Adaptive Biofeedback-Driven Wearable Devices for Enhancing Post-Surgical Patient Recovery: A Machine Learning Approach

Abstract

Recovery after surgery can be slow and is often complicated by secondary health problems. Personalized and timely reminders to patients encouraging them to perform activities associated with faster recovery may help to improve patients’ experience and outcome following surgery. The study aim was to develop an adaptive biofeedback system, which can provide reminders for mobilization at optimal times during recovery and examination of its potential outcome benefit for patients recovering after abdominal surgery. Leveraging adaptive machine learning for decision making, a wearable system was built to monitor posture and provide biofeedback reminders for mobilization only when required. Data collected from abdominal surgery patients (n = 41) during the first 15 days after surgery have been analyzed to determine whether adaptive biofeedback affects patient outcome. The results reveal that adaptive feedback leads to higher patient engagement and faster recovery following first-step ambulation. Despite the preliminary nature of these results, they suggest that machine learning can enable postoperative biofeedback systems to better support patients’ recovery after surgery and are worthy of further investigation in a larger study. The study also provides evidence for the widespread applicability of such adaptive reminder systems—whenever user behavior has an intuitive connection to recovery or outcome.