By: Justine Brooks
16 Nov, 2024
Doctors spend countless hours on administrative tasks such as note-taking and documentation — time that could be spent seeing more patients and providing care. Canada CIFAR AI Chair at Amii and Professor at the University of Alberta Ross Mitchell is part of a team developing a way to ease the burden of some of these tasks for physicians. Their work could improve healthcare delivery, patient satisfaction and physicians’ quality of life.
In a Red Deer, Alberta emergency department, Mitchell and team have deployed a case study of his AI Scribe project, a large language model that automates the process of creating medical notes from healthcare workers’ conversations with patients. A simple interface that can be run on a phone or computer records conversations between doctors and patients, generating transcriptions, summaries, suggestions and more depending on the physician’s needs and specifications. “The doctor will ask the patient for consent to record the conversation to help with note-taking and documentation processes,” says Mitchell. “What we’re hearing from patients is that their experience is better, because their doctors are giving them better face-to-face interactions and eye contact, instead of looking at computer screens. And what we’re hearing from some of the doctors is that they’ll never go back to doing manual notes again.”
System-wide, AI Scribe also shows great promise at improving hospital efficiency. The time saved using this process has already shown a remarkable impact on the number of patients seen at the demo site — approximately three more patients seen per physician per shift. And unlike similar applications, AI Scribe keeps all patient data locally secure. “That means that all the data stays within Alberta; not a single byte travels outside the province. It’s the same level of security that the banking and the military sectors use for their data,” says Mitchell.
With the simplicity and low-cost of use, Mitchell says that the AI Scribe project could be deployed virtually anywhere with access to cloud computing, including in developing nations. The team hopes to launch the free, open-source project more widely by the end of the year with an accompanying paper describing its function and potential uses. They plan to conduct a larger study following its release to more formally evaluate these early observations and their impact on the quality of care and efficiencies in the healthcare system.
Mitchell notes the importance CIFAR funding has had on his work.
“The Canada CIFAR AI Chair offers crucial financial backing for my lab’s personnel and connects me with an expansive network of outstanding scientists and collaborators, including the other Chairs. Additionally, CIFAR’s funding of the Alberta Machine Intelligence Institute has fostered a robust ecosystem supporting AI research in healthcare.”
Ross Mitchell, Canada CIFAR AI Chair, Amii