Gastroenterology

An emergency department-based intensive care unit is associated with decreased hospital length of stay for upper gastrointestinal bleeding

a b s t r a c t

Introduction: upper gastrointestinal bleeding is associated with substantial morbidity, mortality, and in- tensive care unit (ICU) utilization. Initial risk stratification and disposition from the Emergency Department (ED) can prove challenging due to limited data points during a short period of observation. An ED-based ICU (ED-ICU) may allow more rapid delivery of ICU-level care, though its impact on patients with UGIB is unknown.

Methods: A retrospective observational study was conducted at a tertiary U.S. academic medical center. An ED- ICU (the Emergency Critical Care Center [EC3]) opened in February 2015. Patients presenting to the ED with UGIB undergoing esophagogastroduodenoscopy within 72 h were identified and analyzed. The Pre- and Post- EC3 cohorts included patients from 9/2/2012-2/15/2015 and 2/16/2015-6/30/2019.

Results: We identified 3788 ED visits; 1033 Pre-EC3 and 2755 Post-EC3. Of Pre-EC3 visits, 200 were critically ill and admitted to ICU [Cohort A]. Of Post-EC3 visits, 682 were critically ill and managed in EC3 [Cohort B], whereas 61 were critically ill and admitted directly to ICU without care in EC3 [Cohort C]. The mean interval from ED pre- sentation to ICU level care was shorter in Cohort B than A or C (3.8 vs 6.3 vs 7.7 h, p < 0.05). More patients in Co- hort B received ICU level care within six hours of ED arrival (85.3 vs 52.0 vs 57.4%, p < 0.05). Mean hospital length of stay was shorter in Cohort B than A or C (6.2 vs 7.3 vs 10.0 days, p < 0.05). In the Post-EC3 cohort, fewer patients were admitted to an ICU (9.3 vs 19.4%, p < 0.001). The rate of floor admission with transfer to ICU within 24 h was similar. No differences in absolute or risk-adjusted mortality were observed.

Conclusion: For critically ill ED patients with UGIB, implementation of an ED-ICU was associated with reductions in rate of ICU admission and hospital LOS, with no differences in Safety outcomes.

(C) 2021

  1. Introduction

Acute upper gastrointestinal bleeding (UGIB) is a common medical emergency accounting for approximately 300,000 hospitalizations per year in the United States [1]. Thirty-day mortality is reported as high as 10%, and as high as 50% for hemorrhage from esophageal varices [2,3]. The mainstays of management of patients with acute UGIB include resuscitation with IV fluids and blood products, reversing coagulopathy,

? This work has not been presented at another meeting.

* Corresponding author at: Michigan Medicine, Department of Emergency Medicine, Taubman Center B1354, 1500 E Medical Center Dr, SPC 5303, Ann Arbor, MI 48109- 5305, USA.

E-mail address: [email protected] (N.L. Haas).

administration of high-dose proton-pump inhibitor (PPI) therapy, and urgent Upper endoscopy within 24 h of admission [4]. The exact timing (i.e, emergent vs. urgent) of upper endoscopy is debated [5-8].

From the Emergency Department (ED), initial risk stratification of patients with acute UGIB is highly variable despite guidelines and vali- dated risk scores [9]. Disposition from the ED (admission to an intensive care unit [ICU] vs general ward) of patients with UGIB can prove chal- lenging due to limited data points, Short duration of observation, and difficulty anticipating clinical course. The estimated Hospital length of stay ranges from 2 days to greater than 2 weeks depending on complications and re-bleeding, with the average cost of hospitalization ranging from $3500 – $23,000 [10]. Admission to a non-ICU ward with subsequent transfer to an ICU has been associated with higher mortality and longer length of stay for different patient populations [11-15]. Many

https://doi.org/10.1016/j.ajem.2021.07.057

0735-6757/(C) 2021

inpatient ICUs across the United States are capacity constrained, and in- creasing strain has been linked to worse patient outcomes [16]. Thus, both “under-triage” of a critically ill patient with UGIB to a general ward bed, and “over-triage” of a stable patient with UGIB to an ICU bed can have adverse consequences.

One potential mitigation strategy is the ED-ICU model, with the in- tention of delivering early, timely, high quality critical care in the ED set- ting [17,18]. An ED-ICU has previously been associated with reduced patient mortality and decreased inpatient ICU utilization in select pa- tient populations [19-23], though its impact on patients with UGIB is unknown. The objective of this study was to determine how an ED- ICU impacts patient and resource utilization outcomes for critically ill ED patients with UGIB. We hypothesized that implementation of an ED-ICU would be associated with shorter time to ICU level care and fewer ICU admissions with similar or improved safety metrics.

  1. Methods

This is a retrospective observational study, conducted in the United States at a single academic medical center with approximately 75,000 adult ED visits per year. The Institutional Review Board at the University of Michigan reviewed and approved this study. This study is presented in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement [24]. All authors had ac- cess to the study data and reviewed and approved the final manuscript. The Joyce and Don Massey Family Foundation Emergency Critical Care Center (EC3) is an ED-ICU that opened in 2015 with the objective of delivering early, comprehensive, high-quality critical care for patients in the ED. All ED patients are initially managed in the main ED, and crit- ically ill patients can be transitioned to a separate group of EC3 clinicians to assume care, regardless of whether an inpatient ICU bed is available [19]. No specific criteria for transfer to the EC3 exist, and patients are transferred at the discretion of the treating ED clinician based on assess- ment of need for ongoing critical care. Similarly, critically ill ED patients can either be transferred to the EC3 or admitted to an inpatient ICU at the discretion of the treating ED clinician, and strict guidelines or criteria do not exist dictating this decision. The EC3 is staffed 24 h per day by one EM attending physician (with or without critical care Fellowship training) and one or two residents, fellows, or physician assistants, and the patient-to-nurse ratio is typically 2:1. Patient movement out of the EC3 (admission to ICU, admission to general care bed, discharge) is at the discretion of the EC3 clinician, regardless of whether inpatient

ICU beds are available.

To test for differences in patient outcomes and resource utilization before and after EC3 implementation, we analyzed the electronic health record data of patients who presented to the adult ED from Sep- tember 2, 2012 through June 30, 2019 with signs and symptoms of UGIB and underwent an Esophagogastroduodenoscopy in the ED, EC3 or within 72 h of ED presentation during the same visit. The Pre-EC3 co- hort was defined as qualifying visits to the ED occurring from Septem- ber 2, 2012 through February 15, 2015 (1033 qualifying ED visits), and the Post-EC3 cohort was defined as qualifying visits to the ED occurring from February 16, 2015 through June 30, 2019 (2755 qualifying ED visits). This time period was selected as a new EHR was deployed in 2012, EC3 opened on February 16, 2015, and defining the pre-and post-EC3 cohorts in this fashion allowed for a maximum number of qualifying visits in each period. Patients were defined as “critically ill” if they were admitted from the ED to an inpatient ICU and/or transferred from the ED to the ED-ICU. Critically ill patients were divided into three cohorts: Cohort A (Pre-EC3 patients admitted to an ICU), Cohort B (Post- EC3 patients managed in EC3), and Cohort C (Post-EC3 patients admit- ted directly to an ICU without receiving care in EC3). Primary analyses compared these critically ill cohorts, as an ED-ICU would be less likely to impact the care or outcomes of non-critically ill ED patients with UGIB.

Patient demographics (age, gender, comorbidities, anticoagulant use, initial laboratory values), resource utilization outcomes (ED LOS, EC3 LOS [distinct/separate from ED LOS], hospital LOS [time interval from arrival to inpatient bed to hospital discharge], ICU LOS, ED disposi- tion, red blood cell transfusions), and mortality (in-hospital, 7-day, and 30-day) were collected and analyzed. Initiation of ICU level of care was defined as transfer to the ED-ICU or inpatient ICU, whichever occurred first. Rates of missing data were low (missing data on n = 4 ED visits [< 1%] for the “EC3 length of stay” variable; missing data on n = 52 ED visits [1.4%] for the “Hospital length of stay” variable; no missing data on any other variables).

    1. Statistical analysis

Bivariate and multiple variable linear and logistic regression analy- ses were used to test hypotheses about cohort differences in resource utilization (Table 2) and patient (Table 3) outcomes. For all regression analyses, dummy variables were created in order to represent cohort differences, and variables were tested simultaneously in all models. For resource utilization outcomes in Table 2, no additional covariates were included in the regression models. For mortality outcomes in Table 3, we statistically controlled for age, sex, Charlson comorbidity score, ED triage Acuity level, anticoagulant use, and initial hemoglobin level. Because some patients had multiple ED visits, all analyses used cluster-robust standard errors to account for possible within-patient correlation [25]. An alpha level of 0.05 was used for all analyses, and all hypothesis tests were two-sided. Analyses were conducted with the Stata software package (Stata Corp., TX, USA, 2017). Statistical anal- ysis was conducted from January through August 2020.

  1. Results
    1. Participants and descriptive data

We identified 3788 ED visits with signs and symptoms of UGIB and performance of EGD within 72 h of ED arrival; 1033 (27%) Pre-EC3 and 2755 (73%) Post-EC3 (Table 1). Of the 1033 Pre-EC3 patients, 200 (19.4%) were critically ill and admitted to an ICU [Cohort A]. Of the 2755 Post-EC3 patients, 682 (24.8%) were critically ill and managed in EC3 [Cohort B], whereas 61 (2.2%) were critically ill and admitted di- rectly to an ICU without receiving care in EC3 [Cohort C].

Comparing critically ill patients managed in the ED-ICU (Cohort B) to critically ill patients Pre-EC3 (Cohort A), the patients in Cohort B were older (mean age 61.5 vs 57.7 years, p < 0.05) and had higher baseline Charlson Comorbidity Index [representing more chronic comorbidity] (6.1 vs 5.1, p < 0.05). The cohorts were otherwise similar with regards to gender, prevalence of cirrhosis, rate of anticoagulant and antiplatelet use, and initial laboratory values [Table 1]. Comparing critically ill pa- tients Post-EC3 managed in the ED-ICU (Cohort B) to critically ill pa- tients Post-EC3 admitted directly to the ICU without receiving care in the ED-ICU [Cohort C], the patients in Cohort B had higher prevalence of cirrhosis (42.7% vs 24.6%, p < 0.05) and lower initial hemoglobin (8.8 vs 9.8 g/dl, p < 0.05). The cohorts were otherwise similar with regards to age, gender, comorbidities, and rate of antiplatelet and anti- coagulant use.

    1. Main results

Mean ED LOS was shorter in Cohort B compared to Cohort A (3.8 vs 4.3 h, p < 0.05) [Table 2], while ED LOS was similar between Cohorts B and C. Median EC3 LOS (applicable only to Cohort B) was 9.6 h. The mean time interval from ED presentation to ICU level care was shorter in Cohort B (3.8 h) than either Cohort A (6.3 h, p < 0.05) or Cohort C (7.7 h, p < 0.05), and a significantly greater proportion of patients in Co- hort B received ICU level care within six hours of ED arrival (85.3 vs 52.0 vs 57.4%, respectively, p < 0.05).

Patient demographics and baseline characteristics (N = 3788 ED visits)

Pre-EC3

Post-EC3

Comparison

of critically

ill patients

Characteristic

All ED

ED admissions

ED admissions to

All ED

ED to EC3

ED admissions

ED admissions

p value

p value

patients

to ICU [Cohort

non-ICU or

patients

patients

directly to ICU

to non-ICU or

comparing

comparing

(n = 1033)

A] (n = 200)

discharge

(n = 2755)

[Cohort B]

[Cohort C]

discharge

Cohort

Cohort

(n = 833)

(n = 682)

(n = 61)

(n = 2012)

A toB

B to C

Mean age, years

60.3

57.7

60.9

61.2

61.5

61.6

61.1

< 0.05

0.99

Female gender, %

42.3

39.5

43.0

42.5

39.4

27.9

44.0

0.99

0.06

Comorbidities, %

Cirrhosis

20.4

39.5

15.8

28.3

42.7

24.6

23.5

0.46

< 0.05

end stage renal disease

23.7

26.0

23.2

28.7

31.5

41.0

27.4

0.15

0.17

Congestive heart failure

19.2

20.5

18.8

23.3

22.9

32.8

23.1

0.50

0.14

Charlson comorbidity index

4.3

5.1

4.1

5.3

6.1

5.7

5.0

< 0.05

0.45

Anticoagulant use, %

14.6

12.5

15.1

13.9

13.0

14.8

14.2

0.85

0.77

Antiplatelet use, %

19.7

19.0

19.8

19.9

18.9

24.6

20.1

0.98

0.38

Presenting heart rate, bpm,

90.9

98.7

89.0

90.4

95.8

92.8

88.5

0.12

0.33

mean

Presenting systolic blood

123.9

112.3

126.7

124.8

117.8

118.8

127.4

< 0.05

0.43

pressure, mmHg, mean

Initial hemoglobin, mean (g/dl)

9.6

9.0

9.7

9.4

8.8

9.8

9.5

0.34

< 0.05

Initial lactate, mean (mmol/L)

2.1

2.9

1.8

2.1

2.9

3.0

1.7

0.80

0.80

Reversal agent? given during

6.4

8.0

6.0

6.4

12.5

9.8

4.2

0.06

0.52

hospitalization, %

Legend: ED: Emergency Department, ICU: intensive care unit, EC3: Emergency Critical Care Center.

Table 2

Resource utilization (N = 3788 ED visits)

Pre-EC3

Post-EC3

Comparison

of critically

ill patients

Characteristic

All ED

ED admissions

ED admissions

All ED

ED to EC3

ED admissions

ED admissions

p value

p value

patients

to ICU [Cohort

to non-ICU or

patients

patients

directly to ICU

to non-ICU or

comparing

comparing

(n = 1033)

A] (n = 200)

discharge

(n = 2755)

[Cohort B]

[Cohort C]

discharge

Cohort

Cohort

(n = 833)

(n = 682)

(n = 61)

(n = 2012)

A toB

B to C

ED length of stay, hours, mean

4.7

4.3

4.7

5.0

3.8

4.4

5.5

< 0.05

0.19

EC3 length of stay, hours, median

n/a

n/a

n/a

n/a

9.6

n/a

n/a

Presentation to ICU level care, hours,

n/a

6.3

n/a

n/a

3.8

7.7

n/a

< 0.05

< 0.05

mean

Proportion of patients receiving ICU

n/a

52.0

n/a

n/a

85.3

57.4

n/a

< 0.05

< 0.05

level care within 6 h of ED

arrival, %

ED Disposition, %

ICU admission

19.4

9.3

28.6

Floor admission

80.3

89.4

69.1

Discharge

0.3

1.2

2.1

Expired in ED

0

0.1

0.3

Mean time from ED presentation to

29.5

15.8

32.8

30.3

17.2

23.6

35.0

0.26

< 0.05

EGD, hours

EGD performed within 24 h of ED

46.4

80.5

38.3

47.1

74.8

63.9

37.2

0.10

0.06

arrival, %

Location of EGD, %

ED (resuscitation bay, EC3)

4.5

12.0

2.6

16.5

55.4

3.3

3.7

< 0.05

< 0.05

Medical procedures unit

78.9

18.5

93.4

78

37.5

31.1

93.1

< 0.05

0.29

Intensive Care Unit

16.6

69.5

3.8

5.4

6.9

63.9

3.1

< 0.05

< 0.05

Other

0.1

0

0.1

0.1

0.1

1.6

0

ICU length of stay, hours, mean

74.5

99.7a

138.5

0.40

0.10

Hospital length of stay, days, mean

4.8

7.3

4.1

5.1

6.2

10.0

4.6

< 0.05

< 0.05

ICU to floor within 24 h of

4.1

21.0

0.7

1.9

8.2

< 0.05

0.08

admission, %

Floor to ICU within 24 h of

2.7

n/a

3.4

2.3

1.8

n/a

2.5

admission, %

Mean units PRBC transfused during

1.1

2.0

0.9

1.9

2.9

3.1

1.5

<0.05

0.63

hospitalization

Number of units PRBC transfused

during hospitalization, %

0

57.1

31.5

63.3

39.7

22.4

27.9

45.9

< 0.05

0.37

1 – 2

31.1

43.5

28.1

33.2

33.0

26.2

33.5

< 0.05

0.26

3 – 4

8.2

17.0

6.1

17.7

26.0

23.0

14.8

< 0.05

0.60

>= 5

3.6

8.0

2.5

9.3

18.6

23.0

5.7

< 0.05

0.42

Second look scope performed within

4.6

12.0

2.9

4.9

11.1

8.2

2.6

0.74

0.43

72 h, %

Legend: ED: Emergency Department, ICU: intensive care unit, EC3: Emergency Critical Care Center, PRBC: Packed red blood cells, EGD: Esophagogastroduodenoscopy.

a Based on n = 237.

Table 3

Patient outcomes (N = 3788

ED visits)

Pre-EC3

Post-EC3

Comparison of patients

critically ill

Characteristic

All ED

ED admissions to

ED admissions to

All ED

ED to EC3

ED admissions

ED admissions to

p value

p value

patients

ICU [Cohort A]

non-ICU or

patients

patients

directly to ICU

non-ICU or

comparing

comparing

(n = 1033)

(n = 200)

discharge

(n = 2755)

[Cohort B]

[Cohort C]

discharge

Cohort

Cohort

(n = 833)

(n = 682)

(n = 61)

(n = 2012)

A toB

B to C

Unadjusted mortality In-hospital mortality, %

2.3

7.0

1.2

3.1

6.5

13.1

1.6

0.79

0.13

7-day mortality, %

1.6

4.0

1.1

1.6

3.4

4.9

0.9

0.69

0.59

30-day mortality, %

Risk-adjusted? mortality In-hospital mortality, %

5.6

2.3

11.0

6.0

4.3

1.5

5.9

3.1

10.9

5.2

16.4

10.3

3.9

1.7

0.95

0.67

0.26

0.20

7-day mortality, %

1.6

3.3

1.3

1.6

2.7

3.1

0.9

0.63

0.84

30-day mortality, %

5.4

9.9

5.1

5.9

8.9

13.8

3.9

0.68

0.24

Legend: ED: Emergency Department, ICU: intensive care unit, EC3: Emergency Critical Care Center.

  • Statistically controlling for age, sex, Charlson comorbidity score, ED triage acuity level, anticoagulant use, and initial hemogloblin level.

Mean hospital LOS was significantly shorter in Cohort B (6.2 days) than either Cohort A (7.3 days, p < 0.05) or Cohort C (10.0 days, p < 0.05). In the Post-EC3 cohort (n = 2755), fewer patients were admitted to an ICU (9.3 vs 19.4%, p < 0.001), and fewer patients had ICU LOS less than 24 h (0.7 vs 4.1%, p < 0.001). The rate of floor admission with trans- fer to ICU within 24 h of admission was similar in the Pre- and Post-EC3 cohorts (2.7 vs 2.3%, p = 0.47).

The mean time interval from ED arrival to completion of EGD was similar between Cohorts A and B, and was shorter in Cohort B compared to Cohort C (17.2 vs 23.6 h, p < 0.05). In Cohort A and C, most EGDs were performed in the ICU, while in Cohort B most EGDs were performed in the ED/EC3.

Mortality data are presented in Table 3. Comparing critically ill pa- tients (Cohorts A, B, and C), no statistically significant differences be- tween in-hospital, 7-day, or 30-day mortality were observed. Although observed mortality rates were lower in Cohort B, none were statistically significant. Similarly, when statistically controlling for age, gender, Charlson Comorbidity Index, ED triage acuity level, anticoagu- lant use, and initial hemoglobin, risk-adjusted mortality was not signif- icantly different between cohorts. Although observed mortality rates were lower in Cohort B, none were statistically significant.

  1. Discussion

For critically ill ED patients with UGIB, management in an ED-ICU was associated with statistically significant reductions in hospital LOS and rate of inpatient ICU admission. These were observed despite older age and higher Charlson Comorbidity Index in the Post-EC3 cohort.

The observed reductions in hospital LOS associated with manage- ment in an ED-ICU (Cohort B vs Cohort A: 1.1 days shorter; Cohort B vs Cohort C: 3.8 days shorter) demonstrate value via downstream hos- pital bed days saved. This is likely due to the coordinated delivery of early, aggressive critical care, including targeted blood product resusci- tation, frequent vital signs and serial assessments of end organ perfu- sion (mentation, urine output, acid-base status, lactate), reversal of coagulopathy, and bleeding source control via EGD. The unique ED-ICU environment provides the means to rapidly obtain adequate intravenous access, pursue diagnostics, consult specialists (gastroenter- ologists/hepatologists, interventional radiologists, surgeons), and resus- citate efficiently analogous to typical EDs. Simultaneously, it offers lower physician: patient and nurse: patient ratios, frequent vital signs and lab draws, serial and dynamic bedside reassessments, hemody- namic monitoring, and procedural space and resources more compara- ble to typical inpatient ICUs.

A factor that likely contributed to these results was the reduction in time to ICU-level care with implementation of EC3 and greater propor- tion of patients receiving ICU level care within six hours of ED

presentation (Table 2). Previous studies have demonstrated worse out- comes for critically ill ED patients with ED length of stay greater than six hours [26,27]. By creating an ICU environment in the ED and providing means for rapid initiation and delivery of critical care with seamless transition from the ED to the ED-ICU, delays associated with traditional inpatient ICU admission from the ED (ie, no bed availability, cleaning and preparing a bed, verbal report, transport throughout the hospital, etc.) may be avoided [28].

A statistically significant Reduction in admissions to the inpatient ICU was observed Post-EC3 (9.3 vs 19.4%, p < 0.001, Table 2), despite in- creased age and comorbidity in the Post-EC3 cohort. This was likely at- tributable to the reduction in “short-stay” ICU admissions. No universal definition of “short-stay” ICU admission exists, and for purposes of this study we defined one as ICU length of stay less than 24 h. With imple- mentation of an ED-ICU, short-stay ICU admissions for critically ill ED patients with UGIB decreased by 83% (4.1 vs 0.7%). We observed no in- crease in unexpected transfers to ICU for patients admitted to general ward beds within 24 h of admission (2.7 vs 2.3%, Table 2). This observa- tion paired with similar mortality metrics (Table 3) suggest the ob- served reduction in ICU admissions was safe, and was not associated with increased risk for patients on non-ICU wards.

Avoiding short-stay ICU admissions and optimizing ICU utilization for critically ill patients with prolonged critical care requirements may benefit patients and healthcare systems. Doing so may optimize ability to care for decompensating patients on wards and increase availability for outside hospital ICU-to-ICU transfers. Higher ICU capacity strain has been associated with worse patient outcomes [16], and thus strate- gies to avoid ICU strain via preventing avoidable use while not placing patients on non-ICU wards at increased risk are desirable.

We observed no statistically significant difference in mortality of critically ill patients with UGIB between Pre- and Post-EC3 (Cohorts A, B, and C; Table 3). The relatively small sample sizes may have contrib- uted to the lack of statistical significance between cohorts. Similar mor- tality metrics between cohorts, combined with improvED resource utilization (lower rate of ICU admission, shorter hospital LOS, Table 2) likely demonstrates that an ED-ICU can improve the value of care deliv- ered to critically ill ED patients with UGIB.

  1. Limitations

The retrospective nature of this study limits interpretation of results to observations of association, rather than causality. Pre-post cohort studies can be inherently prone to favoring the “post-” cohort, and sec- ular trends in critical care delivery over time may have impacted ob- served results. This study was conducted at a single center in the United States, with a unique ED-ICU care delivery model, and generaliz- ability to other centers is unknown. The uncontrolled, pragmatic nature of this study may contribute to the observed findings. Additional

unmeasured confounding may have contributed to observed results given the observational nature of this study. endoscopic findings and therapeutics/ interventions performed during endoscopy were unable to be collected or analyzed retrospectively, due to electronic medical re- cord limitations. It is thus possible that cohorts differed with respect to rates of variceal bleeding, ulcer, therapeutics, etc., though we are unable to report on this. Similarly, rates of intubation and mechanical ventila- tion during EGD were unable to be extracted, which may also confound outcomes including length of stay.

  1. Conclusion

For critically ill ED patients with UGIB, implementation of an ED-ICU was associated with statistically significant reductions in rate of inpa- tient ICU admission with no difference in key safety outcomes. An ED- ICU model may contribute to improved patient outcomes via more rapid delivery of critical care and avoidance of prolonged ED boarding of critically ill patients, which has been linked to poor outcomes. Avoiding short-stay ICU admissions and optimizing ICU utilization for critically ill patients with prolonged critical care requirements may ben- efit patients and other healthcare systems.

Sources of funding / grant support

none.

Writing assistance

None outside of author group.

Author statement

NLH, RPM, RAH, EDL, MDR, and BSB conceptualized this study, NLH wrote the original draft, and all authors reviewed and edited the final manuscript. JAC provided statistical support and analysis. CB provided administrative support and assisted with data collection / acquisition.

Declaration of Competing Interest

The authors individually and collectively have no conflicts of interest to disclose.

Acknowledgements

The authors would like to acknowledge the Joyce and Don Massey Family Foundation for their generous support in creation of the Joyce and Don Massey Family Foundation Emergency Critical Care Center. The authors would also like to acknowledge Stephanie Laurinec and Amanda Melvin for their administrative support and role in this project.

References

  1. Tielleman T, Bujanda D, Cryer B. Epidemiology and risk factors for upper gastrointes- tinal bleeding. Gastrointest Endosc Clin. 2015;25(3):415-28.
  2. Barkun AN, Almadi M, Kuipers EJ, et al. Management of nonvariceal upper gastroin- testinal bleeding: Guideline recommendations from the International Consensus Group. Ann Intern Med. 2019;171(11):805-22.
  3. Wuerth BA, Rockey DC. Changing epidemiology of upper gastrointestinal hemor- rhage in the last decade: a nationwide analysis. Dig Dis Sci. 2018;63:1286-93.
  4. Barkun AN, Almadi M, Kuipers EJ, et al. Management of nonvariceal upper gastroin- testinal bleeding: guideline recommendations from the International Consensus Group. Ann Intern Med. 2019;171(11):805-22.
  5. Lau JYW, Yu Y, Tang RSY, et al. Timing of endoscopy for acute upper gastrointestinal bleeding. N Engl J Med. 2020;382:1299-308.
  6. Kumar NL, Cohen AJ, Nayor J, et al. Timing of upper endoscopy influences outcomes in patients with acute nonvariceal upper GI bleeding. Gastrointest Endosc. 2017;85 945-952.e1.
  7. Lim LG, Ho KY, Chan YH, et al. urgent endoscopy is associated with lower mortality in high-risk but not low-risk nonvariceal upper gastrointestinal bleeding. Endos- copy. 2011;43:300-6.
  8. Cho SH, Lee YS, YJ K, et al. Outcomes and role of urgent endoscopy in high-risk pa- tients with acute nonvariceal gastrointestinal bleeding. Clin Gastroenterol Hepatol. 2018;16:370-7.
  9. Liang PS, Saltzman JR. A national survey on the initial management of upper gastro- intestinal bleeding. J Clin Gastroenterol. 2014;48:e93-8.
  10. Adam V, Barkun AN. Estimates of costs of hospital stay for variceal and nonvariceal upper gastrointestinal bleeding in the United States. Value Health. 2008;11:1-3.
  11. Liu V, Kipnis P, Rizk NW, et al. Adverse outcomes associated with delayed intensive care unit transfers in an integrated healthcare system. J Hosp Med. 2012;7:224-30. https://doi.org/10.1002/jhm.964.
  12. Escobar GJ, Greene JD, Gardner MN, et al. Intra-hospital transfers to a higher level of care: contribution to total hospital and intensive care unit mortality and length of stay (Los). J Hosp Med. 2011;6:74-80. https://doi.org/10.1002/jhm.817.
  13. Young MP, Gooder VJ, McBride K, et al. Inpatient transfers to the intensive care unit: delays are associated with increased mortality and morbidity. J Gen Intern Med. 2003;18:77-83. https://doi.org/10.1046/j.1525-1497.2003.20441.x.
  14. Robert R, Reignier J, Tournoux-Facon C, et al. Refusal of intensive care unit admission due to a full unit: impact on mortality. Am J Respir Crit Care Med. 2012;185:1081-7. https://doi.org/10.1164/rccm.201104-0729OC.
  15. Cardoso LTQ, Grion CMC, Matsuo T, et al. Impact of delayed admission to intensive care units on mortality of critically ill patients: a cohort study. Crit Care. 2011;15: R28. https://doi.org/10.1186/cc9975.
  16. Wilcox ME, Harrison DA, Patel A, Rowan KM. Higher ICU capacity strain is associated with increased acute mortality in closed ICUs. Crit Care Med. 2020;48:709-16.
  17. Leibner E, Spiegel R, Hsu CH, et al. Anatomy of resuscitative care unit: expanding the borders of traditional intensive care units. Emerg Med J. 2019;36(6):364-8. https:// doi.org/10.1136/emermed-2019-208455.
  18. Jayaprakash N, Pflaum-Carlson, Gardner-Gray J, Hurst G, Coba V, Kinni H, et al. Crit- ical care delivery solutions in the emergency department: evolving models in caring for ICU boarders. Ann Emerg Med. 2020. https://doi.org/10.1016/j.annemergmed. 2020.05.007 ISSN 0196-0644.
  19. Gunnerson KJ, Bassin BS, Havey RA, et al. Association of an emergency department- based intensive care unit with survival and inpatient intensive care unit admissions. JAMA Netw Open. 2019;2:e197584.
  20. Haas NL, Whitmore SP, Cranford JA, et al. An emergency department-based intensive care unit is associated with decreased hospital and intensive care unit utilization for diabetic ketoacidosis. J Emerg Med. 2019. https://doi.org/10.1016/j.jemermed.2019.

10.005 published online Dec 13.

  1. Haas NL, Nafday A, Cranford JA, Yentz SE, Bixby DL, Bassin BS. Implementation of a multidisciplinary care pathway via an emergency department-ICU to improve care of emergency department patients presenting with leukostasis. Crit Care Explor. 2020;2:e0084.
  2. Joseph JR, Haas NL, Joseph JR, Heth J, Szerlip NJ, Bassin BS. Utilization of a resuscita- tive care unit for initial triage, management, and disposition of minor intracranial hemorrhage. Crit Care Explor. 2020;2:e0097.
  3. Haas NL, Larabell P, Schaeffer W, et al. Descriptive analysis of extubations performed in an emergency department-based intensive care unit. West J Emerg Med. 2020;21: 532-7.
  4. Elm Von, Erik, et al.. The strengthening the reporting of observational studies in ep- idemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol. 2008;61(4):344-9.
  5. McNeish DM. Modeling sparsely clustered data: design-based, model-based, and single-level methods. Psychol Methods. 2014;19(4):552.
  6. Chalfin DB, Trzeciak S, Likourezos A, Baumann BM, Dellinger RP, DELAY-ED study group. Impact of Delayed transfer of critically ill patients from the emergency de- partment to the intensive care unit. Crit Care Med. 2007;35(6):1477-83. https:// doi.org/10.1097/01.CCM.0000266585.74905.5A.
  7. Singer AJ, Thode Jr HC, Viccellio P, Pines JM. The association between length of emer- gency department boarding and mortality. Acad Emerg Med. 2011;18(12):1324-9. https://doi.org/10.1111/j.1553-2712.2011.01236.x.
  8. Mohr NM, Wessman BT, Bassin B, et al. Boarding of critically ill patients in the emer- gency department. Crit Care Med. 2020;48(8):1180-7. https://doi.org/10.1097/CCM. 0000000000004385.