More and more health professionals are using artificial intelligence (AI) for the transcription of medical notes during consultations with patients, and Health Quebec intends to grasp the ball. It plans a pilot project which should be launched in 2026 in order to deploy the AI large -scale for medical transcription needs.
Work is underway to assess the solutions available for taking clinical and non -clinical note, Santé Québec said in an email transmitted to the Canadian press.
Currently, only the solutions that have obtained an official Quebec health certification, among other things to guarantee the protection of patient data, can be used in the health network. Plume IA and Coeurway are among these Quebec companies, and they are attracting more and more doctors with their application.
For the moment, Plume IA is mainly dealing with family medicine groups (GMF). The company is currently in the process for the application to be tested by speech therapists, social workers, doctors and nurses “to demonstrate that it works” and possibly that it can have “a license acquisition for a whole department, for a whole hospital”, indicates the co -founder of Plume IA, Dr. James Tu.
He launched Plume IA with Dr. Jasmin Landry barely a year ago, and already, around 10 % of Quebec doctors use their application, which represents around 2,000 doctors.
“The comments we have are really very encouraging. Doctors call us, they write to us, they say that we have changed their practice, they come home much less tired, they have more time with their family. We have anecdotes of family physicians who pushed their retirement this year because they found that the heaviness had decreased, then that it brought a little pleasure in terms of work, “said Dr. You.
The latter is an urgentologist. In his practice, he believes that he is able to see 4 to 6 more patients in a quarter of eight hours.
The time earned varies depending on the practice, but in general, health professionals save one to two hours of file writing per day.
“The application is accessible and walks for all specialties,” says Dr. Tu. […] Those who have a lot of added value using our application are those who have to do a lot of writing. I am thinking in particular of social workers who must write notes on the patient’s demographic data, on the psychosocial aspect. ”
Reread, validate, insert into the file
The Plume IA application has two modes, the first is to record the discussion during consultation with a patient (after having obtained his consent), then the application transforms this discussion into one our medical and structured, a step that clinicians are used to making by themselves.
The doctor can also consult the consultation normally, then at the end, register speaking with the application, which will produce the same kind of structured medical note.
According to Dr. Tu, even if the patient has a pronounced accent or the latter uses certain sentences in another language, the application will be able to produce a reliable medical note.
“This is the strength of artificial intelligence is that it is not a transcription to the word for word necessarily. It is the artificial intelligence that interprets the conversation, […] Who is able to deduce a little more was what the context, ”underlines Dr. Tu.
He recognizes that occasionally, certain errors can still slip into the note, and that the vigilance of clinicians is required. “I think it is a habit that everyone has instinctively to reread the note, and to correct the shells if there is or to add certain information which is more visual or that is implicit. Then, they can correct the note directly in the application. It takes a few seconds and after, they can immediately validate it and insert it in their file, ”explains Dr. You.
For the moment, Plume IA focuses only on transcription, but its co -founder has confidence that technology will one day allow to advance the accuracy and speed of diagnostics.
“The potential is almost infinite and I think it will happen actually,” he says. The Doctor nuances, however, that important issues must be discussed, in particular all ethical questions, the accountability and the diversity of the provenance of the data.
