John Guttag

WRITTEN BY: Matt Busekroos

John Guttag is the Dugald C. Jackson Professor of Computer Science and Electrical Engineering at MIT where he leads the Computer Science and Artificial Intelligence Laboratory's Clinical and Applied Machine Learning Lab (CAML). The group develops and applies advanced machine learning and computer vision techniques to a variety of clinically relevant problems.  Some of Guttag’s most recent research projects include prediction and reduction of adverse medical events, matching patients to therapies and providers, and medical imaging. Guttag has also done research, published, and lectured in the areas of sports analytics, financial analytics, software defined radios, software engineering, mechanical theorem proving, and hardware verification.

CAML’s work medical imaging covers a raft of modalities, including MRIs, CTs, and x-rays. According to Guttag, most of his group’s effort on imaging has been on improving the fundamentals of processing medical images, for example image registration and segmentation. “We want to develop foundational methods upon which others can build a variety of applications,” Guttag said. He points that for many applications image registration and segmentation are critical early steps. “We think that this will have a big impact in the long run, not only on clinical care, but also on research,” Guttag said. 

In addition to his work at MIT, Guttag is also the chief technology officer of a health care company called Health at Scale. Whereas most of his work as a researcher deal medical imaging, signals and electronic healthcare records, Guttag works almost entirely with billing data in the company.

While billing data contains less information than clinical data, it is easier to acquire in volume. “We have billing data, many years of billing data, on well over a hundred million Americans, and billing data on almost every provider in the country,” Guttag said. “Surprisingly, if you do your machine learning correctly, you can construct a pretty good model of someone's health, just from what has been billed.”

One of their products matches patients to physicians, based on a patient’s medical history and physicians’ histories of outcomes with similar patients with similar complaints. “It’s not about ranking physicians, since the best physician for one patient might not be the best for another,” says Guttag. “The data clearly indicates that we can reduce bad outcomes by matching people to the right doctor—in some settings by as much as 18%. This is, obviously good for the patient. It’s also good for the healthcare system as a whole, since reducing bad outcomes dramatically reduces costs.”

Guttag is optimistic about the future of impact of computer science research on healthcare. “I talk to a lot of students who are interested in medicine and health and biology,” Guttag said. “The message I try and give them is that as computer scientists, we have an enormous opportunity to improve people's health, and improve the healthcare system. And we shouldn't forget that it's an opportunity and it's a responsibility. I think computer scientists will probably have more impact on healthcare than any other field over the next decade.”