-positive patients admitted to an ICU. We observed a shortened survival among White patients (after adjusting for confounders) with COVID-19 admitted to the ICU. Importantly, our study was limited to an ICU cohort; as such, we did not evaluate whether there were racial differences in becoming clinically symptomatic or critically ill if infected with COVID-19 (19).
The literature is mixed as to whether racial disparities determine clinically significant outcomes in the ICU. In an older study, Williams et al (20) found a shorter ICU LOS for African-American patients, in contrast to our findings. Conversely, Higgins et al (21) showed no racial differences in ICU LOS. Although no differences in ICU LOS existed between the groups in our study, the overall median ICU LOS in this cohort is similar to other studies in patients with ARDS (22,23).
The literature is also inconsistent with regard to the need for mechanical ventilation among different racial groups. Nanchal et al (24) found an increased rate of mechanical ventilation in African-Americans when compared with White patients with asthma. In ARDS patients, there were no differences between the rates of intubation among the different racial/ethnic groups (25).
We found a decreased 28-day mortality in POC compared with Whites. This effect remained significant after controlling for important confounders including age and common comorbidities. These results seemingly conflict with prior literature that showed either no mortality difference (9,20,26) or worsening (27) in survival among African-American patients. One plausible explanation for this finding is that POC deaths were delayed. Of the nine patients who died in the hospital after 28 days, seven were POC. This could explain the contradictory observation of a lack of inhospital mortality difference despite the 28-day mortality benefit. Furthermore, we did observe significantly longer time to death in the POC group.
We note that our cohort of patients is exclusively individuals admitted to an ICU. Racial disparities in the community may be differentiated by unique social and economic factors that may not be crucial after a patient is admitted to a protocol-driven ICU. Although the protocolized nature of care in the ICU may help mitigate some of these social and economic factors, cultural influences can still significantly contribute to specific clinical outcomes. The longer survival time observed in POC might be explained by POC requesting more critical care resources and life support than Whites (28). Furthermore, POC are less likely to withhold or withdraw care than Whites (29). Finally, POC tend to use palliative care resources less frequently than Whites (30,31).
The overall mortality in our entire cohort was nearly 50%. In patients admitted to an ICU with COVID-19, reported mortality ranges from 8% to 67% (32). Although many of the individual studies had small numbers of patients, the combined mortality was 25%. One explanation for our high mortality rate may be the high proportion of mechanical ventilation and ARDS in our cohort. Almost three-quarters of our patients required mechanical ventilation. Similarly, 70% of our patients had ARDS with approximately 55% of those having severe disease, with a Pao2/Fio2 ratio of less than 100. As such, our cohort represents a very sick population of patients with a novel disease. In addition, our mortality rate is similar to other reported cohorts of patients with ARDS (33).
Despite recent evidence that inhospital use of systemic steroids improves survival in patients with COVID-19-associated ARDS (34–36), we did not find steroids to be a significant predictor of mortality. This was likely because more than three-fourths of patients analyzed in our study received steroids due the early initiation of protocolized care. Additionally, steroid use was less likely to be a confounder in our study, because while Whites received slightly more steroids than POC, the observed survival and mortality advantage favored POC.
There are a few important limitations in our study. First, this is a single-center study performed in a large, urban, tertiary-care hospital. We do not know if these results are applicable in other settings. Detroit is a predominantly African-American community, and our patient population reflects this finding. One of the advantages of our study being a single-center study is that we provide identical resources and protocols for all patients regardless of race.
Our results could conceivably be biased if there was a difference in time to presentation between the groups. Indeed, prior studies in African-American patients with acute coronary syndromes have shown this pattern (37). However, our data show equivalent time to presentation for both groups to our institution.
Another potential limitation is that the definition of POC and White can be subjective. Our data were self-reported, which can often be inaccurate and might skew data (38). However, this is a very common way to document and collect racial/ethnic data (39). In addition, we identified Arabs, Hispanics/Latinos, and Asians as POC. Although this might not be a uniform way of presenting race (17,40), it is acceptable based on a lack of definitive ethnicity norms in the United States (41–43).
Racial disparities are commonplace and an important issue in providing medical care. The literature supports the contention that African-Americans are infected and die of COVID-19 at a rate higher than their population percentage would dictate (13,14). This requires continued investigation as the COVID-19 pandemic continues. Our study investigated whether racial differences affected outcomes in patients with COVID-19 admitted to an ICU. We describe no differences between the persons of color and White patients inhospital mortality, ICU LOS, or time to mechanical ventilation. Conversely, persons of color had a significantly longer survival time when compared with white patients and a decreased 28-day mortality. These results, while intriguing, need to be confirmed beyond this single-center experience. Nonetheless, our findings are encouraging that persons of color might receive similar care to White patients after admission to an ICU.
In critically ill patients infected with COVID-19, POC had a lower 28-day mortality than Whites with no difference in hospital mortality, ICU LOS, or rates of intubation. These findings are contrary to previously held beliefs surrounding the pandemic.
We thank Mohamad Raad, MD, and Mohammed Ferras Dabbagh, MD, for their assistance with data collection.
1. Centers of Disease Control and Prevention. Previous U.S. COVID-19 Case Data. Available at: https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/previouscases.html
. Accessed September 1, 2020
3. Grasselli G, Zangrillo A, Zanella A, et al.; COVID-19 Lombardy ICU Network. Baseline characteristics and outcomes of 1591 patients infected with SARS-CoV-2 admitted to ICUs of the Lombardy Region, Italy. JAMA. 2020; 323:1574–1581
4. Guan WJ, Ni ZY, Hu Y, et al.; China Medical Treatment Expert Group for Covid-19. Clinical characteristics of coronavirus disease 2019 in China. N Engl J Med. 2020; 382:1708–1720
5. Richardson S, Hirsch JS, Narasimhan M, et al.; the Northwell COVID-19 Research Consortium. Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in the New York City Area. JAMA. 2020; 323:2052–2059
6. Arentz M, Yim E, Klaff L, et al. Characteristics and outcomes of 21 critically ill patients with COVID-19 in Washington State. JAMA. 2020; 323:1612–1614
7. Barnato AE, Alexander SL, Linde-Zwirble WT, et al. Racial variation in the incidence, care, and outcomes of severe sepsis: Analysis of population, patient, and hospital characteristics. Am J Respir Crit Care Med. 2008; 177:279–284
8. Danziger J, Ángel Armengol de la Hoz M, Li W, et al. Temporal trends in critical care outcomes in U.S. minority-serving hospitals. Am J Respir Crit Care Med. 2020; 201:681–687
9. Erickson SE, Vasilevskis EE, Kuzniewicz MW, et al. The effect of race
and ethnicity on outcomes among patients in the intensive care unit: A comprehensive study involving socioeconomic status and resuscitation preferences. Crit Care Med. 2011; 39:429–435
10. Galea S, Blaney S, Nandi A, et al. Explaining racial disparities in incidence of and survival from out-of-hospital cardiac arrest. Am J Epidemiol. 2007; 166:534–543
11. Plurad DS, Lustenberger T, Kilday P, et al. The association of race
and survival from sepsis after injury. Am Surg. 2010; 76:43–47
12. Soto GJ, Martin GS, Gong MN. Healthcare disparities in critical illness. Crit Care Med. 2013; 41:2784–2793
13. Yancy CW. COVID-19 and African Americans. JAMA. 2020; 323:1891–1892
14. Chowkwanyun M, Reed AL. Racial health disparities and Covid-19 — caution and context. New Engl J Med. 2020; 383:201–203
15. Henry Ford Health System 2020 COVID-19 internal data.
16. The ADTF. Acute respiratory distress syndrome: The Berlin definition. JAMA. 2012; 307:2526–2533
17. United States Census.gov Quick Facts. Census Data Detroit Michigan. Available at: https://www.census.gov/quickfacts/detroitcitymichigan
. Accessed September 1, 2020
18. 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; 69:458–464
19. Liang W, Liang H, Ou L, et al. Development and validation of a clinical risk score to predict the occurrence of critical illness in hospitalized patients with COVID-19. JAMA Intern Med. 2020; 180:1–9
20. Williams JF, Zimmerman JE, Wagner DP, et al. African-American and White patients admitted to the intensive care unit: Is there a difference in therapy and outcome? Crit Care Med. 1995; 23:626–636
21. Higgins TL, McGee WT, Steingrub JS, et al. Early indicators of prolonged intensive care unit stay: Impact of illness severity, physician staffing, and pre-intensive care unit length of stay. Crit Care Med. 2003; 31:45–51
22. Sheu CC, Gong MN, Zhai R, et al. Clinical characteristics and outcomes of sepsis-related vs non-sepsis-related ARDS. Chest. 2010; 138:559–567
23. Talmor D, Sarge T, Malhotra A, et al. Mechanical ventilation guided by esophageal pressure in acute lung injury. N Engl J Med. 2008; 359:2095–2104
24. Nanchal R, Kumar G, Majumdar T, et al. Utilization of mechanical ventilation for asthma exacerbations: Analysis of a national database. Respir Care. 2014; 59:644–653
25. Kangelaris KN, Ware LB, Wang CY, et al. Timing of intubation and clinical outcomes in adults with acute respiratory distress syndrome. Crit Care Med. 2016; 44:120–129
26. Krishnan V, Diette GB, Rand CS, et al. Mortality in patients hospitalized for asthma exacerbations in the United States. Am J Respir Crit Care Med. 2006; 174:633–638
27. Horner RD, Lawler FH, Hainer BL. Relationship between patient race
and survival following admission to intensive care among patients of primary care physicians. Health Serv Res. 1991; 26:531–542
28. Kwak J, Haley WE. Current research findings on end-of-life decision making among racially or ethnically diverse groups. Gerontologist. 2005; 45:634–641
29. Orlovic M, Smith K, Mossialos E. Racial and ethnic differences in end-of-life care in the United States: Evidence from the Health and Retirement Study (HRS). SSM Popul Health. 2019; 7:100331
30. Johnson KS. Racial and ethnic disparities in palliative care. J Palliat Med. 2013; 16:1329–1334
31. National Hospice and Palliative Care Organization. NHPCO’s Facts & Figures on Hospice Care in America. 2018. Available at: https://www.nhpco.org/research/
. Accessed July 28, 2020
32. Quah P, Li A, Phua J. Mortality rates of patients with COVID-19 in the intensive care unit: A systematic review of the emerging literature. Crit Care. 2020; 24:285
33. Fan E, Brodie D, Slutsky AS. Acute respiratory distress syndrome: Advances in diagnosis and treatment. JAMA. 2018; 319:698–710
34. Fadel R, Morrison Austin R, Vahia A, et al. Early short course corticosteroids in hospitalized patients with COVID-19. Clin Infect Dis. 2020; 71:2114–2120
35. Group RC, Horby P, Lim WS, et al. Dexamethasone in hospitalized patients with Covid-19 - preliminary report. N Engl J Med. 2020 Jul 17. [online ahead of print]
36. Tomazini BM, Maia IS, Cavalcanti AB, et al. Effect of dexamethasone on days alive and ventilator-free in patients with moderate or severe acute respiratory distress syndrome and COVID-19: The CoDEX randomized clinical trial. JAMA. 2020; 324:1307–1316
37. DeVon HA, Burke LA, Nelson H, et al. Disparities in patients presenting to the emergency department with potential acute coronary syndrome: It matters if you are Black or White. Heart Lung. 2014; 43:270–277
38. Klinger EV, Carlini SV, Gonzalez I, et al. Accuracy of race
, ethnicity, and language preference in an electronic health record. J Gen Intern Med. 2015; 30:719–723
39. Council NR. Eliminating Health Disparities: Measurement and Data Needs. 2004, Washington, DC, The National Academies Press,
40. United States Census.gov Quick Facts. Available at: https://www.census.gov/quickfacts/fact/table/US/PST045219
. Accessed September 1, 2020
41. Kayyali R. US census classifications and Arab Americans: contestations and definitions of identity markers. J Ethnic Migration Stud. 2013; 39:1299–1318
42. Allen VC Jr, Lachance C, Rios-Ellis B, et al. Issues in the assessment of “race
” among Latinos: Implications for research and policy. Hisp J Behav Sci. 2011; 33:411–424
43. Morning A. The racial self-identification of South Asians in the United States. J Ethnic Migration Stud. 2001; 27:61–79