Join our daily and weekly newsletters to obtain the latest updates and exclusive content on the coverage of the industry leader. Get more information
Clinical and patient satisfaction (or lack of it) is an important problem in medical care: patients because they can have difficulty getting access to care, doctors because they simply have too much to do.
The Ottawa Hospital (TOH) set out to address this challenge with the integration of Microsoft’s co -pilot Dax last year. And it has already had a great impact: the early results show seven minutes saved by meeting, a 70% reduction in the exhaustion and fatigue informed by the clinician, and 93% of the patient’s report to an equivalent care experience.
“Access to care is likely one of the biggest problems that patients face,” said Glen Kearns, EVP and Cio or Toh, to Venturebeat. “If we can improve performance, only a couple of patients per doctor per turn, multiply it by 10 doctors in an environmental environment, then multiply that by 365, that is not an inconsistive increase in access to care.”
AIENT AI as an active assistant
Toh was the first Canadian hospital in the Microsoft pilot co -pilot, which is directly integrated with the EPIC EPIC platform for electronic health registration widely used (EHR). In March, Microsoft included Dax Copilot with Dragon Medical One (DMO) in the assistant of Integrated Microsoft Dragon Copilot, which is the technological giant he says is used by more than 600,000 doctors.
Dax Copilot, who is out of the box, captures medical patient conversations through a mobile application and generates a real -time clinical notes, Kennel Harper, Microsoft project manager for Dragon, explained to Venturebeat.
“Doctors can start a recording from a mobile phone, leave their phone, examine the patient, talk to the patient, not different from what they do anyway,” he said.
Then, the system extracts details based on the context of the visit (symptoms, diagnosis, treatment plans, follow -up), which is immediately available in the EHR; All that the clinician has to do is review and finish quickly.
“Instead of having to authorize something from scratch and remember all thorough details or write in front of a patient, this is only functioning automatic,” said Harper, noting that they remember a first “precise” first draft. The first draft of the visit is over.
To adjust Dragon Copilot, Microsoft uses a “massive repository of clinical data that have been selected over the years,” Harper explained. Engineers continually refine large language models (LLM) with that data so that the system can reliably understand and summarize medical information.
To further improve precision, models are optimized by specialty, be it an emergency doctor, a dermatologist, a cardiologist or other medical professional.
In the feedback loops, the teams analyze the first draft prepared by the machine and compare it to what and how much the doctor changed later.
“It ensures that, as an average time, as the data continues and we learn from that data, the edition decreases,” said Harper.
How toh is improving visits, reducing doctor’s depletion
According to the Canadian Medical Association, doctors spend about 10 hours per week in administrative tasks, such as the update of the graphics after the patient’s appointments.
Kearns explained that Toh developed a robust evaluation plan for Dax Copilot that includes monthly updates through a Microsoft Power BI board. This incorporates comments from doctors, patient surveys and EPIC data.
“This framework helps us continuously monitor the impact and guide improvements,” he said.
In general, at the end of the day or the turn, doctors must return and then the final documentation of visits to the patient, he explained. But the tool has reduced schedules, graphics and documentation work for “all categories of doctors.” This not only saves them time, but it helps reduce exhaustion because they have less tedious job to do.
The tool has also improved the visits of the cognitive load of the clinician: instead of focusing on the patient’s innocetes and navigating in documents and forms, they can “participate differently and better,” Kearns said. In addition, “we saw a higher level of performance, more patients per shift, by doctor.”
Kearns emphasized that all patients who consent before appointments are registered and have access to the notes on their Mychart patient portal. They are also provided with literature on the program and ensure that their health records are always confidential and safe.
The patient’s receptive also has a “very, very positive” leg, with 97% informing that their experience with the AI tool was as good as a typical or better appointment. “We attribute that to the doctor’s opportunity to commit to them differently and intentionally from the visit, which is compressed by time,” Kearns said.
Looking towards the future, he said that the tool could be used in scenarios such as the detection of biomarkers and the social determinants of health (non -means problems that affect the health of a person, such as nourement or lack or transport, prior authorization in the future is to make dragon co -pilot triggers actions after vision.
“There are many things on the way in the documentation capture space,” Kearns said.
Digital teammates to overcome personnel problems
Another area where toh is incorporating the AI is “fellow digital team”. Last summer, the installation developed use cases with Deloitte and launched Sophie, which speaks several languages.
Interestingly, you can interpret the feeling of the patient and the behavioral response.
“I hate saying this, but patients lie to doctors,” Kearns acknowledged. “You know,” what is your pain scale? “” Oh, I’m doing well, it’s five of 10. “But then Sophie will have the ability to look at your face and say: ‘Well, it doesn’t seem like it was a five.”
She can ask what is the definition of five of the patient and can pivot according to the objective data.
Toh will launch another avatar in spring to help patients navigate and access the health system and participate in prevention.
“I don’t think anyone does not know the crisis of human resources of medical care that exists worldwide,” Kearns said. “We want to try to support and serve patients more robust than we can today.”
For example, the installation performs monitoring patients after certain procedures. However, due to the limitations of resources, they can only follow up with patients with the highest risk. An objective for Kearns is to track each patient, with the avatar that provides clarification of clarification about any patient’s question and confirm if they understood their defaluation orders, the pharmacy could be reached or has been in a bone. The avatar can increase a nurse or the patient’s clinical team if necessary.
“One of the things in which medical care is proud is the human touch,” Kearns said. “This is a way to make sure we are maximizing and optimizing those human tactile resources, but also making sure that patients receive good support through their medical care trip.”
However, he pointed out that it is still early in the game. An important future step is to allow digital teammate to interact with information and epic environment.
“We have much more work to do there, we are still very focused on the side of adoption,” Kearns said. “We are still a health system that is very reactive to the health status of the patients. We would like to reach the place where we are proactive.”
