Next-gen Imaging Workflow and Speech Recognition
Laurens de Groot, Radboudumc, Nijmegen
The Radboud UMC started a two year program back in 2012 to introduce a new electronic health record system from the US vendor Epic. From the beginning, it was clear that end-users would face huge changes in both their habits, workflow and possibilities. It was also an opportunity; there was momentum to realize an integral image management system and introduce the hospital-wide availability of speech recognition.
Next-gen Imaging Worklfow
The search for an enterprise imaging solution that fits the vision, strategy and policy of the
hospital, led to the following requirements:
• safe and accountable patient care,
• durable (effective, cost-efficient and re-usable),
• vendor-neutral (connect with other vendors),
• interoperable (system agnostic),
• best-of-suite (prefer integrated solution over specificity),
• using international standards (DICOM and HL7),
• ready within 9 months.
The architecture has a closed order-loop system, which requires an order for every image-object stored on the platform (effecting procedures and day-to-day workflow). The platform delivers one central viewer for all imaging results coming from several consolidated departmental systems. The outcome is an enterprise wide imaging platform that allows for central storage and accessibility of every type of image (independent of specialty) – thus reducing turnaround times, increasing availability and therefore vastly improving the quality of healthcare.
Having a new imaging platform also invited for new speech recognition solutions. The following solutions were designed based on functional user requirements;
• Speech recognition for Radiology, integrated in the Radiant module of Epic EHR,
• Speech recognition for Pathology, integrated in the VMWare Workplace 2.0 environment,
• Speech recognition for other specialties, independent of integration or software program.
The project included topics like workflow change management, system integration with Epic and VMWare, and principal discussions about the benefits of optimizing user effectiveness by combining different techniques.
With a recognition rate of over 97%, physicians are experiencing faster and more accurate reporting, thus having more time left for seeing the patient and in the meantime delivering safer and higher quality results.
Laurens de Groot MSc PDEng
During his training in Clinical Informatics, Laurens has worked at the Radboud University Medical Center in Nijmegen. He was supervised by dr. Michiel Sprenger (TU/e), dr. Eric Visser and dr.ir. KlaasJan Renema (both Radboudumc). After the training Laurens stayed at Radboudumc as project manager for the Regional Healthcare Information Exchange project.