VoxCare: Studying Natural Communication Behaviors of Hospital Caregivers through Wearable Sensing of Egocentric Audio

2026-03-11

AI Summary

This work addresses the challenge of unobtrusively quantifying natural communication behaviors of healthcare professionals in real clinical settings. To this end, we propose a privacy-preserving wearable audio sensing approach that leverages edge computing to extract first-person acoustic features in real time without storing raw audio. A teacher–student learning framework, guided by a speech foundation model, is introduced to identify valid speech activity and derive interpretable communication metrics. Deployed in an actual hospital environment, our method enables fine-grained, continuous monitoring of interpersonal interactions for the first time, revealing significant differences in communication patterns across shifts and departments. Furthermore, we demonstrate strong associations between communication frequency and duration with clinical workload and stress levels, offering data-driven insights to enhance team collaboration and care quality.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Natural Language Processing: Speech
• Humans and AI: Brain-Sensing and Analysis

Application Category

• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Healthcare professionals work in complex, high-stakes environments where effective communication is critical for care delivery, team coordination, and individual well-being. However, communication activity in everyday clinical settings remains challenging to measure and largely unexplored in human behavioral research. We present VoxCare, a scalable egocentric wearable audio sensing and computing system that captures natural communication behaviors of hospital professionals in real-world settings without storing raw audio. VoxCare performs real-time, on-device acoustic feature extraction and applies a speech foundation model-guided teacher-student framework to identify foreground speech activity. From these features, VoxCare derives interpretable behavioral measures of communication frequency, duration, and vocal arousal. Our analyses reveal how, when, and how often clinicians communicate across different shifts and working units, and suggest that communication activity reflects underlying workload and stress. By enabling continuous assessment of communication patterns in everyday contexts, this study provides data-driven approaches to understand the behaviors of healthcare providers and ultimately improve healthcare delivery.

Problem

• healthcare communication
• natural communication behaviors
• clinical settings
• wearable sensing
• egocentric audio

Innovation

• egocentric audio sensing
• speech foundation model
• teacher-student framework
• on-device acoustic feature extraction
• vocal arousal