Sunshine in medical research

A company hires Yale researchers to investigate its product—and the results are open to all.



View full image

Two years ago, a medical device manufacturer turned to Yale for help.

Facing federal investigations into claims that it manipulated scientific data, the company, Medtronic, might have asked the university’s labs how to improve its bone-growth product. Or it might have hired Yale experts to testify about the effectiveness or safety of the product, called Infuse.

Instead, Medtronic gave the Yale School of Medicine’s Center for Outcomes Research and Evaluation (CORE) a $2.5 million grant and all of its clinical data on Infuse—every patient, every clinical trial. CORE, in turn, hired research groups at two other universities to review the data—independent of each other and of Yale. What’s more, CORE promised that once those reviews were done, it would make the clinical evidence available to any and all outside researchers.

In a world where both industry and academic scientists jealously guard their data sets, and where nearly half of all clinical trial results are never published, the deal represented a breakthrough for open-source science.

“Anyone who didn’t trust the findings, who said, ‘Yale got money from Medtronic’”—and therefore might produce biased results—can check the data for themselves, says Harlan Krumholz ’80. Krumholz is CORE’s director and a professor of medicine, epidemiology, and public health.

Now the results are in. In a series of articles published in the Annals of Internal Medicine in June, reviewers from the Oregon Health & Science University and the University of York in England found that spinal fusion surgeries using Infuse—which consists of a protein embedded in an implanted device—generally worked about as well as the conventional bone-grafting technique, but not better. And the Oregon team raised questions about increased cancer risks in certain Infuse patients.

Medtronic released a statement saying the results “add to a growing body of evidence” supporting Infuse as “a safe and effective treatment option.”

But Krumholz says that “the reviews demonstrated the actual paucity of data” as to whether Infuse, which he notes “was a billion-dollar product,” is safe and effective.

Krumholz does not bash Medtronic, calling it a “partner” in releasing its data. “Many companies have found themselves in similar positions”—accused of suppressing data and funneling payments to outside scientists who publish favorable studies— “and haven’t done anything particularly brave,” he says.

In an Annals editorial accompanying the research, Krumholz and colleagues call the Medtronic studies “A Historic Moment for Open Science.”

Just a few weeks after the articles’ publication, “things have moved faster than I ever thought they would,” Krumholz observes. The Pharmaceutical Research and Manufacturers of America (PhRMA) and its European counterpart announced in July that their members will begin sharing more data with outside researchers and with patients who participate in clinical trials.

“It’s just breathtaking to see that kind of change,” Krumholz says. “I’m really happy to see the [Medtronic] project has sparked this kind of” response.

His next target: academia, where scientists often consider their data proprietary and don’t always publish unexciting results, even if those studies could be helpful to doctors and patients.

“Can Yale guarantee that every trial that’s completed at Yale will have results that are published?” Krumholz asks. “Right now, I don’t know that.”

After that “first step,” he continues, “are we willing to be at the vanguard of sharing data? Are we brave enough to do that? I would love for us to lead.”

The comment period has expired.