WAVES - The Lucile Packard Children's Hospital Pediatric Physiological Waveforms Dataset

Area of Science:

• Biomedical Informatics
• Critical Care Medicine
• Data Science

Background:

• Physiological waveform data is crucial for monitoring patients in intensive and acute care settings.
• Existing datasets may lack the scale, duration, or pediatric focus necessary for comprehensive research.
• High-frequency data capture is essential for detailed physiological analysis.

Purpose of the Study:

• To introduce the WAVES dataset, a large-scale collection of pediatric physiological waveform data.
• To describe the characteristics and organization of the WAVES dataset.
• To highlight the potential research applications of the WAVES dataset.

Main Methods:

• Compilation of 9 years of high-frequency physiological waveform data from a single academic pediatric medical center.
• Inclusion of 1 to 20 concurrent waveforms per patient encounter.
• De-identification, cleaning, and organization of approximately 10.6 million hours of data from over 50,000 patient encounters.

Main Results:

• The WAVES dataset is the largest pediatric-focused physiological waveform dataset available for research.
• It is the second largest overall physiological waveform dataset accessible for research purposes.
• Initial analyses show promise for clinical applications like non-invasive blood pressure monitoring and data imputation.

Conclusions:

• The WAVES dataset represents a significant resource for advancing research in pediatric critical care.
• Its scale and quality facilitate novel investigations into physiological monitoring and data analysis techniques.
• The dataset is poised to drive innovation in both clinical practice and data science methodologies.