STEVE INSKEEP, HOST:
Here's a story of how we learn. We build on the work of those who came before. Musicians make variations on old music. Historians cite or challenge or completely overturn the findings of historians who came before. The same is true in science - or it's supposed to be. It's hard for medical researchers because they often do not share the raw data from their research. Now medical journal editors have called for more access, and that is good news to Harlan Krumholz of Yale, who is an advocate for more openness - a finding that grows out of his experience as a researcher.
HARLAN KRUMHOLZ: Yeah, I've had this experience, especially when I was a young investigator just getting started. I approached a lot of senior investigators and almost resoundingly heard no every time. They had done the studies. They were going to control the data, and they were going to choose who was going to get to touch and use that data. And I was completely walled off from that.
INSKEEP: I can see the importance of this if I just think about journalism, or if you think about research in other fields. People build on the work of other scholars. And you're saying you need the underlying data. You need the basic facts in order to work from and build onto that with your new research.
KRUMHOLZ: Look, one of the basic principles is show your work. Let other people see it. This would be unfathomable in other fields. Imagine an astronomer creating a super-telescope but only a couple of them got to go and look, and everyone else had to hear what the universe was like through the papers that these individuals wrote but never got to see the original data or to actually view the universe with their own eyes. So this is something that other fields have solved a long time ago. And yet in medicine, for some reason, we've developed a culture where people sequester the data. They hold it back.
INSKEEP: You said for some reason this happens. What do you think the motivation is of researchers who hold back data?
KRUMHOLZ: Well, people who do research legitimately spend a lot of time, and the notion for them is that the reward is not only publishing the main paper but having the ability to continue to publish for many years afterwards. So they also may be a little reluctant to allow others to peer into the insides, the inner workings, to see the raw data. It's not that necessarily anyone's done anything wrong, but it's a sense of that. I've also heard some people say, which I think is inexplicable, that only they can understand their own data. And they're worried if other people get in it with their data, they'll draw conclusions that they disagree with, when that's exactly and precisely the point. We shouldn't be afraid of that. In fact, we should be fostering it and encouraging it because it will represent the best of science. And if someone comes out with a conclusion that can't be supported, someone else will have access to the data to be able to demonstrate that.
INSKEEP: Are there drawbacks to having all your information out there?
KRUMHOLZ: Well, first and foremost we have to be sure that patient privacy concerns are respected. I think the concern is that it will add noise to the environment. The people will be going back and forth and no one will know what the truth is. I personally believe that the contrary will be true - that as the data comes out, consensus will coalesce and many people will have seen the data.
INSKEEP: One other thing - if this proposal becomes standard practice, there may come that moment when you receive an email asking for the raw data from your research. Even though intellectually you're going to agree with giving up that data, are you going to have a moment in your heart of just feeling a little tight about it - a little tense about it?
KRUMHOLZ: Well, I've done that. We ran a $10 million NIH trial to look at telemonitoring in patients with heart failure. Within a year of publishing that study, we made that data available. We gave it back to NIH, and we told them they could share it. You know, I wouldn't be honest with you if I didn't say that was hard. And it felt, you know, a little like my baby was being released. I mean, I'd spent so much time on this thing. But we did. And I'm glad we did it. And by the way, many others are doing this. The NIH has a repository of trials. So we're not alone. Many are doing it. So it's not that it's not being done, it's just not being done very much. And this would change that.
INSKEEP: Doctor Harlan Krumholz at Yale University. Thanks very much.
KRUMHOLZ: Thank you so much.
INSKEEP: And you can read his post about data sharing on Shots, NPR's health blog.