Absence of a Weekend Effect on In-Hospital Mortality in Upper Gastrointestinal Bleeding: A Multisite Study from an Integrated Health System.

OBJECTIVES

The "weekend effect"-worse outcomes among patients admitted on Saturdays or Sundays-has been reported in upper gastrointestinal bleeding (UGIB) and attributed to delayed endoscopy and reduced after-hours resources. We examined whether weekend admission is associated with worse outcomes within a contemporary integrated health system with continuous endoscopy availability.

METHODS

We performed a retrospective cohort study of adults (18-89 years) admitted to Northwell Health hospitals with a primary or secondary diagnosis of UGIB between January 2019 and January 2024. Weekend admission was defined as arrival on a Saturday or Sunday. The primary outcome was in-hospital mortality; secondary outcomes were intensive care unit (ICU) admission, length of stay (LOS), and 30-day readmission. Groups were compared using chi-square and Wilcoxon rank-sum tests. A multivariable logistic regression model for mortality included weekend admission (forced entry), age, liver cirrhosis, and shock.

RESULTS

Of 7,640 patients, 1,927 (25.2%) were admitted on a weekend. Baseline characteristics were broadly similar, although liver cirrhosis was modestly more frequent among weekend admissions (8.0% vs 6.5%, P = 0.03). Overall in-hospital mortality was 5.5% and did not differ by admission timing (adjusted odds ratio [aOR] 1.06, 95% CI 0.83-1.35, P = 0.67). Independent predictors of mortality were shock (aOR 19.08, 95% CI 15.33-23.76), liver cirrhosis (aOR 1.81, 95% CI 1.23-2.67), and age (aOR 1.04 per year, 95% CI 1.03-1.05). ICU admission (23.3% vs. 21.9%, P = 0.21), median LOS (4.49 vs. 4.66 days, P = 0.43), and 30-day readmission (1.8% vs. 2.3%, P = 0.18) did not differ significantly. The null association persisted in sensitivity analyses using an off-hours exposure window (aOR 1.02) and holiday reclassification (aOR 1.04), after adjustment for AIMS65 severity (aOR 1.08), and in a prespecified clinically informed model (aOR 1.06).

DISCUSSION: In a large multicenter cohort within an integrated health system, weekend admission was not associated with worse outcomes in UGIB, and this null association was robust across alternative off-hours and holiday exposure definitions, severity adjustment, and a prespecified clinically informed model. Disease severity, particularly shock and cirrhosis, was the dominant determinant of mortality. These findings are consistent with standardized, continuously available GI bleed care pathways mitigating temporal disparities in UGIB outcomes.