How a Dragon Medical One super-user improved efficiency at Guy's and St Thomas'

In the fast-paced environment of modern healthcare, efficiency and accuracy in clinical documentation is crucial. This case study explores how the Musculoskeletal and Trauma and Orthopaedics department at Guy's and St Thomas' NHS Foundation Trust (GSTT) is optimising clinical workflows with Dragon Medical One (Dragon), an AI-powered speech recognition solution, hosted on Microsoft Azure.

GSTT is one of the UK's largest NHS Foundation Trusts with approximately 23,000 staff. In October 2023, a new electronic patient record (EPR) powered by Epic was launched across GSTT in collaboration with King’s College Hospital NHS Foundation Trust (KCH). This joint go-live was one of the biggest ever single go-live of Epic’s EPR anywhere in the world.

Replacing multiple legacy systems with a single EPR is a significant change management challenge. To minimise the impact on the clinical workforce, both hospital trusts decided to launch Dragon before introducing the Epic EPR. This strategy helped familiarise staff with speech-enabled workflows and boosted confidence in the reliability of speech recognition technology ahead of their Epic go-live, while also helping to reduce administrative burden.

Dragon integrates with Epic, enabling users to dictate clinical notes and documents directly into the EPR as well as navigating around Epic using voice commands. It can also be embedded within Epic Haiku, Canto, and Rover mobile apps.

Thousands of doctors, nurses and allied health professionals are now using Dragon Medical across a wide range of clinical settings in GSTT and KCH. This study delves into the journey and impact of speech recognition within Musculoskeletal physiotherapy and Trauma and Orthopaedics.

The journey of a Dragon super-user

Ryan Strother is a Musculoskeletal Physiotherapist and Fracture Clinic Physiotherapy Team Lead at St Thomas’ Hospital. His role involves managing a high volume of patients in both the Fracture Clinic and the Musculoskeletal (MSK) Clinic. The Fracture Clinic handles same-day walk-in appointments with up to 20 patients in the morning, while the MSK Clinic manages 30–40-minute appointments with up to 17 patients per day.

Ryan’s journey with Dragon began in May 2023, with his initial focus just on dictating clinic notes and letters to help alleviate the burden of typing. After the Epic EPR went live in October 2023, Ryan sat with a Dragon trainer for 45-minutes and learnt about the advanced features such as step-by-step voice commands.  

Using Dragon voice commands along with Epic SmartTools can help simplify workflow and make administrative processes more efficient. Dragon has the potential to speed up documentation and has supported Ryan to take more detailed notes, allowing him to better represent his patient’s narrative.

Ryan advised, “Understanding workflow is critical—any solutions developed for clinicians must be purpose-built to integrate with the way they work. That’s why co-design with clinicians is essential; involving them from the outset ensures the technology truly supports clinical practice.”

Using voice commands to enhance workflow

Ryan described how Dragon voice commands has streamlined large portions of his typical MSK appointment workflow, reducing much of the typing and clicking usually required.

Ryan measured that the old appointment workflow took him approximately two minutes to complete, excluding notes, and 30 mouse clicks. In comparison, the new Dragon workflow he created takes only 20-25 seconds and replaces the 30 clicks with just four voice commands. (The time savings from dictating clinical notes instead of typing were not included in these measurements as it varies significantly from patient to patient.)

Figure 1. Example of MSK appointment workflow: Replacing ~30 mouse clicks with 4 voice commands.

That original appointment workflow in Epic also required clinicians to know which clinic codes to use, for example when doing orders for appointments. With the new Dragon enabled workflow, no prior knowledge of these codes is required as the orders are prebuilt on EPIC. This helps reduce the risk of patients being accidentally booked onto the wrong code.

The second compelling example Ryan shared was a streamlined MSK vetting process he created. His department typically receives 2,500-3,000 referrals per month from primary and secondary care and each one must be vetted by clinicians. There are 75 different vetting decisions so it’s quite a complex process for the clinicians.

Ryan explained that following the clinical decision, vetting a referral used to take 90 seconds on average, involving approximately 25 clicks and typing. With the process he developed using Dragon, he has reduced the time to only 17 seconds with just two voice commands.

Figure 2. Example of MSK referral vetting workflow: Replacing ~25 mouse clicks with 2 voice commands.

As a Dragon superuser, Ryan led a quality improvement project that involved sharing the commands he created with the rest of the department alongside staff training and dissemination of user support materials. The project has achieved a 90% adoption rate* across the department and has led to an estimated time saving of 60 hours of clinician time per month (based on 2800 referrals).

Streamlining workflow beyond the EPR

These examples illustrate how advanced speech recognition solutions can improve clinical workflows, extending beyond just dictating clinical notes. It also highlights the transformational role of superusers and how effective training, and continuous support, can help ensure the technology is used to its full potential.

The first step is to understand the basic manual processes or clinical workflows as defined by the EPR system and the NHS Trust. The aim is not necessarily to change these workflows—correct procedures need to be followed—but to streamline those workflow steps using voice commands and automated text, templates, orders, and so on. This approach not only saves time but can also improve documentation quality and reduce the likelihood of errors.

Dragon's speech recognition technology can be applied beyond clinical workflows in the EPR, including tasks like drafting emails, navigating Microsoft Office documents, and conducting web searches. Users can create customised voice commands to streamline their routines. For example, when Ryan logs on to his computer, he must open the same applications, files, and spreadsheets each time. He has created a single voice command called ‘morning routine’ which opens everything all at once. This typically saves him 1.5-2 minutes every morning. He has created another voice command to close everything down for him too.