Introduction

A novel respiratory pathogen, coronavirus SARS-CoV-2, [1] caused an outbreak of respiratory illnesses beginning in Wuhan, China in December 2019 [2] Within weeks, cases had been reported in the United States [3] and Europe,[4] heralding a global pandemic of over 600,000 cases and 28,000 deaths by the end of March 2020[5]. To date, in the United States, there have been over 111 million cases of coronavirus and over 1.2 million reported deaths [6]. Researchers in medicine, public health, and nursing have been reporting the disparities in health care outcomes of minority populations [7]. Some etiologic factors of these disparities that have been identified include race-associated comorbidities, [8-9] destitution/socioeconomic status (SES), [8-12] longstanding distrust of health care providers and health care systems,8 lack of access to health care for treatment of chronic conditions, [9,11] language barriers, [8-10,13] multiple generations living in the same household, [8,9] having jobs that cannot be performed remotely, [9] need to use public transit, [9] low level of education, [10] poor housing conditions, [10] and knowledge deficits about symptoms to report and how to mitigate the spread of COVID-19 [9].

Each of these variables leads to adverse COVID incidence due to the potential for close proximity to others with coronavirus, which leads to increased transmission risk to vulnerable individuals [14]. Non-remote workers potentionally were less able to protect themselves from COVID-19 [15]. Health care disparities have been further intensified due to COVID-19 and its associated sequelae. [8,10,16] Black, indigenous, and people of color (BIPOC) including Hispanic people have reported higher infection rates [16-17] and complications of COVID-19. Data are not consistent regarding infection rate in these racial groups [18] as well as Asian people. According to the Centers for Disease Control and Prevention, [19] the rate of hospitalization for BIPOC patients is 2.3 times higher than White persons. These data are corroborated by others. These latter data compared racial minority populations with majority groups [20-26]. The death rate is up to 1.7 times higher in BIPOC people than non-Hispanic White persons [8,20]. Similarly, Hispanic or Latino persons have a 2.2 times higher hospitalization rate and 1.8 to 3.2 times higher death rate compared to White persons. Data are lower for persons of Asian, non-Hispanic persons with a 0.8 times higher rate of hospitalization and death than White persons [8,16,24]. The risk of death from COVID-19 among Asian populations, however, is similar to that of White populations.

There are inadequate data to discern differences among other racial/ethnic groups.16 Data on differences among various race and ethnic groups (non-Hispanic Black, Asian/Pacific Islander, American Indian/Alaskan Native) are corroborated by others [27]. The mortality rates were similarly lower in non-Hispanic White persons than the other racial groups with an increased risk of death reported in those younger than 65 years of age.27 Data on mortality rates and severity of illness vary among states [28] and are inconsistent [10]. In another study, BIPOC patients accounted for more than 50 percent of in-hospital patients who died from COVID-19. However, Asian persons in this study had more severe cardiopulmonary disease. The authors attributed these latter findings to fear of being attacked based on their race, language barriers, immigration status (which can result in lack of health insurance) and lower SES, the latter of which can impact receipt of high-quality health care [29]. Other data, albeit inconsistent, acknowledge the relationship among these socioeconomic variables, COVID-19 infection rate, and patient outcomes [10,30].

The increased rate of COVID-19 infections and death rate are most evident in rural areas [11]. This is partly attributed to limited testing and vaccination rates. Those in rural areas tend to be older, have inadequate or no health insurance, and have more comorbidities, which can put them at increased risk for poorer outcomes from COVID-19. Several of these comorbidities (e.g., hypertension, diabetes, obesity) are common in BIPOC patients. Persons living in rural areas have limited access to health care because of the limited availability of primary care providers [8]. Data also suggest that those living in rural areas were less likely to adhere to COVID-19 prevention methods such as wearing a mask and social distancing [31]. There is also a higher incidence of ‘vaccine hesitancy’ in persons from rural areas [32]. BIPOC patients living in rural areas have a reported higher infection, hospitalization, and mortality rates when compared to non-hispanic White people [17,31,33-35] and are more vulnerable to COVID-19 infections during surges associated with new variants [7].

Methods

Study Design

A retrospective chart review was performed on 100 patients at least 18 years of age who were admitted to the intensive care unit (ICU) of an academic medical center in the southeast United States and who tested positive for COVID-19 via polymerase chain reaction. All patients admitted to the unit study site were positive for COVID-19; the unit was the first designated COVID-19 intensive care unit of the hospital. The first 100 patients were evaluated for this study. Given a sample size of 100, we were powered at 80% and 5% level of significance to detect moderate effect sizes for all bivariate associations between the measures (r=0.25-0.30 for continuous measures, ω=0.26-0.31 for categorical/dichotomous measures); and powered to detect f2=0.11 to 0.17 for a linear prediction model with up to 5 predictors. For a dichotomous outcome (such as death with an expected rate of 20%), we were also powered to detect odds ratios of 1.95 to 2.32, where an odds ratio of 2.5 is considered to be a moderate effect size.

Study Procedure

This study received exempt status from the Institutional Review Board. Patients were admitted over an 8-month period during a COVID-19 surge. Patients were hospitalized in October 2020 through May 2021. Select patient demographics and study variables (Table 1) were collected from the electronic medical record. Data were recorded as they were reported by the patient or family member (if the patient was unable to self-report). Payor information was collected from a pre-selected group. Medicare and Medicaid are listed as one option (US government sponsored healthcare plan) despite differences in eligibility for individuals and impact of socioeconomic status. Patient race was similarly documented based on patient self-report. There is no specific delineation of types of Black patients (e.g., African, Caribbean immigrant, Black Hispanic) in the hospital’s admission database.

Study Instrument

The Sequential Organ Failure Assessment (SOFA) was developed in 1996 to evaluate patients’ severity of illness based on the extent of organ dysfunction. It is commonly used in the intensive care unit setting. Good validity and reliability have been reported [36,37]. An interclass correlation coefficient of .889 was established among physicians [36].

Determination of whether patients lived in a vulnerable zip code was derived from the OASIS Online Analytical Statistical Information System website, which delineates the demographic clusters in the state of Georgia https://oasis.state.ga.us/gis/demographiccluster/democlusters2011.htm. A description of the address county is provided including where residents rank in terms of the state average income. Based on the patient’s reported primary address, the zip code was entered into the website. These data were used to categorize patients into socioeconomic groups.

Table 1: Descriptive Statistics of Sample.

Data Analysis

All data were reviewed for completeness, missing values, outliers, and possible typographical errors or other unusual or invalid values. Descriptive statistics were run for all demographics. Comparisons tests (t-tests for continuous data and chi-square tests for ordinal and categorical data) were performed for all demographics. Regression analyses (linear or logistic as needed for each outcome) were performed to determine the relationships among the independent and dependent variables. SPSS v.28 (IBM Corp. Released 2021. IBM SPSS Statistics for Windows, Version 28.0. Armonk, NY: IBM Corp.) Statistical analysis software were used for all analyses with 5% level of significance used for all tests.

Results

The one-hundred COVID patient ages ranged from 25 - 99 years with an average age of 62.4 (SD 17.1) years. Slightly more were male (57%), and African American (AA) (60%). The majority had Medicare/Medicaid (60.0%) as their payor with 21% having a health maintenance organization (HMO), 12% having government insurance and 7% having either commercial or private (Blue Cross Blue Shield) insurance. Forty-four percent of the patients lived in a vulnerable zip code. Sequential Organ Failure Assessment (SOFA) scores for the 100 patients ranged from 1 to 14 with a median score of 5. The majority survived (77%) despite 73% having organ failure (Table 1).

There was a significant association between survival and organ failure with none of the patients without organ failure dying (p < .001). However, there were no other significant differences between the patients who survived and those who did not for zip code of primary residence, age, sex, payor, nor race. However, descriptively, the patients who died were more non-Caucasian; a higher proportion of AA and Asian, Hispanic, or unknown (Table 2).

Table 2: Survival-Comparison Characteristics.

Additionally, there were no significant differences between patients who had organ failure versus those who did not for age, sex, race, zip code of primary residence, and payor. However, patients who had organ failure were more likely to have Medicare/Medicaid as their payor (Table 3). The only significant difference was for SOFA scores. That is, patients with organ failure had significantly higher SOFA scores (median 6) compared to patients without organ failure (median 3), (p = .003.) (Table 3). When looking at length of stay, there was no significant effect for age nor SOFA scores and no significant differences between sex, race, payor, and zip code of primary residence categories.

Table 3: Organ Failure-Comparison Characteristics.