Patterns in continuous pulse oximetry data prior to pulseless electrical activity arrest in the general care setting

Area of Science:

• Critical Care Medicine
• Biomedical Engineering
• Clinical Monitoring

Background:

• Pulseless electrical activity (PEA) arrest poses a significant risk in general inpatient settings.
• Early detection of PEA is crucial for timely intervention and improved patient outcomes.
• Continuous monitoring of vital signs offers potential for predictive analytics.

Purpose of the Study:

• To determine if features derived from continuous pulse oximetry data can provide early warning of PEA arrest.
• To analyze SpO2 and pulse rate data for predictive patterns preceding PEA events.

Main Methods:

• Retrospective analysis of continuous pulse oximetry data (SpO2 and pulse rate) from PEA (n=38) and control (n=42) patient cohorts.
• Calculation of central tendency and variation measures over various time intervals (1 min to 1 hour).
• Logistic regression to assess the predictive ability of derived features for future PEA events.

Main Results:

• PEA arrest patients exhibited lower mean SpO2 and higher mean pulse rates compared to controls.
• Increased SpO2 and pulse rate variability was observed in the PEA cohort, starting hours before the event.
• Significant differences in SpO2 and pulse rate features were detected up to 20 minutes prior to PEA rescue events.

Conclusions:

• Features from continuous pulse oximetry data show significant differences between PEA and control groups.
• Automated calculation and clinician notification systems integrating these features may enhance patient safety through early PEA detection.