Limitless. That's the potential that the American Medical Association says artificial intelligence can offer to the way health care is currently delivered. This machine learning is already in use in many hospitals, but as AI continues to evolve, so to too are the ways in which it is transforming the very practice of medicine.
At SUNY Oswego, Computer Science Professor Daniel Schlegel is applying to a common issue. When doctors prepare a plan for how to treat a patient's disease, they are often informed by clinical practice guidelines evidence-based blueprints for specific afflictions. The problem, says Schlegel, is that there are a lot of them and many of the guidelines conflict. That can present an issue for doctors who are treating patients who suffer from multiple diseases.
"There is a lot of effort needed on the side of the physician to read all of these guidelines, figure out how they interact, where they intersect, and come up with a plan of attack for each individual patient. It would be ideal to have the computer do a lot of that work for the doctor."
That's where Schlegel's computer science credentials come in. He's developing an algorithm that could automate the process of taking hundreds of pages from multiple clinical practice guidelines and come up with a way that a doctor can resolve these conflicting recommendations. Schlegel says that does more than save time.
"There’s sort of an important safety issue here where you can actually stop issues before they begin," He said.
Projects like these are why the American Medical Association says AI in healthcare is more accurately referred to as augmented intelligence. The focus currently is on aiding humans rather than replacing them according to Iman Hajirasouliha, an assistant professor of computational genomics at Weill Cornell Medicine.
"AI is more like a tool that can help doctors make better decisions, make things more standard, and make things more automated and fast," Hajirasouliha said.
Hajirasouliha recently created an algorithm that can accurately determine whether an in vitro fertilized human embryo has a high potential to progress to a successful pregnancy. Researchers hope it could improve the success rate of in vitro fertilization.
"Machines or algorithms can handle large data sets very quickly and having a large data helps us," Hajirasouliha said. "There is a lot of potential in using algorithms and computational techniques for a wide range of problems in biomedicine – not only on imaging applications and sequencing, but a wide range of applications."
AI is not be new to healthcare, yet it's taking off rapidly right now alongside the evolution of new and faster technology. The increasing frequency of digitized data in healthcare is also responsible for this boom says Schlegel.
"With the sort of mandate that hospitals start using electronic healthcare record systems, there’s a lot more electronic data and that opens up the avenues for doing something with it," Schlegel said.
But at least for now, humans are still in the driver's seat in both writing the code that these machines learn and in making the ultimate calls on how to treat patients.
