Project 3: An automatic speech recognition tool to assess, feedback and improve team communication in operating room teams

Recruiting organisation: University of Bern, Bern, Switzerland 

Supervisors: Prof. Dr. Guido Beldi, Prof. Dr Tanja Manser, Dr Sandra Keller

Doctoral Candidate: Lilien Schewski 

Secondments:  Bonn University Hospital (UKB) in Germany and REGION HOVEDSTADEN (Region H) in Denmark

Summary: Based on automatic speech analysis, we will facilitate the development of an automated tool that can assess and provide feedback to surgeons on their communication patterns, with a particular focus on the expression of negative emotions and stress in their intra-operative communication.  

The project has following aims: (1) Identification, in audio-recorded surgical operations, patterns of speech that coincide with sequences stress and experienced negative emotions. (2) Establish a validated methodology to automatically assess negative emotions and stress based on audio-recordings. (3) Prototype a feedback system to inform surgeons on their communication patterns assessed based on audio-recordings of operations. 

Background

Teamwork in the operating room is a key aspect for patient safety. However, teamwork is highly affected by team communication and the surgeon`s emotional state. Negative emotions and stress can impact both surgeon`s technical performance and nontechnical performance, including communication and decision-making. Furthermore, negative emotions and stress expressed in surgeon`s speech can negatively impact teamwork because of emotional contagion (conscious or unconscious induction of emotional states and behavioral attitudes by a person or group to other team members). Emotional contagion of negative emotions and stress has the potential to affect the quality of collaborative work in teams, as emotional and behavioral stress responses influence team performance and patient outcomes. 

While communication in the OR has been analyzed in various studies, no tool currently exists to objectively and automatically measure communication patterns based on video and audio recordings. This project will specifically focus on negative emotions and stress in the communication within the OR team, exploring methods to objectively assess these communication patterns based on speech. 

Approach

In a first step, we will identify reliable acoustic markers of stress and negative emotions in speech by conducting a systematic review on the literature. Additionally, we will conduct temporal analyses of communication sequences between anesthetists and surgeons, based on intra-operative video recordings. Finally, we will conduct real-time analyses on communication patterns in operating room team members based on acoustic markers. These studies will help in facilitating the development of a tool to automatically assess and feedback communication patterns to operating room team members.  

University of Bern and University Hospital

The main workplace will be University Hospital of Bern, department of visceral surgery. The University Hospital of Bern works in close collaboration with the University of Bern. The research group at the Department for Visceral Surgery and Medicine has extensive experience in investigating teamwork in the operating room and is lead by Prof. Guido Beldi.

During the PhD project, there will be collaborations with medical data scientists at IHU Strasbourg, France