Geographic distribution of bed occupancy during the COVID‐19 epidemic in the United States: A nationwide study

Full text

RESEARCH LETTER
Geographic distribution of bed occupancy during
the COVID-19 epidemic in the United States: A
nationwide study
Kate E. Trout | Li-Wu Chen
Department of Health Sciences, School of Health Professions, University of Missouri-Columbia, Columbia, Missouri
Correspondence
Kate E. Trout, Department of Health Sciences, School of Health Professions, University of Missouri-Columbia, 510 Lewis Hall, Columbia, MO 65211, USA.
Email: kate.trout@health.missouri.edu
Funding information
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
1|INTRODUCTION
Since the SARS-CoV-2 pandemic started, the United States has
become the global leader in confirmed cases. As of 10 Decem-
ber 2020, the United States had 15 271 571 confirm coronavi-
rus disease 2019 (COVID-19) infections and 288 762 deaths.
The epidemic threatens overwhelming our healthcare system,
especially in areas and communities that have been well-
documented to have limited access to care. Several models
have aimed to predict the burden on the healthcare system in
the early stages of the pandemic,
1,2
but do not adjust for
many important population characteristics that impact access
to care, geographic variability, or the ability of the local pro-
vider infrastructure to deliver care (eg, bed capacity). To the
best of our knowledge, no papers have been published using
longitudinal, retrospective data describing bed occupancy or
geographic variability during the COVID-19 pandemic in the
United States. As a nation, we do not have a good under-
standing of how the direct and indirect consequences of the
COVID-19 epidemic have impacted access to healthcare in var-
ied geographic regions or for underserved populations. The
objective of this paper is to provide preliminary information
describing the geographic disparities of bed occupancy across
the United States during the early pandemic. In addition, we
will identify states that are experiencing ongoing capacity
issues.
2|METHODS
2.1 |Data and sample
This longitudinal, observational study used data from the Department of
Health and Human Services for COVID-19 estimated patient impact and
hospital capacity by state from 1 April to 31 October 2020.
3-6
The daily
time series datasets are estimations of inpatient, COVID-19 inpatient,
and intensive care unit (ICU) beds occupied using facility-level granularity
from: (a) HHS TeleTracking and (b) healthcare facilities reporting to HHS
Protect by state/territorial health departments. Due to data availability
and access to the datasets, there are gaps in the data from July 8-12 and
August 14-30. The final dataset included 50 states and the District of
Columbia (DC) for a total of 10 067 observations. We merged the data
with the United States COVID-19 cases and deaths by state.
7
2.2 |Bed capacity mapping
We mapped each state based on their highest daily occupancy over
the time period to determine the geographic maldistribution on their
level of risk of reaching full bed capacity. States that reached greater
than 70% occupancy for either inpatient OR ICU beds in the time
period were considered “elevated risk,”greater than 70% occupancy
for both were “high risk,”and >99% occupancy for inpatient or ICU
beds were “at or above capacity.”Reasons for using a 70% threshold
Received: 4 February 2021 Revised: 12 May 2021 Accepted: 23 May 2021
DOI: 10.1002/hsr2.315
&C?JRFѥ1AGCLACѥ0CNMPRQ
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any
medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
© 2021 The Authors. Health Science Reports published by Wiley Periodicals LLC.
Health Sci Rep. 2021;4:e315. wileyonlinelibrary.com/journal/hsr2 1of5
https://doi.org/10.1002/hsr2.315

were 3-fold: (a) the highest US national occupancy for the time period
was near 70%; (b) similar 70% thresholds are used to describe states
with the highest occupancy
8
; and (c) our data indicate fluctuations of
30% occupancy are possible as the epidemic progresses. States that
have predominantly older adults are at greater risk of hospitalization
from COVID-19 and account for 53% of ICU admissions.
9
Our map
includes information identifying the top 10 states with the highest
populations of older adults living in rural counties that do not have ICU
beds.
10
2.3 |Statistical analysis
Descriptive statistics were produced for every state over the time period
to determine bed occupancy. We estimated the impact of new cases on
inpatient and ICU occupancy by conducting a bivariate regression analy-
sis. To improve data quality, we included data from May 01 to October
30 to avoid the correction errors in new cases at the beginning of the
reporting period. Our primary dependent variables were the number of
inpatient beds occupied by COVID-19 patients and ICU beds occupied,
and the independent variable was the number of new COVID-19 cases.
We used a fixed-effects generalized least squares regression with robust
option to obtain heteroskedasticity-robust standard errors by the state
to determine the impact of COVID-19 incidence on inpatient and ICU
occupancy. Model selection was verified by conducting Hausman test
where the unique errors (ui) were correlated with the regressors
(P< .001).
11
In addition, we performed bivariate linear regressions to
analyze the demand of non-COVID-19 inpatient beds over time. We
considered P< .05 to be significant and report 95% confidence intervals
(CIs) for coefficients. All analyses were computed using StataMP
version 16.
3|RESULTS
3.1 |States' greatest occupancy experienced
during the epidemic and risk of reaching hospital bed
capacity
From 1 April to 31 October 2020 the United States average for inpatient
beds occupied by COVID-19 patients was 7.3%, 62.9% for total inpatient
beds occupied, and 64% for ICU beds occupied. The states with the
highest daily inpatient beds occupied were Rhode Island (102.2%),
followed by Hawaii (99.3%) and Washington (91.9%). The DC (100%),
Texas (99.0%), Rhode Island (94.1%),andWashington(93.8%)hadthe
highest daily ICU bed occupancy. Out of the 50 states and DC, 72.5%
(n =37/51) had reached a maximum inpatient occupancy of 70% or
greater during the time period, and 90.2% (n =46/51) for ICU occu-
pancy, respectively (Figure 1). There are geographic clusters of states fac-
ing a high risk of reaching capacity in the external regions of the United
States. Mississippi, Missouri, Wyoming, and Vermont are identified at
elevated risk of reaching capacity but are also states with a high popula-
tion of older adults living in rural counties without ICU beds.
9
West
Virginia, Maine, Arkansas, South Dakota, North Dakota, and Montana
are also identified as the top 10 states with the highest population of
older adults living in rural counties without ICU beds,
9
and are at high
risk of reaching capacity for both inpatient beds and ICU beds.
Using nationwide estimates, for every 100 new COVID-19 cases
we can expect 56 patients (95% CI 43.9-69.1) will be admitted to an
inpatient bed (P< .001; R
2
=0.629). For every 100 new COVID-19
cases we can expect 8 patients (95% CI 3.5-13.8) will be admitted to
an ICU (P=.001; R
2
=0.448). At the state level, there is a
maldistribution of COVID-19 inpatient occupancy. The highest daily
COVID-19 inpatient occupancy was 46.2% (NJ), followed by 44.8%
(NY) and 41.2% (AZ), and were clustered in the south and northeast
regions. Texas (99.0%) and the DC (100%) were at capacity for ICU
beds during the time period and were the top 10 states for COVID-19
inpatient occupancy. Rhode Island (102.2%) reached capacity for inpa-
tient beds and was a top 10 state for COVID-19 inpatient occupancy.
Arkansas was one of the top 10 states for COVID-19 inpatient occu-
pancy, face large populations of older adults living in rural counties
without ICU beds, and have greater than 70% occupancy for inpatient
and ICU beds.
3.2 |States with the highest COVID-19 occupancy
have a growing demand for non-COVID-19 admissions
Arizona, Georgia, Massachusetts, Maryland, New Jersey, and
New York had the highest daily inpatient beds occupied by COVID-
19 patients. The northern states including Massachusetts, Maryland,
New Jersey, and New York experienced peaks in COVID-19 inpatient
admissions early months of the epidemic from April to June. In con-
trast, the southern states including Arizona and Georgia experienced
peaks in COVID-19 inpatient admissions toward the middle of the
time period between June and September. Temporal trends in
COVID-19 inpatient beds occupied declined overall among these
states, but the same declines were not seen in the percent inpatient
beds occupied. Figure 2 shows the temporal change in occupancy for
inpatient beds occupied by COVID-19 patients, overall inpatient occu-
pancy, and ICU occupancy.
The nation is experiencing increased demand as the epidemic con-
tinues. Every day there was a 0.09% (95% CI 0.08-0.10) national increase
in non-COVID-19 inpatient occupancy (P< .001; R
2
=0.78). In addition,
the bottom frame of Figure 2 demonstrates that non-COVID-19 related
inpatient occupancy increased over the course of the epidemic for all
6 states experiencing a high hospital burden of COVID-19 patients. We
found a significant positive linear relationship between non-COVID-19
inpatient occupancy and time for all 6 states (P< .001). Five of the six
states had higher projected daily increases in non-COVID-19 admissions
than the national estimate, including New Jersey (0.20%), New York
(0.17%), Maryland (0.13%), Massachusetts (0.12%), and Georgia (0.11%).
Arizona experienced 0.05% increase over the time period, which was
below the national average.
Alabama, DC, Georgia, Maryland, Rhode Island, and South Caro-
lina were the states with the highest occupancy over the time period.
2of5 TROUT AND CHEN
&C?JRFѥ1AGCLACѥ0CNMPRQ

Over the time period, the non-COVID-19 inpatient occupancy had a
significant positive linear relationship for all 6 states (P< .001). Rhode
Island (0.20%), Maryland (0.13%), DC (0.16%), Georgia (0.11%), and
South Carolina (0.11%) experienced increases higher than the national
estimate of 0.09% per day in percent inpatient occupancy by non-
COVID-19 patients.
4|DISCUSSION
Our results show that hospital inpatient and ICU bed occupancy is a
major concern across the United States. From 1 April to 31 October
2020, 72.5% of states and territories had reached a maximum daily
inpatient occupancy of 70% or greater, and 90.2% of them had
reached a maximum ICU occupancy of 70% or greater. Four of the
states/territories reached at or above capacity during the time period
for ICU or inpatient beds. States with the highest sustained occupancy
for inpatient and ICU beds were Alabama, DC, Georgia, Maryland,
Rhode Island, and South Carolina. Georgia and Maryland were not
only among the states with the highest sustained inpatient and ICU
occupancy but also had the highest COVID-19 inpatient occupancy.
Immediate contingency planning and resource allocation are needed
for states reaching capacity for either ICU or inpatient beds, which will
have negative consequences on health outcomes and the healthcare
system.
Based on this national data, 56 patients will be admitted inpatient
and 8 patients will be admitted into the ICU for every 100 new cases
of COVID-19; though these estimates were not able to be adjusted
by important population characteristics, it may inform the overall bur-
den on the healthcare system. Despite inpatient occupancy for
COVID-19 declined over the time period, trends in overall occupancy
did not, indicating there may be indirect consequences of the pan-
demic on population health outcomes from delayed access to care.
Non-COVID-19 inpatient occupancy has increased nationally by
0.09% per day over the course of the epidemic (P< .001). States
experiencing the highest COVID-19 inpatient burden also are
experiencing higher rates of non-COVID-19 inpatient admissions than
the national average. New Jersey and Rhode Island had the highest
daily increases in non-COVID-19 admissions. Local and state
healthcare systems should monitor admissions to see if they are
experiencing a “double burden”on bed capacity, such as in New Jer-
sey. New Jersey is one of the highest states for COVID-19 inpatient
admissions and had the highest national increase in non-COVID-19
inpatient admissions (0.20% daily increase). In addition, states with
FIGURE 1 States' risk of reaching capacity and top 10 states with the highest daily burden for COVID-19 inpatient occupancy from 1 April to
31 October 2020
TROUT AND CHEN 3of5
&C?JRFѥ1AGCLACѥ0CNMPRQ

the highest inpatient COVID-19 admissions were not always the
states with the highest bed occupancy overall. Rhode Island was not
one of the top states with the highest COVID-19 inpatient admissions
but did experience the highest daily increase for non-COVID-19
admissions in the nation with ongoing capacity issues. More research
is needed to determine the reasons for admissions that are not due to
COVID-19, and the added burden that these admissions will have on
the healthcare system.
Research needs to determine the epidemic's impact on smaller
geographic areas and those from rural communities, older adults,
minorities, immigrants, low income, uninsured, and those located in
federally designated health provider shortage areas and medically
underserved populations/areas.
1,10,12
Many rural areas in the United
States have little to no access to acute or critical care. States that have
the highest population of older adults (60 years and older) that live in
rural counties with no ICU bed should have special consideration,
10
especially since these states are already experiencing a high and ele-
vated risk of reaching capacity. Special attention should be given to
Arkansas because it was one of the top 10 states for COVID-19 inpa-
tient occupancy, face large populations of older adults living in rural
counties without ICU beds, and reached greater than 70% occupancy
for both inpatient and ICU beds.
FIGURE 2 States with the
highest inpatient bed occupancy
from COVID-19 patients by state
from 1 April to 31 October 2020
4of5 TROUT AND CHEN
&C?JRFѥ1AGCLACѥ0CNMPRQ

4.1 |Limitations
The aggregated state-level estimates limit our ability to adjust for
important population characteristics in our model or identify smaller
geographic areas where there are COVID-19 hotspots or that experi-
ence limited bed capacity. This study provides preliminary evidence of
the state-level bed occupancy during the epidemic. Therein, states
that were not identified at risk of reaching capacity may also
experience capacity issues within smaller geographic areas.
5|CONCLUSION
Even though inpatient occupancy for COVID-19 declined over the
time period, trends overall for inpatient and ICU occupancy did not.
Careful attention needs to be given to admissions that are non-
COVID-19 related as the pandemic progresses. Further overburdening
already overwhelmed healthcare facilities will likely have a negative
impact on the quality of care, access to care, and health outcomes, in
the short- and long-terms. As a consequence, we may expect to see
increases in emergency room admissions and preventative hospitaliza-
tion. The long-term impact of the overburdened healthcare system is
yet to be determined but could be related to the delay of care for
chronic diseases.
ETHICS STATEMENT
This research has been deemed not human subject research by
U.S. Department of Health and Human Services, and uses publicly
available aggregated statistics at the state level.
CONFLICT OF INTEREST
The authors have no potential conflicts of interest to declare regard-
ing the work involved in this article.
AUTHOR CONTRIBUTIONS
Conceptualization: Kate E. Trout, Li-Wu Chen
Data Curation: Kate E. Trout
Formal Analysis: Kate E. Trout
Methodology: Kate E. Trout, Li-Wu Chen
Visualization: Kate E. Trout, Li-Wu Chen
Writing –Original Draft Preparation, Review, & Editing: Kate E. Trout,
Li-Wu Chen
All authors have read and approved the final version of the
manuscript.
Dr. Kate E. Trout had full access to all of the data in this study
and takes complete responsibility for the integrity of the data and the
accuracy of the data analysis.
TRANSPARENCY STATEMENT
Dr. Kate E. Trout affirms that this manuscript is an honest, accurate,
and transparent account of the study being reported; that no impor-
tant aspects of the study have been omitted; and that any discrepan-
cies from the study as planned have been explained.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are openly available in
HealthData.gov for hospital ICU and inpatient beds occupied by
the state.
ORCID
Kate E. Trout https://orcid.org/0000-0002-6633-7349
REFERENCES
1. Miller IF, Becker AD, Grenfell BT, Metcalf CJE. Disease and
healthcare burden of COVID-19 in the United States. Nat Med. 2020;
26(8):1212-1217.
2. Bartsch SM, Ferguson MC, McKinnell JA, et al. The potential health
care costs and resource use associated With COVID-19 in the United
States: a simulation estimate of the direct medical costs and health
care resource use associated with COVID-19 infections in the United
States. Health Aff. 2020;39(6):927-935.
3. HHS Protect. Department of Health and Human Services (DHHS).
Estimated Inpatient Beds Occupied by State Timeseries [Data file].
HHS Protect, DHHS: Washington, DC; 2020. https://healthdata.gov/
dataset/covid-19-estimated-patient-impact-and-hospital-capacity-state.
4. HHS Protect. Department of Health and Human Services (DHHS).
Estimated Inpatient Beds Occupied by COVID-19 Patients by State
Timeseries [Data file]. HHS Protect, DHHS: Washington, DC; 2020.
https://healthdata.gov/dataset/COVID-19-Estimated-Inpatient-
Beds-Occupied-by-COVI/py8k-j5rq.
5. HHS Protect. Department of Health and Human Services (DHHS). Esti-
mated ICU Beds Occupied by State Timeseries[Data file]. HHS Protect,
DHHS: Washington, DC; 2020. https://healthdata.gov/dataset/COVID-
19-Estimated-ICU-Beds-Occupied-by-State-Time/7ctx-gtb7.
6. Center of Disease Control (CDC). National Healthcare Safety Net-
work (NHSN). National and state estimates NHSN COVID-19 Module
data [data file]. https://www.cdc.gov/nhsn/covid19/report-patient-
impact.html.
7. Center of Disease Control (CDC). United States COVID-19 Cases and
Deaths by State over Time [Data file]. Case Task Force, Center of Dis-
ease and Control: Atlanta, GA; 2020. https://data.cdc.gov/d/9mfq-cb36.
8. Center for Disease Control National Healthcare Safety Network
(CDC NHSN). Current Hospital Capacity Estimates–Snapshot.
Updated July 16, 2020. https://www.cdc.gov/nhsn/covid19/report-
patient-impact.html#anchor_1587406850.
9. CDC COVID-19 Response Team. Severe outcomes among patients
with coronavirus disease 2019 (COVID-19) - United States, February
12–March 16, 2020. MMWR Morb Mortal Wkly Rep. 2020;69(12):
343-346.
10. Davoodi NM, Healy M, Goldberg EM. Rural America's hospitals are
not prepared to protect older adults from a surge in COVID-19 cases.
Gerontol Geriat Med. 2020;6:2333721420936168.
11. Greene WH. Econometric Analysis. 6th ed. Upper Saddle River, NJ:
Prentice Hall; 2008.
12. Lakhani HV, Pillai SS, Zehra M, Sharma I, Sodhi K. Systematic review
of clinical insights into novel coronavirus (CoVID-19) pandemic:
persisting challenges in U.S. rural population. Int J Environ Res Public
Health. 2020;17(12):4279.
How to cite this article: Trout KE, Chen L-W. Geographic
distribution of bed occupancy during the COVID-19 epidemic
in the United States: A nationwide study. Health Sci Rep. 2021;
4:e315. https://doi.org/10.1002/hsr2.315
TROUT AND CHEN 5of5
&C?JRFѥ1AGCLACѥ0CNMPRQ