(원문: 여기를 클릭하세요~)

 

Science  20 Jul 2018:
Vol. 361, Issue 6399, pp. 212
DOI: 10.1126/science.361.6399.212

 

Under pressure to be more transparent about the results of drug testing studies, some companies have begun to share anonymized patient data from clinical trials with approved researchers on secure websites. An online platform launched this week aims to expand such efforts by offering a one-stop clearinghouse for those seeking to mine these data for new insights.

The platform, created by Vivli, a nonprofit based in Cambridge, Massachusetts, debuts with access to more than 4000 clinical trial data sets from eight companies and nonprofits. It also features tools for combining and analyzing the data. “This is the first time it’s all going to be available in one place,” Vivli Executive Director Rebecca Li says.

Vivli, which spun out of a policy think tank at Harvard University–affiliated Brigham and Women’s Hospital in Boston, is part of a push to encourage drug developers to share trial data—even negative results, findings that show a treatment has no benefit. Companies seeking U.S. regulatory approval for a drug, as well as investigators funded by the National Institutes of Health, must post limited, summary results on ClinicalTrials.gov. But many researchers and policy analysts believe sharing detailed raw data on individual patients, stripped of identifying information, would be valuable. Researchers could confirm that a drug works, look for side effects, or explore new questions.

Starting this month, the International Committee of Medical Journal Editors—which includes the leaders of many major journals—will ask submitting authors to include a data sharing plan that can include patient data. Such sharing remains controversial. In 2016, Jeffrey Drazen, editor of The New England Journal of Medicine, worried in an editorial that it would embolden “research parasites”—scientists who request others’ data and quickly publish papers, preempting the scientists who generated the data. But Drazen ultimately endorsed the committee’s plan.

Some companies have already responded. Drug giant Johnson & Johnson allows researchers to request patient data at a 5-year-old site called YODA, sponsored by Yale University, whereas GlaxoSmithKline and 13 other firms share data at ClinicalStudyDataRequest.com.

Vivli aims to streamline researchers’ ability to find, request, and combine data from these and other sites, Li says. It will both list data deposited elsewhere and eventually host data sets. GlaxoSmithKline, for example, is allowing Vivli to list more than 2000 of its data sets. Vivli will have an independent panel review some requests, but refer others to the sites that hold the data. Because of patient privacy concerns, users often won’t be able to download the data to their own computers, but will use the Vivli platform.

Companies can purchase memberships to have Vivli share their data. Academic researchers will pay $2000 to $4500 per study for storage and sharing services. Data miners can freely use the site’s basic tools for a year, but after that will pay a daily fee of $12. Both the Bill & Melinda Gates Foundation and Harvard plan to help researchers cover data submission costs.

At the Gates Foundation in Seattle, Washington, officials anticipate that many grantees will deposit their clinical trial results in Vivli in order to meet the foundation’s data-sharing requirements. And Harvard officials will be encouraging faculty to add their clinical data sets, including hundreds from already completed studies.

Some data sharing advocates are pleased by Vivli’s arrival. “We need to get everyone behind one platform instead of having a proliferation of these things,” says epidemiologist Evan Mayo-Wilson of Johns Hopkins University in Baltimore, Maryland. There is uncertainty about demand, however. A 2016 study by researchers at Duke University in Durham, North Carolina, found that scientists had requested access to just 16% of more than 3200 patient data sets available on three platforms. One obstacle was the difficulty finding the data, says Duke cardiologist Eric Peterson, an author. Vivli could resolve that problem, he says, by serving as a clinical data “card catalog.”

  • * With reporting by Elizabeth Gamillo.

 

 

 

 

Leave a Reply

Your email address will not be published. Required fields are marked *