← Becker’s Healthcare Podcast
Becker’s Healthcare Podcast
Ngan KN MacDonald, Chief of Data Operations at the Institute for Artificial Intelligence in Medicine at Northwestern University
00:07:12
Loading summary
Transcript13 lines
- [00:00]
A
This is Scott Becker with the Becker Healthcare Podcast. I'm thrilled today to be joined by a brilliant leader, Nand McDonald. And Nan's going to talk to us about her system, what she's doing, where she's most focused, and a lot more. Nand, can you take a moment and introduce yourself?
- [00:18]
B
Sure. Hi, Scott. Thanks for having me on. So I serve as the chief of Data Operations for the Institute for AI and Medicine at Northwestern University's Feinberg School of Medicine. And I also have a second role as Director of Data Innovation at Mathematica's Data Innovation Lab. So At Northwestern, I aim is basically an academic institute that we launched in 2020 under Feinberg School of Medicine, with this vision of bridging computational methods and human expertise to advance human health. AI. We think about AI as augmented intelligence, which is designed to enhance clinical decision making and not replace it. And so we're focused on both research and education. We have seven centers around things like collaborative AI in health care, advanced molecular analysis, deep phenotyping, precision therapeutics, bioethics, computational and social sciences, and then medical education and computational imaging and signal analytics. It's a lot, but we essentially draw our leadership from across different disciplines, including medicine, engineering, ethics, data science, and social science.
- [01:30]
A
Talk to me about that. Talk to me. What are the core goals that you're trying to reach at the Feinberg School of Medicine, which is one of the great, great academic medical centers, Northwestern, in the world, in the country, every place. Tell us a little bit about what is it you're trying to accomplish.
- [01:45]
B
Well, our goal is oftentimes what we see in AI is that people are doing it in their little silos. And when we were launched in 2020, it was really the idea of being able to span across disciplines, so across our engineering school, our social science research, as well as our school of medicine. So really it was about creating scalability, because we can't really do that if we don't actually include all of these different disciplines within our research and within our education. We're the first school that, one of the first medical schools that integrated education for our medical students around digital health and data science in all four years of medical medical school. So, you know, we really feel that there is a collaboration that needs to exist for AI to be successful between both industry and academic research.
- [02:41]
A
And talk about that. It was schools, medical schools, integrating AI and data analytics into their programs. How important is that to the future of producing great doctors and great medicine? I mean, talk about that for a moment.
- [02:55]
B
Sure. You know, we started this journey because we got feedback from, from our medical students, our graduates that said, hey, while the medical education is great, what they found was that they weren't quite ready for the day to day of interfacing with electronic health records and all of the other technologies that were part and parcel of day to day healthcare delivery. And so we intentionally started integrating parts and pieces of our curriculum into all four years of medical school to make sure that the future doctors like that they are trained around AI and digital.
- [03:35]
A
Health right from the beginning and so, so important. What are you most excited about and focused on now? Non, when you look at the rest of this year, where are you most focused and excited?
- [03:47]
B
Well, so the things that I'm really excited about are this idea of agenic AI. And you know, we have been experimenting with basics of AI, things like ambient listening, clinical documentation, remote monitoring, but those are all about increasing efficiency and sort of the bread and butter of AI. But when we go beyond that, we can scale AI to be able to do things that allow us to integrate data and interoperability in a more intentional way and automate processes that are beyond just, you know, what you do with personal efficiency.
- [04:29]
A
Thank you. When you look at sort of chief data operations or governance of data and integrating AI in medical school education, what advice would you have for others? What's your sort of starting point of advice you'd give to other leaders trying to do what you're doing?
- [04:46]
B
Yeah, I think the first point is that the data is the fuel for your AI. So if you don't have really good understanding of your data, what the level of data quality is, why you collected that data, where the gaps might be, you're not going to return a great AI system. And so I talk about that a lot. I think I just came out with a blog that talks about that as well. And you know, you can start with all of the training of the models is based on what kind of data you have. So obviously when you don't have good data and it's not representative of your population, you can't apply that model to your system. You have to be able to take that, a model that might have been trained somewhere else and retrain it for your own health systems data. In order to make it work well, you kind of need processes around that whole model deployment and being able to monitor it and then bring it back in and retrain it periodically. So all these things that people talk about, like model drift, like the tendency for us to overfit the model and make it too specific as opposed to kind of really understanding that, you know, oftentimes there's a little bit of human intervention and a little bit of art into AI.
- [06:05]
A
Non, talk to us a little bit about. There's a lot of concern that AI is going to replace humans, replace doctors, replace nurses. Can you give us your thought there? Is it going to be more of a supplement or replacement? A substitute or replacement? What do you see?
- [06:20]
B
I don't think people should see AI as a replacement. I think of AI as augmented intelligence. Right. The human brain is limited to the amount of data it can hold, but it's unlimited in its ability to be creative and imagine something else. So our AI is built upon what we've done in the past, but there's so much more of what we can imagine it being applied towards in the future. So we really need to have those two superpowers harnessed together. AI's ability to understand huge amounts of information and then our human ability to imagine a different solution and a different future.
- [06:58]
A
Thank you so much. What an amazing career you're having. And what you're doing at Northwestern is social critical. So thank you so much for taking time to join us today. Thank you very, very much.