Biorhythms derived from consumer wearables predict postoperative complications in children

Area of Science:

• Biomedical Engineering
• Pediatric Surgery
• Digital Health

Background:

• Postoperative complications in children present significant health risks.
• Current methods for detecting complications after discharge rely on subjective reports, delaying diagnosis.
• Wearable devices offer objective, continuous monitoring for early complication detection.

Purpose of the Study:

• To investigate the relationship between biorhythm metrics and pediatric postoperative recovery.
• To assess the utility of wearable-derived biorhythms for predicting complications in children.
• To explore the potential of unobtrusive monitoring for enhancing pediatric surgical care.

Main Methods:

• 103 children undergoing appendectomy were monitored with wearables for 21 days post-surgery.
• Biorhythm metrics (circadian and ultradian rhythms) were extracted from activity and heart rate data.
• A machine learning model was developed to predict postoperative complications using biorhythm data.

Main Results:

• The machine learning model accurately predicted postoperative complications up to 3 days before clinical diagnosis.
• The model achieved 91% sensitivity and 74% specificity in complication detection.
• Wearable-derived biorhythms demonstrated a strong correlation with recovery status.

Conclusions:

• Wearable-derived biorhythms provide a promising, objective method for monitoring pediatric postoperative recovery.
• This technology can enable earlier detection of complications, improving patient outcomes.
• The approach has significant potential for pediatric health monitoring in diverse clinical settings.