Drugmakers Mine Data for Trial Patients

To lower their testing costs, some buy access to hospital records

Pharmaceutical companies can easily spend years—and more than $1 billion—bringing a new drug to market, in part because they can’t find enough patients to do the required testing of the compound. Such delays can cost up to $1 million a day, fritter away valuable months of patent protection, and allow rival developers to catch up. One remedy: pay hospitals to sift through the health records of their patients.

Five big drugmakers, led by Pfizer, are planning to use electronic health data gathered from patients of 13 hospital systems across New York State to help them identify and enroll participants in drug studies. The effort, which begins testing this month, is projected to make $75 million a year for the hospitals and save the pharmaceutical companies time and money in developing new products. “This is going to be a game changer, making medicine more of a science and less of an art,” says John Murphy, senior director of clinical analytics for Quintiles Transnational, which helps drugmakers conduct trials.

The program, called the Partnership to Advance Clinical Electronic Research, or PACeR, is among scores of such initiatives created nationally as medical providers, software vendors, and health data businesses seek ways to profit from the flood of clinical data now being gathered electronically. Besides Pfizer, companies helping fund PACeR include Merck, Roche, Johnson & Johnson, and Bayer HealthCare. Quintiles and Oracle are developing the system.

Federal law bars medical providers, hospitals, and insurers from disclosing identifying information such as names, addresses, and Social Security numbers. So drug companies that want to test a new product or compound would pay PACeR to query the records systems of participating hospitals to compile a list of patients who match a trial’s requirements. Each query would cost between $50,000 and $200,000.

Once they determine how many patients might qualify and where they’re located, and get approval from a local ethics board, the hospitals would contact the patients’ doctors. A drugmaker would have access to personal information only if the patient consents, says David A. Krusch, head of PACeR’s leadership team and director of medical informatics at the University of Rochester Medical Center. “There is no central database,” Krusch says. “We’re not dumping big buckets of de-identified data anyplace. Pfizer will not have a network connection, say, to the University of Rochester, or to any other participant in the network.”

Still, some civil liberties groups worry about data breaches and the potential for third parties to reconnect names to the data. “In a world where so much data is being retained, exchanged, and sold, being able to protect the privacy of individuals is a lot more difficult,” says Lillie Coney, associate director of the Electronic Privacy Information Center, a Washington-based advocacy group. Even data scrubbed of someone’s name and Social Security number can be “re-identified,” Coney warned, citing a 2006 case where AOL released information about people’s search queries, which others were able to combine with publicly available data to identify an elderly Georgia woman who had used her AOL account to research medical conditions.

Consumer and ethics groups who advised the New York project say they see patient benefits as long as confidentiality is protected. “One of the problems with clinical trials is it’s very hard to get the information out to the average physician, or to the average cancer patient being treated in the community,” says research scholar Karen Maschke of the Hastings Center, a bioethics group that consulted on PACeR. Maschke says that while she supports the PACeR concept, “I’m also a skeptic. There probably should be conversations at the national level about best practices for these endeavors to guard confidentiality.”

Such discussions are expected to increase around the country as the use of electronic health records increases dramatically over the next few years, spurred by $27.4 billion set aside in the 2009 U.S. stimulus to pay doctors and hospitals to adopt and use them.

Drugmakers are big potential customers of aggregated health data since drug development times have more than doubled in the past 20 years without a marked increase in the percentage of compounds successfully brought to market. “If pharmaceutical companies can make this happen faster and more cheaply, they’re big winners,” says C. William Schroth, a former consultant for the New York State Health Dept., who first broached the concept behind PACeR to a group representing New York hospitals and also to drug companies.

Doctors and hospitals liked the idea because it makes them more attractive research partners. It wasn’t a hard sale for Big Pharma, either: Every day of delay during a Phase 3 trial costs drugmakers more than $1 million, with an average delay of 90 days, says David Leventhal, director of clinical innovation at Pfizer, citing a Deloitte Consulting analysis. “Even if that number is only 25 percent right, it’s still a compelling message.”


    The bottom line: A group of 13 New York hospitals will sell access to patient data to drugmakers for $50,000 to $200,000 per search.

    Before it's here, it's on the Bloomberg Terminal.