Overwhelmed by thousands of documents describing the inner workings of pharmaceutical companies, the Drug Industry Document Archive wants to enlist the help of crowds.
Documents uncovered during lawsuits against drug companies could be made searchable to the public, just like documents from tobacco company lawsuits. The problem is that there are more files than DIDA's own workers can handle. Until they're processed, they can't be properly searched. Crowdsourcing the project could speed the database's growth.
"In the long run, it's not feasible to get grants to add the documents, and we want to do it sooner rather than later," said Kim Klausner, the Archive's manager.
DIDA is an offshoot of the University of California, San Francisco's Legacy Tobacco Documents Archive, which was born from the 1998 legal settlement between tobacco companies and 46 states that sued them. As a condition of the settlement, all industry documents uncovered during the trial had to be made readily available to the public. It now contains 11 million documents numbering some 60 million pages.
In 2006, two UCSF professors who had served as expert witnesses in a lawsuit against pharmaceutical company Parke-Davis, accused of illegally promoting their anticonvulsant drug Neurontin for unapproved uses, approached the Tobacco Documents Archive about setting up a similar project for drug industry information. Thus began DIDA, which has since gathered thousands of documents from attorneys and journalists involved in drug company lawsuits, and plans to gather millions more.
But whereas tobacco companies had to hire people to enter document metadata — document type, authorship and other information that makes an archive usefully searchable — the drug company records still need to be indexed, evaluated and entered. Until then, they're of limited use.
Klausner envisions an internet army of students, journalists and concerned citizens helping, in much the same way as the Guardian newspaper invited the public to catalogue records of government-expense violations and the National Library of Australia enlisted crowds to correct errors made by automated scanners.
Data entry can be tedious, but it's a chance to get a first-hand look at potentially revealing information, said Klausner.
In a batch of internal AstraZeneca communications uploaded in February, for instance, a company researcher says the antipsychotic drug Seroquel is "now the responsibility of sales and marketing." In other documents, marketing strategists give instructions on massaging unfavorable data, researchers talk about how to "spin" and "de-emphasize" weight gain caused by the drug,
Astrazeneca is currently being sued for hiding the diabetes risk posed by Seroquel, which has been used by 19 million people worldwide.
Before DIDA is ready for the crowds, however, its interface needs work. Programmers are welcome to help; those with experience with J2EE and Spring MVC are welcome to contact Klausner.
"There are thousands, if not millions of documents that could be added as a result of these lawsuits," she said.
Image: DIDA
See Also:
- Crowdsourcing for Plants
- Crowdsourcing the Flu Vaccine
- Open Data: Help Migratory Bird Observations Fly into the Digital ...
- Open Data: Shuttle Impacts From Space Junk
Brandon Keim's Twitter stream and reportorial outtakes; Wired Science on Twitter. Brandon is currently working on a book about ecological tipping points.
