By Tom Ulrich
Ed. note: This is the first in a two- part series on making clinical trial data more transparent.
2013 was the year when big data became, well, big. Everyone from investment companies to public utilities to security agencies – including medical researchers – are now clamoring for as much data on as many subjects and topics as they can get their digital hands on.
But while data in other fields are becoming ever more open, clinical trial data – especially from corporate-sponsored trials – are relatively hard for medical researchers to obtain.
Finding a balance point
The results of most clinical trials are supposed to be publicly available already, at least in summary form. A federal law passed in 2007 called for trial sponsors to deposit summary trial data in ClinicalTrials.gov, a publicly available website and database maintained by the National Library of Medicine (NLM) originally founded to improve patients’ access to clinical trials.
But as noted last year in the British Medical Journal, summary data aren’t really enough to advance pharmaceutical science and safety. And even those data that are deposited in a database like ClinicalTrials.gov or the European Medicines Agency’s EU Clinical Trials Register or EudraCT databases don’t necessarily make it to academic publication. A study late last year in PLoS Medicine noted that half of all clinical trials conducted in the United States go unpublished.
At the same time, some players within the medical industry are taking steps to making data more accessible. Last year in the New England Journal of Medicine (NEJM), Harvard School of Public Health policy researcher Michelle Mello, JD, PhD, and collaborators noted that companies like Roche, GlaxoSmithKline and Medtronic are all creating policies for opening access to participant-level data from their clinical trials.
Why hold data close to the vest? Concerns that trial participants – who are supposed to be anonymous – could be re-identified based on their raw data constitute one reason. Another is maintaining competitive advantages over competitors. If you put all of the raw data about a trial out in the open, companies argue, competitors could potentially glean insights about your future business plans.
"Pharmaceutical companies are interested in curing disease, but they also have responsibilities to their boards and shareholders," explains Fernando Valles, JD, who negotiates clinical trial agreements with companies for Boston Children’s Hospital’s Technology and Innovation Development Office (TIDO). "They want to hold their competitive advantage and come out with a drug before anyone else."
Ellis Neufeld, MD, PhD, director of the Clinical and Translational Study Unit at Boston Children’s and a hemophilia researcher, cites the risk of inappropriate re-analysis and patient re-identification as barriers to greater transparency. "It’s happened more than once that investigators supported by one company have tried to trash the data generated by trials in another."
"There are overwhelming benefits to society when you have access to big data."
"Even NIH-funded data sets are sometimes censored," he adds, noting that in trials of rare diseases, the risk of patient re-identification is even higher.
The importance of being open
Alexa McCray, PhD, who developed ClinicalTrials.gov while at the NLM and who is now associate director of the Countway Library of Medicine at Harvard Medical School, has heard these objections before.
"The arguments made today against open sharing of clinical trial data are the same ones made 15 years ago against public registration of trials themselves," she says. "But having everyone both register their trials and deposit their results levels the playing field between companies while helping researchers and patients stay aware of what research is taking place. Everybody wins."
In their NEJM paper, Mello and her colleagues noted two key benefits to data transparency: better assessment of drug safety and effectiveness, and greater opportunity for scientific discovery.
"There are overwhelming benefits to society when you have access to big data," she says. "There is a treasure trove of data in each trial that can be used to verify the results, explore new questions, ensure accountability and speed innovation by illuminating new avenues and seeing new ways of conducting clinical research."
McCray’s counterpart Isaac Kohane, MD, PhD – director of both Countway Library and of Boston Children’s Informatics Program – agrees on all counts, stating, "[Openness] creates opportunities to make sure that a study was done in the right population. Given how many disappointments we’ve had recently where trials and postmarketing surveys have come up with contradictory results, it’s inevitable that we’ll have to have greater transparency."
"Lack of transparency can slow discovery," Neufeld adds. "Look at the experience with Vioxx. There were clinical events within the trial data that, if analyzed openly and early, could have revealed the drug’s safety issues much sooner."
Is there a right answer? Are there ways to address companies’ concerns while allowing patients, clinicians and researchers to make best use of these rich sources of data that could impact the health of millions?
We’ll address that in post #2.