Time-Varying Associations Between Selected Clinical Indicators And In-Hospital Mortality During The Early Period Of...

Area of Science:

• Critical Care Medicine
• Clinical Epidemiology
• Health Services Research

Background:

• Patients in intensive care units (ICUs) often experience clinical instability in early care.
• Predicting in-hospital mortality risk in critically ill patients requires understanding dynamic clinical variables.

Purpose of the Study:

• To describe how the association between clinical variables and in-hospital mortality changes over short, sequential time periods during early ICU care.
• To identify time-varying risk factors for mortality in the initial phase of intensive care.

Main Methods:

• Retrospective analysis of 19,439 ICU encounters from 17,769 patients (2018-2023).
• Encounters segmented into time intervals based on complete blood count results (up to 20 intervals).
• Multivariable logistic regression used to estimate odds ratios for in-hospital death associated with nine clinical variables and life-support treatments.

Main Results:

• One-third of in-hospital mortality occurred within the first 20 time intervals of ICU care.
• Odds ratios for age, sex, platelet count, hemoglobin, and transfusion showed minor variations over time.
• Odds ratios for ventilator use, dialysis, pressors, and bilirubin exhibited wider fluctuations, ranging from 1.43 to 4.75.

Conclusions:

• The strength of association between clinical variables and in-hospital mortality is time-dependent during early ICU care.
• Dynamic monitoring of clinical variables is essential for accurate mortality risk assessment in critically ill patients.