It’s a familiar scenario: A group of computer science and AI specialists on one side of the table says “this project is going to change cardiology,” and cardiologists on the other side of the table say “this is totally useless.” This is the struggle AI specialists and medical professionals have navigated for years, and a problem that Chicago-based Northwestern Medicine’s AI clinician trainee program is working to resolve.
“The idea was to put the AI cardiology fellow in the middle, so they could work on meaningful projects together and bring things along,” James Thomas, MD, director of the center for heart valve disease at Northwestern and medical liaison to the AI fellowship, told Becker’s.
This middleman method allows for better collaboration between the two specialty fields.
“This isn’t AI telling medicine what to do, and it’s not medicine telling AI what to do,” Kristian Hammond, PhD, professor of computer science and director of Northwestern’s master’s program in AI, told Becker’s. “It’s what happens when you actually work in partnership, understanding that each field has its own tools, techniques and values, and that creating a generation of practitioners in one field who have the skills of another is the result of that partnership. We’re creating AI skills and ambassadorship — students who are going to do both clinical and research work in cardiology, but with the tools and techniques available today in artificial intelligence mapped onto that.”
The AI clinician trainee program gives select cardiology fellows a master’s in AI and trains them how to be the bridge between medicine and technology. Some fellows enter the program in their third year, doing it alongside their fellowship, while others take a year off to focus on their AI training. The master’s in AI program usually takes about 15 months, but AI clinician trainees complete it in one year and complete the program alongside computer science majors.
The program accepts two to three cardiology fellows each year, many with a technical background in majors such as biomedical engineering, physics and programming. So far, about 12 fellows have graduated from the program and gone on to implement AI projects into clinical workflows. For example, one of the first fellows to graduate in the program developed a technique to diagnose COVID-19 using an EKG, an advancement published in a top radiology journal.
“We’re not doing this because we can find interesting problems in artificial intelligence, we’re doing this because we can solve meaningful problems in cardiology. That’s the goal,” Dr. Hammond said.
Once fellows complete the program, they come out equipped to tackle machine learning and AI problems with the skills of an expert.
“We need people who don’t just pay lip service to it but can get in the trenches, write their own code, and know how to pre-process a million echocardiograms to answer real questions,” Dr. Thomas said. “As time goes on, cardiology is going to have more and more AI algorithms as part of its toolkit.”
At the Becker’s 32nd Annual Meeting: The Business and Operations of ASCs, taking place October 29-31 in Chicago, ASC leaders, surgeons and healthcare executives will explore strategies to drive growth, enhance operational performance, navigate reimbursement challenges and prepare for the future of ambulatory surgery. Apply for complimentary registration now.