Improving non-clinical tasks using ML-enabled provider workflow


Making the Rounds

Improving non-clinical tasks using ML-enabled provider workflow

Dec 15, 2023

In this episode of Making the Rounds, cardiology fellow at the Texas Heart Institute, Matthew Segar, MD, MS, talks about his research submission for the AMA Research Challenge: “Improving non-clinical tasks using a machine learning-enabled provider workflow.”  View his poster (PDF).

Learn more about the AMA Research Challenge.


  • Matthew Segar, MD, MS, cardiology fellow, the Texas Heart Institute
  • Brendan Murphy, senior news writer, American Medical Association


  • Todd Unger, chief experience officer, American Medical Association

Listen on the go to the full episode on Apple Podcasts, Spotify or anywhere podcasts are available.

Student-focused benefits

Become an AMA member to unlock:

  • FREIDA™ residency dashboard
  • Unlimited access to the JAMA Network™
  • Lenovo: Save up to 30% on the latest tech

Unger: Welcome to Making the Rounds, a podcast by the American Medical Association. Today’s interview features one of the finalists for the AMA Research Challenge, which is the largest national research event of its kind. Join us for the main event in February where the five finalists will present their research and compete for a ten-thousand-dollar grand prize, sponsored by Laurel Road. For the full details, visit Here’s AMA Senior News Writer, Brendan Murphy.

Murphy: Hello and welcome to Making the Rounds, a podcast by the American Medical Association. I'm Brendan Murphy, senior news writer at the AMA. Today, we're continuing our series featuring each of the AMA Research Challenge finalists. The semifinals of the event took place in October, and our top five research projects will compete in the final event, which you can watch on February 6.

Joining us today is Dr. Matthew Segar, a cardiology fellow at the Texas Heart Institute. How are you today, Dr. Segar?

Dr. Segar: I'm doing very well. Thank you sincerely for the opportunity.

Murphy: We appreciate you being here with us. Your submission to the research challenge is entitled, "Utilizing an ML-enabled EMR provider workflow to improve non-clinical tasks." We're excited to learn more about it. Can you tell us about the topic, why it appealed to you and how you got involved in the Research Challenge?

Dr. Segar: Awesome. So I think the overall theme is that as a physician, you're just inundated with the number of non-clinical tasks that are required of you. And these can be a range of things from paperwork to answering messages and message inboxes to just overall forms in general. And so, seeing this throughout my training, even going from a medical student to a resident where you are involved in a primary care clinic, you see a lot of these non-clinical tasks that are required of you and it really takes away from the things that we as physicians enjoy most, and that's taking care and helping patients. So as a fellow, I had a really amazing opportunity to work alongside very smart engineers and computer scientists to come up with a solution to try to help physicians, specifically primary care physicians, with these completion of non-clinical tasks. So we hypothesized that machine learning, some natural language processing and other things that I'm sure we'll get into could help streamline this, and we've been pretty happy and successful with the progress so far.

Murphy: What were some of the challenges you encountered in doing this research?

Dr. Segar: Great question. So I think every step of the process when you're talking about computers can be challenging. The biggest challenge is just getting data from the EMR. You would think it's very easy to just go and get a note, but every single EMR does things differently. There's different permissions. You have to, of course, make sure that things are secure, HIPAA compliant. Understandably, hospitals just don't want to be giving away patient records.

So working with hospitals, working with clinics to try to find avenues and ways to do it securely, effectively—and frankly—easily was the biggest challenge.

Murphy: What were your prior research experiences? How is the research you've done in your graduate medical training different from research you may have conducted as a medical student?

Dr. Segar: Yeah, great question. So I've been very fortunate to have many mentors throughout my educational and medical training. Starting as an undergraduate, I was a computer science major. I was a programmer prior to medical school. I then got my master's in bioinformatics where I programmed software for DNA sequencers, epigenetic software, epigenetic assays. And so that experience really led me to the field that I'm in now, and that's...using technology to help medicine. The things that I'm doing now are a little bit different. I'm not focusing on DNA anymore. I'm focusing more on the macro, big picture, full user environment, which I'm really enjoying. But still the troubleshooting, the programming, the getting down and dirty with the code is still something that I do on a daily basis.

Murphy: What advice do you have for students, residents and fellows who are conducting research on a project like this? And it is a unique project in that it's not necessarily clinical.

Dr. Segar: I think it is clinical. I think you're finding, first off, that you're finding this new wave of physicians that do have a technology background. And if you can use the expertise and intuition and just overall knowledge that you have of the clinic, medical education in general, you can find these opportunities where technology can be leveraged to support your workflow. So I really do think this is a clinical opportunity.

I think the biggest advice is twofold. First, find a good question. Find something you're passionate about. When you're young in your career, it's very easy to find a project that may not be as interesting to you, but maybe because a mentor or someone is doing it that you kind of jump on. But if you can find a project that you really enjoy, it's a topic that you really want to learn more about, you're going to do so much better. And the second is find a good mentor.

Find somebody that's used to working with students, residents, fellows, that has projects and goals that are at your level. And that if publishing or presenting is one of your goals, find somebody that can help support you in doing that. You can sometimes get stuck in projects that take four or five years. And when you're a resident—and for internal medicine specifically, you're only there for three years—it can feel like you're not making as much progress. So finding somebody that can help with your specific goals that understands your circumstances can make a world of a difference.

Murphy: Do you envision research staying a part of your skill set, a part of your career as you move from fellowship to practice? And what does that look like to you, if so?

Dr. Segar: I love that question. I think as a physician scientist, in some ways, I think every physician is a scientist and that you're very used to seeing a question, an unanswered question, and just having this innate sense to want to solve it. So that can be from very basic sciences to health care implementation. But for me specifically, I see my future being more on the health care system side, trying to get opportunities in the physician's hands to implement the day-to-day care of patients. I don't see myself as a very basic scientist in the lab four or five days a week. I want to take care of patients. I want to do procedures. I want to be a physician above all else, but I think there's definitely opportunities to do research, to find opportunities, to enact change, to improve the well-being of people all around us.

Murphy: What were some of the findings you had in this project and how do you apply it?

Dr. Segar: Yeah, great question. So the two big things that we found in our research, our project was that one, physicians typically under code. So you can come up with diagnoses, but it's a completely different language to then try to translate that diagnosis and problem into what we call an ICD-10 code, International Classification of Diseases. In the insurance world, that's what's used for billing, for reimbursement.

And so you can do all this work, but it's crazy. If you don't select that one button of what is that ICD-10 code, you can get your reimbursement cut. You can lose the overall panel of your severity of patients can go down because you're not classifying them appropriately. And so, we found that computers are very good at picking up those nuances to try to recommend to physicians what codes they could be using in their notes. And that second, that this type of work can actually improve patient care because there's very discrete and distinct guidelines that the USPTF Preventative Task Force puts out for us to recommend for certain conditions what we should be doing next. And what computers can help us is that when you have a patient with this condition that have these diagnoses, we can recommend they get preventative and appropriate screening. And so, that was the second big thing that we picked up is that we're actually helping improve the screening of very treatable conditions through our work.

Murphy: And what are the next steps in this work?

Dr. Segar: The biggest next step is expansion. We had a pilot for one site, we've now expanded to two. We're trying to get more involvement to see a different range of patients. We went from Medicare Advantage population, trying to move to see if this works in private insurance and then working with other hospital systems. Like I said, every time you work with a different EMR, that note extraction process is different. So coming up with an even more intuitive note agnostic data extraction process is going to be really important for us.

Murphy: What else would you like to tell us about your journey in medicine?

Dr. Segar: My biggest thing is I actually applied three times to get into medical school. I failed the first time, failed the second time. It wasn't until the third time that I got in two weeks before classes started when I'd already committed to a completely different track, a PhD program. But I wanted to become a physician. I was not going to settle for anything less. So I'm very happy of how my medical journey's gone. And I think that perseverance has really paid off.

Murphy: Well, that's certainly inspiring. Another potential source of inspiration is the $10,000 prize that comes with the AMA Research Challenge. What if you won that money? What would you do with it?

Dr. Segar: I would love to say that I would spend it on something fun and noteworthy, but I think with everybody that's a finalist, loans are definitely something that is on the back of all of our heads. So definitely going a little bit towards that. But I think there's things that I could definitely find to buy, whether it's some computer upgrades or maybe some more hard drive space with all the data that we're collecting, need somewhere else to store it or cloud server costs, there's always technology costs that pop up here and there.

Murphy: Reinvesting in the work. Now that is commitment. Well, Dr. Segar, thank you so much for sharing your research with our listeners.

Dr. Segar: Exactly. Thank you again for the opportunity. It's been very fun.

Murphy: And to find out more about Dr. Segar's research and the other finalists' research, tune into the AMA Research Challenge on February 6. You can to get more information on the contest. You can also find a link to that page in the description of this podcast. This has been Making the Rounds. I'm AMA Senior News Writer, Brendan Murphy. Thanks for listening.

Unger: Don’t forget to tune in to the AMA Research Challenge finals on February 6. Get the details at Subscribe to Making the Rounds today.

Disclaimer: The viewpoints expressed in this podcast are those of the participants and/or do not necessarily reflect the views and policies of the AMA.