Sean Green is trying to train a computer to think like a doctor.
Sean developed an expertise in machine learning – the way computers acquire knowledge – while earning his PhD in Engineering at Carnegie Mellon University.
At IHME, Sean is using that knowledge as part of the Causes of Death research team. He is trying to create a software program that will be able to make the same kinds of diagnoses that a doctor would based on surveys of people who have had a member of their family die, known as verbal autopsies.
“In these low-resource areas, you don’t have death certificates that list the cause of death, so we are trying to fill in that gap through verbal autopsies,” Sean said. “If we had to rely on physicians to review all the surveys to determine the causes of death, it would be prohibitively expensive, and it also perts these doctors from where they are most needed – doing clinical work with patients.”
Sean was attracted to IHME because of its dual focus on broad, population-level health improvements and narrow, practical solutions. While working in India as part of an internship devoted to improving water sanitation in poor areas, he became increasingly interested in the pathogens carried in the water supply and the various ways clean water is vital to a healthy population.
“If you are in a country just trying to fix one problem, and you don’t know how to assess the other problems, you won’t have much of an impact,” Sean said. “At IHME, we are trying to look at the inpidual problems and also where policymakers can make the biggest difference by taking on the primary threats to public health.”
Sean hopes to use what he learns at IHME to create a “reverse brain drain” between low-income and high-income countries. The son of immigrants from Barbados, Sean recalls returning to that country for a visit and wondering why so many people with advanced educations had left. Sean hopes to work in the developing world and to find ways to attract others with similar skills to do the same.
Published Works
Flaxman AD, Vahdatpour A, Green S, James SL, Murray CJL, the Population Health Metrics Research Consortium (PHMRC). Random Forests for verbal autopsy analysis: multisite validation study using clinical diagnostic gold standards. Population Health Metrics. 2011; 9:29.
