The FDA recently awarded Johns Hopkins University a three-year contract to develop advanced analytical strategies to mine the data residing within multiple FDA databases. The goal is to find ways to allow drug and diagnostic firms access to anonymous premarket data (from submissions, failed clinical trials, etc.) related to safety and efficacy for specific patient subpopulations. The project will, as FDA's Dr. Seyfert-Margolis notes, allow the agency to use "data mining with an eye towards sub-stratification of patients for efficacy and toxicity predictors."
Meta-analysis and data mining is actively encouraged in other product development fields outside of the life sciences. New product development specialists even have a term for such endeavors, "bookshelving." The intent is to flag previously gathered data from past products, experiments, failed innovations, consumer feedback, and so on, in order to take one of three possible actions: skip portions of the development process, avoid dead-ends, or better target a product toward a specific set of consumers. In my book, Get to Market Now! Turn FDA Compliance into a Competitive Edge in the Era of Personalized Medicine, I walk readers through how to incorporate this type of meta-analysis and data mining in preclinical and clinical development, as well as include it in quality by design implementation and submissions.
A study published by the Tufts Center for the Study of Drug Development found that less than half of pharmaceutical and biotechnology firms today are pursing personalized medicine projects in their pipeline. Given the slow - admittedly steady - progress of the FDA to publish guidelines associated with various aspects of personalized medicine development and commercialization, it is a testament to the industry's belief in personalized medicine that so many projects are underway without firm FDA rules in place.
Unfortunately, FDA's contract with Johns Hopkins University does not also cover all the postmarket data in FDA's possession. While some postmarket data is available already publicly, there is little doubt that the agency has access to a great deal more. And to realize the greatest return on investment in data mining, bookshelving and personalized medicine development, as much data as possible needs to be included in any meta-analysis. Forcing firms to seek out other individual postmarket databases will not only slow their meta-analysis, it will jeapordize the integrity of any results. Afterall, what if the one postmarket database containing critically relevant information is ommitted from the data mining? How will firms know?
Until the FDA can make all of its premarket and postmarket personalized medicine-related data available, firms may need regulatory affairs data specialists, adept in finding, collating, synthesizing, analyzing data across multiple sources - but do such experts exist?