Association Between Levels of Intensive Care and In-Hospital Mortality in Patients Hospitalized for Sepsis Stratified by Sequential Organ Failure Assessment Scores*

Abstract

OBJECTIVES: To assess the association between levels of intensive care and in-hospital mortality in patients hospitalized for sepsis, stratified by Sequential Organ Failure Assessment (SOFA) score at admission. DESIGN: A nationwide, propensity score-matched, retrospective cohort study. SETTING: A Japanese national inpatient database with data on 70–75% of all ICU and high-dependency care unit (HDU) beds in Japan. PATIENTS: Adult patients hospitalized for sepsis with SOFA scores greater than or equal to 2 on their day of admission between April 1, 2018, and March 31, 2021, were recruited. Propensity score matching was performed to compare in-hospital mortality, and patients were stratified into 10 groups according to SOFA scores. INTERVENTIONS: Two exposure and control groups according to treatment unit on day of admission: 1) ICU + HDU versus general ward and 2) ICU versus HDU. MEASUREMENTS AND MAIN RESULTS: Of 97,070 patients, 19,770 (20.4%), 23,066 (23.8%), and 54,234 (55.9%) were treated in ICU, HDU, and general ward, respectively. After propensity score matching, the ICU + HDU group had significantly lower in-hospital mortality than the general ward group, among cohorts with SOFA scores greater than or equal to 6. There were no significant differences in in-hospital mortality among cohorts with SOFA scores 3–5. The ICU + HDU group had significantly higher in-hospital mortality than the general ward among cohorts with SOFA scores of 2. The ICU group had lower in-hospital mortality than the HDU group among cohorts with SOFA scores greater than or equal to 12. There were no significant differences in in-hospital mortality among cohorts with SOFA scores 5–11. The ICU group had significantly higher in-hospital mortality than the general ward group among cohorts with SOFA scores less than or equal to 4. CONCLUSIONS: Patients hospitalized for sepsis with SOFA scores greater than or equal to 6 in the ICU or HDU had lower in-hospital mortality than those in the general ward, as did those with SOFA scores greater than or equal to 12 in the ICU versus HDU.

Current Clinical Practice of Laboratory Testing of the Hemostasis and Coagulation System in Patients with Sepsis: A Nationwide Observational Study in Japan

“…Organ dysfunction was determined if the SOFA subcomponent score was 2 or more. To determine the source of infection, we utilized the definition of ICD-10 codes reported previously ( Supplementary Table 1 ) (17) . We also collected in-hospital mortality as a general outcome.…”

Does Critical Care Really Make A Difference?*

“…In an era where artificial intelligence will be increasingly employed to determine treatment plans and public policy, it is especially important to fully understand the circumstances underlying the published finding. The study by Ohbe et al (8) offers some support for managing sepsis patients without organ dysfunction in a non-ICU setting. This should not be misinterpreted as a cost-saving mandate to manage such patients outside the ICU.…”

“…In this issue of Critical Care Medicine, the study by Ohbe et al (8) provides novel insight into the value of high-intensity care on in-hospital mortality in patients with sepsis. Unlike prior studies, the focus is on the location of care (ICU vs. high-dependency unit [HDU] vs. acute care ward) rather than specific intensivist staffing, although nursing and physician input did vary by location.…”

“…A major strength of Ohbe's et al (8) study is that it captures over 70% of Japanese ICU admissions. But Japan may not be representative of other jurisdictions.…”