Highlights | Creating the future of medical data science
- The Institute for Medical Data Science was developed to implement artificial intelligence (AI) and machine learning into healthcare and to study it.
- Medical data can be harnessed to help the care team make clinical decisions faster and easier and even improve clinical outcomes.
- The Institute has a unique opportunity to study operational models in practice to improve patient health.
Three years ago, an organization-wide committee convened to discuss the future of medical data science at UW Medicine. The committee drafted a proposal to create an institute to implement and study artificial intelligence (AI) and data science in healthcare. Just last year, that proposal was approved by the University.
Enter: The Institute for Medical Data Science.
Overseen by the UW School of Medicine, College of Engineering and School of Public Health, this collaborative and cross-functional team aims to lead the way in medical data research and implementation into patient care.
“The formation of this new institute puts us in a position to be world leaders in innovations for leveraging data, putting AI into practice, and providing better healthcare outcomes,” says Sean Mooney, PhD, interim director of the Institute for Medical Data Science.
Mooney has spent his career managing the development of collaborative electronic systems supporting biomedical research. And at UW Medicine, he’s dedicated his time in the field to teaching and research — he’s a professor in the Department of Biomedical Informatics and Medical Education in the School of Medicine, the chief research information officer of UW Medicine, and the faculty director of Biomedical Informatics at the Institute of Translational Health Sciences.
This program has been a goal and dream for Mooney since he started working at UW Medicine in early 2015. Now that it’s finally here, he and the other members of the Institute’s executive team — Shwetak Patel, PhD; Jonathan Liu, PhD; Lurdes Inoue, PhD; Dushyant Sahani, MD; and Peter Tarczy-Hornoch, MD — are excited about the future.
What is artificial intelligence?
If you own a smartphone, you are using AI daily — most of us just don’t think about it. The notifications we get on our phones and the advertisements we see are connected to our search preferences and how we use our apps; that’s all generated by AI.
“Machine learning is good at synthesizing a lot of data and detecting patterns that could otherwise go unnoticed,” says Mooney.
Technology and business fields have been using AI methods for decades — and healthcare is just starting to catch up.
“Healthcare systems generate enormous amounts of data — and this data is being largely underutilized,” says Mooney. “Simple AI tools have the potential to help make clinicians’ jobs easier and more effective.”
For example, when a patient comes into the emergency room, we conduct different tests, take vitals and do imagining, which is a lot of medical data to review in an urgent patient situation. And that is exactly what AI is good at — quickly sorting through a lot of data.
Machine learning programs could help utilize the data to provide clinicians with the tools they need to deliver care better and faster, says Mooney. With AI monitoring the data, it can flag a serious problem earlier, allowing the care team to react faster. That’s the future of AI.
Pairing research and action for better patient outcomes
The pipeline of publishing research to enact change is a slow process in healthcare. And, according to Mooney, it leaves out what’s essential: the doctor and patient experience. The only way to get that is by putting the research into action, not just in a journal.
“It’s not enough to publish a paper based on data and say, ‘This method is accurate,’” he says. “We want to implement the methods and find out how healthcare changes. How do clinical workflows change? How do providers react? How do patients react to that method being implemented?”
And the Institute was built to do just that. From developing new methods to studying the downstream effects of implemented AI and data science, the Institute’s aim is to have the right mix of research and action.
In guiding the organization to use new AI technologies, the Institute is dedicated to investigating new data types that improve methods and outcomes and to improving patient health and provider satisfaction regionally.
It’s a lofty list of goals, but it’s one that Mooney and the medical data science community at UW Medicine and UW are passionate about achieving.
Equity in AI learning
As the teams at the Institute start their work around care delivery transformation, discovering technology innovations, building business and healthcare operations, understanding the societal impacts of AI, and advancing diagnostics and education — they recognize the importance of leading with an equity lens.
“Data can be biased in many ways, and we want to make sure that, as we start implementing machine learning into patient care, it doesn’t increase potential disparities and inequity,” says Mooney. “And that’s something we want to understand better so we can develop tools or processes that might improve care outcomes for all patients.”
Normalizing medical data science
“I hope AI methodology will be seamless and we don’t notice it being implemented,” says Mooney. “This field continues to grow, and we have so much technology in our area and the medical and academic institutions to help enable these new methods to be developed.”
He acknowledges that thinking about machine-run intelligence can make people nervous. One reason is that they often don’t understand what it really means in practice. This is one of the roles of the Institute: to train the next generation of data scientists and quantitative researchers and help make the community aware of why AI is valuable and how it improves patient care.
If you are interested in learning more about AI in healthcare, the Institute hosts monthly and yearly programs for employees that bring together thought leaders, faculty, students, and industry stakeholders in the field.
- 2022-23 Medical Data Science Seminar: Held virtually on the first and third Monday of the month from 1-2:30 p.m.
- Medical Data Science Journal Club: Every two weeks on Wednesdays, 1-2:30 p.m., led by Aaron Lee, MD.
- Medical Data Science Symposium: Feb. 27 and 28, 2023, at the UW Hub. The first day includes an introduction to medical data science. Tickets still available, register today.
- Request to join the mailing list: email@example.com.