Heart rate complexity helps mortality prediction in the intensive care unit: A pilot study using artificial...

Area of Science:

• Critical Care Medicine
• Artificial Intelligence in Healthcare
• Cardiovascular Physiology

Background:

• Accurate prediction of mortality in intensive care units (ICUs) is vital for patient management.
• The Simplified Acute Physiology Score II (SAPS-2) is a common tool but requires extensive clinical and laboratory data.
• This study aimed to develop an AI model using hemodynamic parameters for ICU mortality prediction.

Purpose of the Study:

• To develop and validate an artificial intelligence (AI) model for predicting ICU mortality within 24 hours.
• To compare the performance of the AI model against the SAPS-2.
• To explore the utility of hemodynamic parameters and heart rate curve structures in mortality prediction.

Main Methods:

• Analysis of hemodynamic parameters and heart rate curve structures in 1888 ICU patients.
• Development and validation of a machine-learning AI model.
• Comparative analysis of AI model performance versus SAPS-2 based on accuracy, calibration, and generalizability.

Main Results:

• Key mortality predictors identified: Glasgow Coma Scale, heart rate complexity, age, diastolic blood pressure duration, heart rate variability, and blood pressure thresholds.
• The AI model demonstrated comparable or superior performance to SAPS-2 in predicting ICU mortality.
• Observed mortality rate in the study cohort was 24.0%.

Conclusions:

• An AI model integrating heart rate and blood pressure analysis with clinical data offers a novel method for ICU mortality prediction.
• This AI approach provides a potentially simpler alternative to existing tools reliant on extensive data.
• Integration into ICU monitoring systems could streamline mortality prediction and patient management.