Recently I wrote about a fascinating project that was monitoring the calls made to emergency dispatch teams for signs of cardiac arrest. The company, known as Corti, have developed an AI assistant that will analyze calls to ambulance dispatchers in Denmark to listen out for clues that point to the caller suffering a heart attack.
Whilst such voice analyses are far from commonplace, there are signs that they are becoming more widespread. The latest example comes via Nuance Communications, who have recently announced a report documenting their work with the Emergency Department (ED) at South Tees Hospitals NHS Foundation Trust.
The paper highlights how clinicians felt that speech recognition technology was 40% faster than writing clinical documentation by hand. When this saving was extrapolated over the course of a full year, it would equate to gaining nearly two full-time clinical staff.
Streamlining note taking
Medical notes are a huge part of any health professionals job, with estimates that nearly 50% of a doctor’s time is spent documenting clinical processes. What’s more, a further 52 minutes is spent searching for information that either isn’t captured in the record or isn’t easily searchable.
I suspect few would doubt the potential time savings of having an automated voice annotation system for medical notes, but question marks have lingered over the accuracy of such systems. If time is required to fact check for accuracy of annotations, it defeats the very point of the systems.
The research saw clinicians at the hospital use Nuance Dragon Medical speech recognition software to speed up the recording of notes into the electronic patient record system used in the ED. The average time saving was around 3 minutes per patient, with the clinicians happy with the quality of the annotation. Some 86% of the clinicians using the software thought it resulted in more comprehensive notes being taken, with 90% saying that notes improved in quality compared to when they were hand written.
“Speech recognition has transformed our ED, releasing our doctors and nurses from the shackles of clinical documentation and enabling them to spend more time treating patients,” the hospital say.
Other findings from the report include:
- Speech recognition is now the clinicians’ preferred means of capturing clinical documentation into the EPR across ED
- All clinicians reported that the speech recognition deployment in ED had a positive impact
- Most (67%) of clinicians believed patients had no concerns with their use of speech recognition