Improving voice recognition technology in medical laboratory services

November 16, 2016

Dr. Jill Wooff

Earlier this year, the health system piloted Fluency Direct voice recognition technology to medical laboratory services personnel in the Prince Albert Parkland and Regina Qu’Appelle health regions. Using Fluency Direct, pathologists, pathology attendants, and medical laboratory technologists dictate their words, see them appear on-screen, edit their dictations, and release reports in real time.

However, since the technology was introduced, users in both regions have reported a number of technical and workflow-related issues that are causing systems not to perform as expected. The project team is right now working with technology vendors, healthcare practitioners, and health system leaders to resolve these issues together.

In late October, M*Modal and 3sHealth worked with pathologists in RQHR to discuss their successes and challenges using Fluency Direct in conjunction with Soft Path, the current laboratory information services system. Four key issues emerged from those discussions:

  • A specific voice command was not working properly. This issue will soon be resolved through a new version release.
  • Microphone settings were contributing to misrecognized speech for some pathologists. This issue will also be addressed in the new version release.
  • Some pathologists did not want background noise filters to be enabled. This feature was disabled for a number of pathologists and for others it was enabled.
  • Integration issues between Soft Path and Fluency Direct were causing formatting issues in reports. The Soft Path vendor (SCC) and M*Modal met on Nov. 10 to research the issues and start to work out a resolution.

Other steps being taken include:

  • Roughly 1,700 words will be added to the Fluency Direct dictionary by Nov. 18 to aid in better dictations. A process for adding pathology terms is also being worked out.
  • We continue to explore with our partners the need for an editor role to address workflow issues around the creation and dissemination of reports.
  • 3sHealth has also been working with the laboratories to value stream map the current state to better understand the impact that voice recognition technology will have on operations.
  • We have begun surveying stakeholders weekly to determine their level of satisfaction with the transformation and better engage the community of users. User feedback is critical to improving the provincial transcription service, and the team deeply appreciates stakeholders’ candour and engagement.

Our goal is to continue to work together with partners and vendors to find solutions that work for care providers and patients.

No Very

Captcha Image