What’s next:
Healthcare

×

5 tips for rolling out successful healthcare speech recognition projects

“Speech recognition should be considered a project in and of its own right not just about adding a usability feature. It’s a disruptive technology which must be deployed as a full-blown project,” insists Olivier Boussekey, IT manager at Saint-Joseph Hospital, Paris who rolled-out a successful paperless healthcare project. Based on the feedback from this and other customer and partner projects we have identified five tips you should follow to achieve your own successful speech recognition deployment.

By
hands and fingers

 

1. Map out and engage ALL your stakeholders including your medical secretaries and back-office support team

    Speech recognition rollouts are usually part of a broader healthcare efficiency strategy which will involve new service models and ways of working. Historic workflows involving clinicians, medical secretaries and other back-office resources should all be considered and reviewed when planning the new workflows.  The early engagement of all your support and back-office resources are essential. Why? Because the devil is often in the detail. Medical secretary and other administration resources will have detailed knowledge of processes, links to interfacing organisations, current snags and great ideas about how to overcome them. Also, they are often the personal interface with the patient.  Their positive involvement can be the difference between project success and failure.
     

    2. Put yourself in your clinicians’ shoes

      The relationship between clinicians and IT departments is not always easy. Clinicians criticise IT for not understanding their daily lives and vice versa. Several practical examples have revealed the benefits of the IT team immersing themselves in the daily life of the clinicians. At Saint-Joseph Hospital in Paris, actively listening to the needs of medical practitioners was a priority for the IT team who spent time in each department, observing the uses of speech recognition in medical practice. They were able to detect problems that the project team had not previously raised: clinicians were accustomed to leaving an audio or paper note for the secretary advising them of a specific action to take. With front-end speech recognition clinicians weren’t able to do this in the same way. Instead, the IT team created a “Post-it” in the EPR that delivered the same result.
      Get down onto the ‘shop floor’, get close to your clinicians and put yourself in their shoes. Technology should serve the clinicians – not the other way around.
       

      3. Training, training, training

        At the first level, speech recognition training is quick and easy. At St Joseph it took less than 4 weeks for 100 clinicians to be operational. However, for clinicians to maximise the full potential and power of speech recognition in their day-to-day work they must master a multitude of standard and customised features and commands. For example, for a medical note or report, rather than dictating and correcting as you go along it is preferable to dictate the entire text first and then correct any errors by voice once that segment of dictation has ended. In this way the software learns from its errors. Without correction, just like any individual, the software will repeat its error.
        Training doesn’t have to take place in a classroom environment. If the basics are in place then innovative ways of delivering training on-the-job, as at Alderhey Children’s Hospital, can be efficient and effective.
         

        4. Support, support, support

          Deploying a dedicated support team to help clinicians in real-time is a common feature of successful projects. In all healthcare organisations that have successfully completed their projects, the IT team have set up a system to quickly respond to clinicians’ requests, identify bottlenecks and speedily resolve any user problems. At Saint-Joseph Hospital the IT Director was part of the support team and included himself on the ‘hotline’ for clinicians to report difficulties or request support. direct to his mobile phone. Finding and taking ownership of a solution always requires supervision, even when the solution is easy to use.
           

          5. The end of the beginning

            Speech recognition is not a secret sauce although the new generation of technology based on deep learning and neural networks is truly amazing in terms of its immediacy, precision and quality. Once your clinicians are using speech recognition on a daily basis and are beginning to understand and feel the benefits it is time to refine the use of automatic templates and voice commands customised to their workflows and specialties. At CHU Nantes, the project team organise regular meetings with clinicians in order to share experiences and answer outstanding questions. At St-Joseph Hospital in Paris they distribute newsletters to share tips and tricks from colleagues.

            Each project is unique but in our experience these 5 simple tips hold true for all successful healthcare speech recognition projects.

            Read full article

            More from the editor

            AI tops Gartner list of 2018 tech predictions
            The RCGP says GPs spend too much time ticking boxes and completing documentation
            Be a part of our research with PracticeBusiness
            Where there’s a challenge, there’s often a solution
            Make it quick and easy for clinicians to capture the patient story
            Top tips for implementing speech recognition in busy pathology departments
            Because the ‘patient powered decade’ is almost here
            The Digital Health CCIO Summer School shows the role of data in the transformation journey
            Data output and analysis is big up the to do list of healthcare.
            panorama and sunset
            AI tops Gartner list of 2018 tech predictions
            patient and doctor smiling at a patient record
            Where there’s a challenge, there’s often a solution
            Speech recognition can ensure clinical data is recorded accurately and in a timely manner
            Because the ‘patient powered decade’ is almost here
            Male nurse and boy using stethoscope
            The RCGP says GPs spend too much time ticking boxes and completing documentation
            NHS Digital must do its homework on clinical documentation
            Make it quick and easy for clinicians to capture the patient story
            Digitising the NHS and role of data
            The Digital Health CCIO Summer School shows the role of data in the transformation journey
            Be part of our research into the digital patient
            Be a part of our research with PracticeBusiness
            Top tips for implementing speech recognition in busy pathology departments
            Data output and analysis is big up the to do list of healthcare.
            Show more articles