Sept. 13, 2007 | The same mathematical foundations used by banks to identify fraud are now being applied to the business of mining adverse event data for signals of true drug safety problems. In light of “more mature” signal detection and management technology — and encouragement from drug safety regulators to adopt it — pharmaceutical companies “have to be more proactive,” says Sanket Agrawal, chief strategy officer at Irvine, CA-based Relsys International.

“Regulatory agencies are building their own database for signaling and endorsing  certain techniques,” says Agrawal. “Peer pressure” and consumer expectations are spurring more technology-enabled pharmacovigilance, adds Agrawal. Companies also have a lot more “technologically savvy” biostatisticians, epidemiologists, and clinical researchers in their drug safety department to champion that approach.

But even if a signal is detected, the question then becomes how to manage the information. Should it be sent to an independent safety monitoring board or ethics committee? When should regulators be told? “A lot of what is detected is not real,” says Agrawal. “There are a lot of false positives.”

How signals get managed needs to be documented and auditable, if only to counter legal and regulatory inquiries that may later emerge, says Agrawal. In terms of liability, “it is not just that a product caused damage, but whether or not a company could have known about it sooner and done something about it.”

The “unique value” of Argus Perceptive, the recently developed signal management toolkit of Relsys, is that it addresses all three key elements — signal detection, signal management, and signal work-up — of drug safety surveillance, says Agrawal. The suite of applications works with any adverse event reporting system.

Multiple companies are in various stages of implementing Argus Perceptive, says Agrawal.  What-if modeling built into Argus Perceptive helps companies correctly define detection criteria, based on the relative volume of false positive results.

“Most companies are adopting these capabilities incrementally,” says Agrawal. At the moment, signal detection and documentation are the primary areas of interest. It may take one to two years before most pharmaceutical companies have a “structured methodology” for risk and signal management that Argus Perceptive makes possible. Industry and regulatory agencies are  in the “experimentation stage,” says Agrawal. “Our interests are all the same at the end of the day.”

Subscribe to Bio-IT World  magazine.