How AI Can Support the Medical Community

Dr. Hawkins Gay used to work 100 hours per week as an investment banking analyst. He was fresh out of college and liked working hard and feeling productive, but he wasn't motivated by the work. He didn't feel fulfilled. That realization sent him on a path toward a medical career, and today he is a cardiology fellow at Northwestern Memorial Hospital.  

Northwestern Memorial Hospital is also where Dr. Alex Leidner completed his residency and later served as chief nephrology fellow. He and Dr. Gay have more in common than working at the same hospital — both are students in Northwestern Engineering's Master of Science in Artificial Intelligence (MSAI) program. 

Dr. Leidner and Dr. Gay both see the powerful potential for AI and machine learning in the medical community. They recently took time to talk about that, and their goals as students in MSAI.

"AI can provide doctors with augmented decision making to help identify patients that could use new treatments and help monitor those patients' side effects appropriately." — Dr. Alex Leidner

How can the medical community benefit from machine learning?

Dr. Leidner: Cardiovascular disease and hypertension are referred to as silent killers. Hypertension does not result in negative outcomes for everyone, but it remains a leading risk in developing heart disease. Using machine learning, we can identify latent characteristics to make it clear whose hypertension will cause issues later on. It can also identify those people with subtle abnormalities and assist public health attempts to identify and treat these diseases.

Dr. Gay: We are living through an exciting period of advanced computing with artificial intelligence systems and machine learning algorithms that have the capability to consume large amounts of data and to learn from that data in real time. Most medical institutions have transitioned from paper records to electronic health records, and these electronic systems collect clinical, laboratory, and demographic information on patients in both the inpatient and outpatient settings. This results in massive amounts of catalogued data that can be used efficiently or inefficiently. If we could access all of the clinical data at our fingertips, we can use these systems to help us make decisions and to care for patients. We can also increase the speed of medical discovery without having to collect new data for each idea.   

How do you think AI could be incorporated to speed up the process of innovation in medicine and get new treatments to patients faster?

Dr. Leidner: One of the hardest parts about making new treatments available to patients is making sure they are ready to be prescribed and provided by the entire healthcare system. Your doctor must be comfortable with new side effects, your pharmacist must know how to monitor the medicine, and your insurance company must cover the medication. AI can provide doctors with augmented decision making to help identify patients that could use new treatments and help monitor those patients' side effects appropriately, alleviating much of this problem.

What is your goal for your time in the MSAI program?

Dr. Leidner: My goal is to be able to analyze data in new ways in order to find patterns that were previously undetectable by older statistical methods. There is so much medical data that has been analyzed and stored in the past 40 years, and inside that data are key discoveries that could help people avoid illness and complications.

Dr. Gay: I have a couple of research projects I am involved in that focus on applying machine learning algorithms to different clinical information for the purpose of making medical diagnoses that clinicians would have a difficult time making on their own.  A lot of my course work in the MSAI program is directly focused on gaining a practical and working knowledge of these systems.   

How do you hope to incorporate the lessons you learn in the program into your medical work?

Dr. Gay: I would like to continue to be involved with medical research, and my current training should help me think of creative solutions to difficult problems I encounter in my medical practice. Ultimately, developing an appropriate way to test these solutions is as important as the technology itself. We want to be confident that the advanced tools we are capable of building have a positive impact on patient care. Having a background in public health and AI will be beneficial for both the development and deployment of these solutions.