AI scribe for doctors tested across Montreal region

MONTREAL – AI Scribe, a time-saving digital tool that helps doctors take notes and summarize their consultations with patients, is being field-tested by physicians across CIUSSS West-Central Montreal for its potential to improve efficiencies and lighten the administrative burden for physicians.

In addition to having AI Scribe tested by 10 physicians, the healthcare network is intent on determining where, how and at what pace the CIUSSS should be widening its implementation of AI to improve access and the quality of care.

To that end, professor Maxime Cohen (pictured), a leading Canadian expert on AI, has been mandated to serve as the CIUSSS’s chief of AI strategy, acting as a consultant on a part-time basis.

With more than 15 years’ experience in the field, Prof. Cohen holds a Ph.D. from the Massachusetts Institute of Technology and is the Scale AI Chair Professor at McGill University.

This past December, he began assisting the network in developing a cohesive AI strategy, setting priorities for more extensive use of AI, identifying promising partnerships with those in the technology sector, and warning of possible risks and pitfalls.

According to Dr. Lawrence Rosenberg, the network’s president and CEO, this makes CIUSSS West-Central Montreal one of the rare healthcare organizations in Canada to closely collaborate with an AI specialist.

“Think of Prof. Cohen as our AI scout,” Dr. Rosenberg says. “He will scan the frontier and propose directions in which we might move. Underlying every avenue of inquiry will be the insistence that any type of AI under consideration must make a direct contribution to upgrading the quality of care.”

AI is not entirely new to the CIUSSS, but so far, its use has been limited. For example, during the COVID-19 pandemic, an AI-based program was fed demographic data about known cases of the illness in the area served by the network.

This enabled the tool to predict the neighbourhoods – and sometimes even individual city blocks – where a surge in cases was likely to occur.

However, professor Cohen said, due to the recent spike in interest in AI tools like ChatGPT, increasing numbers of people – including some in the healthcare sector – have been naturally curious about whether and how these tools might be put to practical use.

Professor Cohen’s short-term goal is to help the CIUSSS focus its efforts on achieving what he calls “quick wins” – that is, the relatively prompt and easy completion of two or three projects of modest size to demonstrate AI’s value and strong, positive impact.

Potentially, AI Scribe is one such quick win. For many (if not most) physicians, making notes and summaries about their consultations with patients is a painful chore. In fact, Dr. Cohen said, it is one of the major contributors to clinicians’ stress and burnout.

To ease this burden, some doctors use a human scribe, who sits in on the consultations and prepares notes and summaries about each patient, which the doctor reviews, approves and places into the medical record.

AI plays the same role, professor Cohen explains, but it finishes the task much more quickly and, often, more accurately. It makes an audio recording of the session with the patient, while carefully safeguarding the patient’s privacy and any sensitive data.

Then, quickly and automatically, the tool generates a summary of the consultation notes, which the physician revises (if necessary) and approves.

In a paper that he co-wrote on the subject last year, professor Cohen concluded: “By automating this task, AI medical scribes can dramatically reduce the time that clinicians spend on documentation, thereby allowing them to reallocate their time to what they do best – namely, seeing and diagnosing patients.”

In addition, he will be helping the CIUSSS to develop clear guidelines and policies on how AI should and should not be used. Since patient confidentiality and the security of patients’ data are central to this effort, he is working closely with the IT and Legal Affairs Directorates.

More broadly, he reports directly to Dr. Rosenberg and is coordinating with Justin Cross, chief digital health officer; Elliott Silverman, associate director of digital health; Dr. Lawrence Rudski, medical director of virtual care; and Erin Cook, director of quality, transformation, evaluation, value, clinical and organizational ethics, and virtual care.

Looking farther down the road, professor Cohen says AI holds enormous potential to streamline and improve quality, efficiency and productivity in various aspects of healthcare. Among them:

• Prediction: As was the case during the pandemic, if AI is fed high-quality data, it can accurately predict a variety of outcomes, such as how many people will arrive in the Emergency Department at specific times, and how many patients will fail to show up for their appointment in specific departments.
• Scheduling: AI can help guide the efficient deployment of various types of employees on specific days.
• Suggesting a diagnosis: Although physicians always have the final word in making a diagnosis, AI can assist them by reviewing the medical data about a patient and proposing a possible diagnosis, complete with explanations. Even if some of these suggestions are rejected by the physician, AI can learn from these experiences how to keep improving the accuracy and reliability of future proposals. Before the consultation takes place, AI can even issue a customized questionnaire to each patient, so that physicians can read AI-generated summaries to prepare for the patients’ visits.

“The important thing,” professor Cohen says, “is to do everything safely and securely, with every possible safeguard in place. This will allow us to proceed according to a carefully structured plan, based on the immediate needs of the CIUSSS’s users of healthcare and social services.

“At the same time, we need to identify AI use-cases with the strongest benefits and the best return on investment. AI is not something we can turn our back on – not if we want to play a prominent part in the next phase of healthcare.”