Thomas Trikalinos

Image

Thomas Trikalinos

Associate Professor of Health Services, Policy, and Practice

Frank Mullin/Brown University
Thousands of studies and research reports — sometimes apparently contradictory — can be brought to bear on any clinical issue. Tom Trikalinos and colleagues use mathematical approaches to pursue clarity despite different methodologies, patient groups, and other factors.

Tom Trikalinos always loved math, but in Greece when you are a good high school student, you are expected to go to the toughest college program around. In the lakeside town of Ioannina, that was the medical school at the University of Ioannina.

“I liked studying medicine, but I don’t think I was meant to be a clinician,” Trikalinos said. “I realized that very soon.”

So what to do you when you are a physician who’d rather wield a slide rule than a stethoscope? You practice mathematical medicine. That’s essentially what Trikalinos does as the director of Brown’s new Center for Evidence Based Medicine.

EBM is a statistics-heavy field that has grown from the notion that clinical decisions (such as what to tell a patient about taking vitamin D supplements to prevent heart disease) should be guided by published evidence of efficacy, cost, risk, and other factors. Often the relevant research is scattered among dozens of studies, all with subtle variations in methodology, patient populations, and other aspects of their design. Only skillfully developed and applied mathematical techniques can reconcile and combine the data to deliver a single, definitive basis for an answer.

Trikalinos has been doing this kind of work since he was a medical student and through his Ph.D. years. In 2006, he graduated and left Greece for New England to work with Dr. Joseph Lau and his team at Tufts Medical Center, where he also met Chris Schmid. Lau took him under his wing, and they have worked together ever since.

This spring, however, Lau, Trikalinos, and Schmid decided to move together to Brown along with Dr. Issah Dahabreh and computer scientist Byron Wallace. They are now the founding members of the new center that Trikalinos leads.

They are off to a fast start. In August, Brown won designation by the federal Agency for Healthcare Research and Quality as an Evidence-Based Practice Center, which qualifies Brown to compete for contracts to do EBM research. When AHRQ has a question, for instance about the best way to diagnose a disease, Brown will now be among about a dozen places they look to for an answer.

Trikalinos is quick to note that the move from Tufts Medical Center to Brown University was not merely a matter of cut and paste. In Brown’s more academic context, the researchers want to expand the scope of their work.

“This center is a continuation-slash-reinvention-slash-reincarnation of our previous center at Tufts,” Trikalinos said. “We realized we should be in a university environment because there is a greater ease to reach out and collaborate with other disciplines and to attract students.”

Broader academic collaboration is important to Trikalinos because the application of EBM to specific clinical questions is only part of what his research is about.

“Our main interest is in methodology,” he said. “This research tries to port methodologies from computer science and applied math to optimize the process of doing the applications.”

At Tufts, for example, he and his colleagues worked on ways to bring software to bear on the enormity of doing systematic reviews of studies. It’s a monumental task to search the entirety of the world’s medical literature to make sure that every relevant study can be incorporated into the analysis. It means reviewing thousands of abstracts, and then evaluating scores of potentially relevant papers more deeply.

The team developed an artificial intelligence program that employs “machine learning” to mimic the evaluation process of EBM researchers. In very early testing the software is showing that it can do as good a job as a human researcher at finding relevant studies among the thousands of possible candidates. The team has also developed the Systematic Review Data Repository, freely available software for aiding the process of reconciling data among studies.

That’s not to say that people don’t remain valuable. In fact, Trikalinos said, he’ll hire additional people at the new center. Some will help handle new inquiries coming via the AHRQ designation. Others will expand the center’s expertise into new areas of EBM: making decisions and weighing tradeoffs amid uncertainty; cost-effectiveness analysis; and judging which studies would be most helpful for reducing the uncertainties that prevent important clinical questions from being answered.

So while Doctor Trikalinos won’t be in any single exam room when health care is delivered, his center’s efforts to provide doctors with evidence means that in a sense, he could be in thousands of exam rooms around the world.