Comparing COVID-19 Critical Care Admissions by Minority Populations to Pre-COVID-19 Critical Care Utilization

Background: Several months into the COVID-19 pandemic, reassessing intensive care unit (ICU) utilization, specically with regional impact on diverse populations, should be a priority for hospitals planning for critical care resource allocation. In our study, we reviewed the impact of COVID-19 on a community hospital serving an urban region, comparing the sociodemographic distribution of ICU admissions before and during the pandemic. Methods: We executed a time sensitive analysis to see if COVID-19 ICU admissions reect regional sociodemographic populations as well as ICU admission trends prior to the current pandemic. Collected sociodemographic variables included sex, race, ethnicity, and age of adult patients (age 18 and older) admitted to the hospital’s medical and cardiac ICUs, which were converted to COVID-19 ICUs. The time period selected was 18-months, which was then dichotomized into pre-COVID-19 admissions (December 1, 2018 to March 13, 2020) and COVID-19 ICU admissions (March 14, 2020 to May 31, 2020). Variables were compared using Fisher’s exact tests and Wilcoxon tests when appropriate. Results: During the 18-month period, 1861 patients were admitted to the aforementioned ICUs. The mean age of the 1861 patients was 62.75 + 15.57 years old, with the majority of these patients being male (52.23%), White (64.43%), and non-Hispanic/Latinx (95.75%). There were differences in racial and ethnic distribution comparing pre-COVID-19 admissions to the COVID-19 admissions. Compared to pre-COVID-19 ICU admissions, there was an increase in African American versus White admissions (p=0.01) and an increase in Hispanic/Latinx versus non-Hispanic/Latinx admissions (p<0.01), during the COVID-19 pandemic. Discussion: During the rst three months of admissions to ICUs, there was a rise in admissions among Hispanic/Latinx and African-American patients, while non-Hispanic/Latinx and White patient admissions declined to the pre-COVID ndings support development of strategies enhance ICU with regards to socio-demographics. An increase in minority populations based on race and ethnicity was evident in COVID-19 ICUs, while non-minority populations saw a decrease in ICU admissions. While much has been said regarding the disproportionate impact in minorities due to COVID-19, we have shown, using time-lapse evaluations, that this impact is signicantly different than pre-pandemic ICU admission trends. Therefore, as strategies are discussed to curb the incidence of COVID-19 in minority populations, monitoring if such a reduction results in a decrease in ICU admissions should be emphasized as a public health priority.

19 ICU admissions, there was an increase in African American versus White admissions (p=0.01) and an increase in Hispanic/Latinx versus non-Hispanic/Latinx admissions (p<0.01), during the COVID-19 pandemic.
Discussion: During the rst three months of admissions to COVID-19 ICUs, there was a rise in admissions among Hispanic/Latinx and African-American patients, while non-Hispanic/Latinx and White patient admissions declined compared to the previous pre-COVID year. These ndings support development of strategies to enhance allocation of resources to bolster novel, equitable strategies to mitigate the incidence of COVID-19 in minority populations.

Background
Optimizing the utilization of intensive care units (ICUs) is a signi cant healthcare priority in the United States. 1 This optimization requires a critical understanding of variables that contribute to regional usage of adult ICUs. For instance, geographic and sociodemographic factors (e.g. prevalence of minority race and ethnicity, poverty, educational status) 2,3 and the presence of certain medical conditions 4,5 are associated with adult ICU utilization and admissions 5,6 . When assessing common diagnoses present on admission to adult medical ICUs, such as sepsis and acute respiratory distress syndrome (ARDS), these diagnoses cluster in regions with high prevalence of certain medical conditions (e.g. hypertension and diabetes) 7 and with socioeconomic factors (e.g. race, ethnicity, poverty) 5,8 . Insight into the aforementioned variables would ensure e cient use of critical care resources for hospitals, communities, and patients.
The current pandemic caused by the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) has created unprecedented demands on acute health care, necessitating the development of triage algorithms typically assigned to battle elds to determine who receives medical treatment or access to medical equipment, and in which order. 9 Speci cally, SARS-CoV-2 resulting coronavirus infectious disease 2019  frequently results in sepsis and ARDS 10,11 , which warrant ICU admissions. Having reached a point several months into the pandemic, reassessing ICU utilization, speci cally regional impact, should be a priority for hospitals in an effort to assist in ongoing public health policies to mitigate the spread of the virus as well as planning for critical care resource allocation.
In our study, we reviewed the impact of COVID-19 on a community hospital serving an urban region. Speci cally, we executed a time sensitive analysis to determine if COVID-19 ICU admissions re ect regional sociodemographic populations as well as ICU admission trends prior to the current pandemic.

Patient Population
We obtained 18-months of admissions (December 2018 to May 2020) to the medical intensive care unit (MICU) and the cardiac intensive care unit (CICU) at Johns Hopkins Bayview Medical Center (JHBMC). JHBMC is one of three academic hospitals of the Johns Hopkins Health System, located in Baltimore City and has a potential capacity of 500 beds. The reason for choosing MICU and CICU is that both were converted to specialized COVID-19 intensive care units (ICU) to provide care for critically-ill patients con rmed to have COVID-19 infection and persons under investigation (PUI). We reviewed all adult patients (18 years

Time Period
An 18-month review of data for this investigation provides an entire scal year of ICU admissions, allowing secular trends to be identi ed in how speci c populations utilized critical care resources over a speci c time. In addition, by capturing the rst three months of 2020, we could assure that trends observed between March to May 2020 were a) not speci c to the current year and b) were not typical seasonal trends (as our data captures March to May 2019).
The 18-month data were then dichotomized into pre-COVID-19 admissions (December 1, 2018 to March 13, 2020) and COVID-19 ICU admissions (March 14, 2020 to May 31, 2020). March 14, 2020 was chosen because it was the rst reported case of COVID-19 in Baltimore City 13 and this timeframe correlated with an increase in COVID-19 admissions in our hospital.

Variables of Interest
We collected individual sociodemographic data including age, sex, race, ethnicity, and preferred language spoken (identi ed as "English not preferred") of all patients admitted to the MICU and CICU during the aforementioned timeframe. We de ned elderly as age 65 years and older. We also documented length of stay in the ICU as well as ICU-mortality.

Statistical Analysis
All continuous variables are presented as mean ± standard deviation or median (interquartile range). Categorical variables were summarized as counts and percentages. Variables were compared using Fisher's exact tests and Wilcoxon tests when appropriate. For comparisons of more than two groups, an analysis of variance (ANOVA) was performed. Conditional density plots were utilized to display population admissions over the 18-month period. Statistical analyses were conducted with R software (V.0.99.903).

Results
Over the course of 18 months, the MICU and CICU saw a total of 1861 admissions. The mean age of the 1861 patients was 62.75 ± 15.57 years (range 19 to 100 years). The majority of these patients were male (52.23%), White (64.43%), and non-Hispanic/Latinx (95.75%). Of the 1861 patients, 419 (22.51%) of them did not survive their hospitalization. A complete list of sociodemographic and hospital variables is provided in Table 1.

Discussion
During the rst three months of admissions to a COVID-19 ICU at an urban, academic, community hospital, there was a rise in admissions among Hispanic/Latinx and African-American patients, while non-Hispanic/Latinx and White patient admissions declined over the recent 3-month period. Compared to non-Hispanic/Latinx patients admitted to the ICU, Hispanic/Latinx admitted to the COVID-19 ICU were younger. As expected given the age differential, Hispanic/Latinx had greater survival rates compared to non-Hispanic/Latinx. African American patients had worse mortality outcomes during the pandemic as compared to non-COVID-19 ICU admissions, especially elderly African American patients. Overall, this shift in both admissions and outcomes based on race and ethnicity was a signi cant change during the COVID-19 pandemic for this community hospital in comparison to the previous year's ICU admissions.
Disparities in critical care diagnoses and admissions have been well documented. 5,6,14 And in the case of the COVID-19 pandemic, disparities continue to exist in the proportion of minorities that are impacted by  In the analysis by Price-Haywood et al, they found a disproportionate impact on African Americans with COVID-19: 76.9% of the patients hospitalized were African American, and 70.6% of the patients who died were African American. 16 The data reviewed spanned the rst 6-weeks of the public health crisis of COVID-19 in Louisiana (March 1 to April 11, 2020). Such a disproportionate impact was also seen in African Americans in our hospital's admissions. However, our data showcases that the pandemic had differential effects, increasing admissions of African American patients while simultaneously decreasing admissions of White patients. This nding is new, as evident by our review of critical care admissions the previous years, which were consistently stable for both African Americans and Whites, as well as the same pattern for Hispanic/Latinx patients versus non-Hispanic/Latinx patients.
Optimization of regional ICU utilization has been a priority during the COVID-19 pandemic. 17  such an effort is warranted as this population has the highest rate of hypertension-related critical care admissions 19 . ICU-level information rea rms the need to promote testing and identify resources to facilitate isolation and quarantine for African Americans 16,20,21 and Hispanic/Latinx 20-22 who are experiencing COVID-19-related health disparities. Such strategies will require partnerships with neighborhood leaders to leverage community-based resources and venues to effect these collaborative changes, as well as structural changes at levels of policy and advocacy.
This study has several limitations. First, we did not evaluate more granular clinical data, such as severity of symptoms on presentation and pre-existing conditions. Second, we did not investigate what COVID-19 related syndromes warranted ICU admissions (e.g. sepsis, ARDS, heart failure). While these limitations are important to address in future studies and are necessary overall to understand the complete clinical presentation of COVID-19, we believe such information was not as relevant for implementing immediate community engagement initiatives towards populations impacted disproportionately by the pandemic. Finally, this information is relevant for our community hospital based in an urban region; it is unclear if similar ndings would occur in community hospitals serving rural regions. This should be investigated in order to understand urban versus rural differences on populations as related to COVID-19 incidence and ICU admissions.
In the rst months of the pandemic, there was a signi cant shift in persons admitted to an urban community hospital's ICU with regards to socio-demographics. An increase in minority populations based on race and ethnicity was evident in COVID-19 ICUs, while non-minority populations saw a decrease in ICU admissions. While much has been said regarding the disproportionate impact in minorities due to COVID-19, we have shown, using time-lapse evaluations, that this impact is signi cantly different than pre-pandemic ICU admission trends. Therefore, as strategies are discussed to curb the incidence of COVID-19 in minority populations, monitoring if such a reduction results in a decrease in ICU admissions should be emphasized as a public health priority. Consent for publication. Not applicable as data is not of an individual person that could be identi ed based on the above ndings.

Abbreviations
Availability of data and material. The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Competing interests. The authors declare that they have no competing interests. Funding. Not applicable as funding was not utilized to perform this analysis and review. Authors' contributions. PG, ESC, and SHG were involved in conceptualization, data analyses, manuscript writing (original draft), and manuscript editing. KRP, SC, SS, and VJ were involved in formal analyses and manuscript editing. TT, FHB, and SS were involved in conceptualization and data analyses. JM was involved in data curation and manuscript editing. Comparion of intensive care unit admissions by minority populations before and after SARS-CoV-2 admissions.

Figure 1
Comparion of intensive care unit admissions by minority populations before and after SARS-CoV-2 admissions.