The challenge
London Health Sciences Centre (LHSC) is one of Canada’s largest acute-care teaching hospitals, operating the Michael Gunning Simulation Centre where medical professionals train through realistic clinical scenarios. Simulation-based training is critical for developing clinical skills, teamwork, and communication under pressure.
A key area of research at the Simulation Centre focuses on communication breakdowns during medical scenarios. These breakdowns can lead to adverse patient outcomes in real clinical settings, making their identification and analysis essential for improving training protocols. However, analyzing simulation recordings manually is extraordinarily time-consuming. Researchers had to watch entire recordings, often hours long, to find specific communication moments.
- Lengthy recordings: Medical simulations generate hours of audio and video content. Manually reviewing these recordings to find specific communication events consumed enormous research time.
- Communication breakdown identification: Researchers needed to identify specific moments where communication failed, which required careful review of every interaction in a simulation.
- Cost and efficiency: Traditional manual transcription and review was expensive and slow, limiting the volume of simulations that could be analyzed for research purposes.