New England Surgical Society

Back to 2020 Abstracts


Does Size of Intensive Care Units Affect Mortality Among Sepsis Patients? Results of a Nationwide Retrospective Analysis
*Leon Naar, *Majed W. El Hechi, *Mohamad El Moheb, *Jason Fawley, *Jonathan Parks, *April E. Mendoza, *Noelle N. Saillant, *George Velmahos, Haytham Kaafarani, *Jarone Lee
Massachusetts General Hospital, Harvard Medical School, Boston, MA

Objective: Regionalization of sepsis centers is hypothesized to improve outcomes. Therefore, benchmarking quality of intensive care units (ICU) care is critical. We sought to compare mortality among patients with sepsis admitted to smaller versus larger ICUs.Design: Retrospective nationwide study.Setting: eICU Collaborative Research Database.Patients: Patients admitted to the ICU for sepsis with lactate ≥2 mmol/L within 24 hours of admission. ICU size was stratified according to the hospitalís total ICU capacity: Small-to-Medium(1-15 beds), Large(16-30 beds), and Extra-Large(>30 beds).Intervention(s): N/AMain outcome measure(s): ICU mortality was assessed using a multivariable regression model adjusting for demographics, APACHE-IV, qSOFA scores, source of sepsis, surgical patients, and ICU size. Regional clustering was analyzed using a mixed-effect model.Results: We identified 7,148 patients admitted to different hospitals with the following ICU capacity; Small-to-Medium:1,801(25.2%), Large:3,342(46.8%), and Extra-Large:2,005(28.1%). Mean APACHE scores were 68, 75, and 75 (p-value<0.001), respectively. Median qSOFA score was 2. Most common source of sepsis was the respiratory system(36.0%). Overall ICU mortality rate was 14.0%. Kaplan-Meier curves for the unadjusted ICU mortality are shown in the Figure. Mixed-effect multivariable regression showed that smaller ICU size was not associated with higher mortality among patients with sepsis (Table).
Conclusions: Mortality among patients with sepsis requiring ICU care was similar regardless of ICU size in the US. Therefore, regional sepsis centers might not be required, and sepsis care should continue within our current system.

Mortality in the ICUOdds RatioP-value95% Conf. Intervals
Age1.010.029100-1.01
Sex (female)1.050.4430.92-1.20
Race (Ref: Caucasian)
African American0.820.0270.70-0.98
Hispanic0.680.3760.29-1.60
Asian1.200.6090.60-2.37
Other0.850.0770.71-1.02
Apache score1.04<0.0011.03-1.04
qSOFA score (Ref: 0)
12.770.0011.50-5.14
23.28<0.0011.84-5.83
35.47<0.0012.56-11.69
Surgical patients0.840.5550.47-1.50
Source of sepsis (Ref: Urinary tract)
GI tract2.04<0.0011.84-2.26
Respiratory system1.72<0.0011.54-1.91
Cutaneous and soft tissues1.380.0421.01-1.89
Gynecology1.180.9010.09-15.76
Other1.61<0.0011.40-1.86
ICU size (Ref: >30 beds)
16-30 beds0.960.7900.69-1.33
1-15 beds0.960.8960.54-1.70


Back to 2020 Abstracts