Computer Scientists Have 'Moral Obligation' To Lead Health Care Research, MIT Prof. Says

For the past couple of days, I've been tagging along at my husband's annual work retreat, a gathering on the Cape of big thinkers from MIT's Computer Science and Artificial Intelligence Lab.

This year, health is on the agenda.

Among those pushing for more research related to health care is John Guttag, a professor in the department of electrical engineering and computer science whose current work focuses on applying advanced computational techniques to medicine. (We featured his recent studies on computational biomarkers that predict heart-attack deaths here.) Guttag told me he believes computer scientists (including those at MIT) have a "moral obligation" to undertake far more research in the areas of health and medicine, indeed, they should lead the way with work that has real-world applications.

In his talk yesterday (which I didn't attend, but Guttag kindly recounted) he essentially issued a challenge to his colleagues:

Over the next decades, computer science can contribute more to improved healthcare than any other discipline...It is simply unacceptable to be a leading academic computer science research lab without being a leader in research at the intersection of healthcare and computer science.

He added that there's a lingering perception among computer scientists that their highly technical research shares little common ground with the all-too-real world of health care. "There is often a tension between application pull and technology push," Guttag said in a followup email. "Academic computer scientists have an easy time applauding new computing techniques, but often struggle at deciding whether novel applications of computing constitute legitimate C.S. research."

http://www.youtube.com/watch?v=Fpv0CWLouzc&feature=player_embedded

In an example of how a more multidisciplinary approach might work, Guttag cites a project in which researchers (including graduate student Michael Rubinstein, recent alumni Hao-Yu Wu and Eugene Shih and professors William Freeman and Fredo Durand) "make the invisible visible through video."

Using video software, the team is able to magnify certain vital signs (blood pressure, pulse, etc) in humans so they can be evaluated without the intrusion of being touched. (This can come in handy in many settings, for instance, dealing with premature babies in the NICU.) Guttag describes the work this way:

As you can see, it is about magnifying color changes and motions in video. We started this work because we were interested in developing a non-contact way to measure vital signs, but it clearly can have many non-medical applications as well. Going forward, we will be looking for ways to improve the techniques so that they can be used for more sophisticated medical applications, but that's a ways off.