Icon Legend

This session is not in your schedule.

This session is in your schedule. Click again to remove it.

Session Icons

Thought Leader Session

Guidelines Session

Aging Adult Track

Additional Registration Fee

Research Snapshot Theater: Quality and Patient Safety/Future of Critical Care, Adult I

Incidence of Ultra-Short ICU stays (< 24 Hours): A Retrospective Study of the Cosmos Database

Sunday, March 22, 2026

12:15 PM - 1:15 PM Central Time

Location: Connections Central - RST 11

Meagan D'Antone, DO (she/her/hers)

Emory University, Georgia

Disclosure(s): No relevant financial relationship(s) to disclose.
Sharaf Khan, M.D.

Emory Critical Care, Georgia

Disclosure(s): No relevant financial relationship(s) to disclose.
DM

Donal Maberry, MD

Assistant Professor
Emory University School of Medicine

Disclosure information not submitted.
SS

Samantha Strickler, DO, FACEP

Assistant Professor
Emory University School of Medicine

Disclosure information not submitted.
DP

Deepa Patel, MD

Associate professor
Emory University School of Medicine

Disclosure information not submitted.
CZ

Chet C. Zalesky, MD,MSc

Assistant Professor
Emory University School of Medicine

Disclosure information not submitted.

Introduction: From 2008 to 2019, there has been a decline in ICU admissions but an increase in short ICU stays 2-7 days in length. Previous single-center studies have also identified that 20-30% of ICU admissions lasted < 24 hours. Using the nationally representative Cosmos Database (Epic Systems), we hypothesized that 20% of ICU admissions were < 24h in length across a variety of care settings.

Methods: This is a retrospective analysis of adult ICU admissions within the Epic Cosmos database from January 1, 2017 to September 30, 2024. All ICU admissions for patients 18 years or older with only one ICU stay during their hospitalization were included. Pediatric hospitals and ICU stays of less than 4 hours were excluded. An ultrashort ICU stay was defined as an ICU stay of < 24 hours. Hospital categories (academic, community, safety net, critical access) are defined by Epic according to ACGME training programs, location, and payor make-up. Logistic regression analyzed how hospital type affects ultrashort ICU stay odds.

Results: 3,897,787 ICU admissions occurred in the sample timeframe. Among all ICU admissions,24.8% (n=968,552) were ultrashort. Survival to hospital discharge was 81% and was similar between ultrashort and other ICU admissions. The median ICU length of stay (LOS) was 1.87 days (1.0-3.5), with the median LOS for ultrashort ICU stays was 0.72 days (0.47-0.87). Median overall hospital LOS was 5 days (3.0-9.0). Academic safety net hospitals had the fewest ultrashort ICU stays (22% n=204,256/943,942) while community (26%, n=377,564 of 1,425,464) and academic hospitals (27% n=107,741 of 400,720) had the most. When controlling for department volume and payor class, academic safety net (OR 0.78, 95% CI of 0.77-0.78, p < 0.001) and community safety net (OR 0.93, 95% CI 0.92-0.93, p < 0.001) hospitals had the lowest odds of an ultrashort ICU stay compared to general community hospitals. Academic hospitals were the only group with an increased odds of a short ICU stay (OR 1.03, 95%CI 1.02-1.04, p < 0.001).

Conclusions: Within the Cosmos Database, 24.8% of all ICU stays were < 24 hours in duration. Additionally, academic and community Hospitals had the highest proportion of ultrashort ICU stays. These admissions represent a potential area to improve patient care and ICU utilization.