Nonlinear Association of Hospital Length of Stay and In-Hospital Mortality of Elderly Patients Admitted Emergently for Colon Cancer: Analysis of 40,421 Patients

Author Type(s)

Faculty

Document Type

Abstract

Publication Date

11-2021

DOI

10.1016/j.jamcollsurg.2021.08.086

Journal Title

Journal of the American College of Surgeons

Department

Surgery

Abstract

INTRODUCTION: The study aimed to evaluate the predictors of mortality among elderly patients with colon cancer who underwent emergency admission.

METHODS: Data were obtained from the National Inpatient Sample, 2005–2014, to evaluate elderly patients with colon cancer who underwent emergency admission. Demographics, etiology, clinical course, and hospital length of stay were collected. Multivariable generalized additive model (GAM) and multivariable logistic regression with backward elimination were used to identify the associations of predictors and in-hospital mortality. The models were adjusted for age, sex, race, income quartile, severity of illness subclass, and the associated comorbidities.

RESULTS: 40,421 elderly patients were included. Their mean (SD) age was 78.7 (7.95) years. 57.5% were females. Mean hospital length of stay (HLOS) was 10.5 days and 5.65% died in the hospital. The multivariable regression GAM identified a significant and direct linear association of age and mortality (EDF=1.67, p

CONCLUSION: The key findings of the study included the linear associations of age and mortality, and the U-shaped association of HLOS and mortality in elderly patients with colon cancer undergoing emergency admission. For each additional day after day 15, the mortality odds increased by 1.8%.

Link to Full Text

Share

COinS