The New England Journal of Medicine quickly backtracked and clarified its wholehearted support of data sharing, but the hashtag #Iamaresearchparasite quickly rippled across the Internet, with people as prominent as the White House’s chief data scientist and the editor of the prestigious journal Science claiming membership in this select society of scientific bottom-feeders.
But research published in the Journal of the American Medical Association this month suggests we actually don't have enough of these so-called parasites. The study found that, among three online portals that allow scientists to request the data behind thousands of trials from pharmaceutical and medical device companies' trials, only 15.5 percent of the trials had ever been requested. That follows another study that found that although nearly half a million clinical trials could be found on a database called MEDLINE, there had been only 37 published reanalyses of randomized, controlled trials.
This is a problem because data sharing ensures that findings -- such as the study that got a drug approved -- can be re-examined and checked. Data sets can also be combined to try and answer questions that didn't occur to the people who originally designed the study. Data parasites, scientists argued, were anything but parasites and were actually crucial to the process of science moving forward and taking full advantage of the evidence gathered.
“Much of what we know about the large-scale features of this planet is apparent thanks to widespread data-sharing practices and the early establishment of data banks in the geosciences,” Marcia McNutt, the editor-in-chief of the journal Science wrote in an editorial. “Many fields, including the biomedical sciences, are now benefiting from meta-analyses of data to better understand the big picture.”
Vice president Joe Biden even mentioned data sharing as bullet point two in his effort to enable a cancer moonshot:
Despite the considerable lip service that data sharing gets, the researchers behind the JAMA study determined there are relatively few people actually taking advantage of the data.
“In the research community, there’s been a lot of calls in the last several years -- and they're getting louder -- to increases access to clinical trial databases,” said Ann Marie Navar of Duke University Medical Center, the lead author of the study. “What was surprising to us is first, in general, the lack of utilization of these platforms.”
In other words, an "if you build it, they will come" approach may not work that well when it comes to data sharing.
To fully understand the debate, it's important to note the reasons people do these studies-of-studies. The data underlying one finding contains volumes of data that might also yield the answer to questions that never occurred to the original researcher.
For example, researchers trying to understand the strength of the placebo effect in a particular disorder could look across many studies of different drugs in which placebos were used, to try and measure the effect. Researchers could attempt to unpack why small populations of people within multiple studies responded to a treatment that was otherwise unimpressive. Transparency in data could also allow researchers to re-analyze studies and confirm their results or point out flaws in the original analysis.
To Jennifer Miller, an assistant professor in the division of medical ethics at the New York University School of Medicine and the president of Bioethics International, a nonprofit committed to transparency, the amount of utilization is small but encouraging.
“It’s so early in the data-sharing movement,” Miller said. She noted that among phase-three clinical trials, the most advanced and largest trials, the rate of requests for data was higher. About a quarter of those trials had been requested.
To Navar, the findings suggest there should be increased incentives for people to reexamine studies -- meaning grants that support such work. She also hopes for greater awareness among researchers that this data exists.
In an accompanying editorial, the editors of JAMA argued that while transparency and openness are fundamental to science, there’s another reason to support this kind of sharing: to respect the contributions of the patients who have participated in clinical trials, putting themselves at risk in order to contribute to science.
“The social contract for taking these risks and experiencing these harms impose an ethical obligation that the results lead to the greatest possible benefit to society,” the editors wrote. “The absence of data sharing … constitutes a failure of the obligation of researchers to the study participants, and therefore a failure of the ethical underpinnings of conducting clinical trials.”
More from Wonkblog: