Predicting Subjective Recovery from Lower Limb Surgery Using Consumer Wearables.

Area of Science:

• Digital health
• Wearable technology
• Rehabilitation medicine

Background:

• Monitoring rehabilitation requires long-term individual baseline data for objective functional recovery assessment.
• Consumer-grade wearable devices offer continuous tracking of daily functioning before medical events.

Purpose of the Study:

• To assess the utility of long-term wearable data for monitoring functional recovery after lower limb surgery.
• To identify patterns in activity, heart rate, and sleep related to surgery type and recovery time.
• To evaluate the predictive accuracy of individual recovery trajectories using wearable data.

Main Methods:

• Collected Fitbit data (steps, heart rate, sleep) from 1,324 individuals 26 weeks pre- to 26 weeks post-lower limb surgery.
• Subgrouped participants by surgery type: fracture repair, tendon/ligament repair, joint replacement.
• Utilized linear mixed models and XGBoost for analyzing activity changes and predicting recovery time.

Main Results:

• Wearable data trajectories (12 weeks pre- to 26 weeks post-surgery) captured individual activity changes relative to baseline.
• Recovery trajectories varied significantly by surgery type and reflected age-related effects.
• Individual long-term recovery could be predicted one month post-surgery with high accuracy (AUROC 0.734) using baseline data.

Conclusions:

• Long-term, passively collected wearable data facilitates relative assessment of individual recovery.
• This approach is a foundational step towards data-driven interventions in rehabilitation.
• Individualized baseline data is crucial for accurate prediction of functional recovery.