FDA Resurrects Data Mining Project Once Left for Dead
Posted 01 April 2014 | By
The US Food and Drug Administration (FDA) has resurrected a once-canceled project intended to better mine data about adverse drug reactions from literature reports.
FDA regularly puts out requests for proposals and solicitations on the Federal Business Opportunities (FBO) website, and has regularly expressed a keen interest in methods of monitoring adverse event data.
In April 2013, for example, it put out a notice indicating it wants a partner to help it build a safety prediction and data mining tool to help it better assess drug adverse event safety signals as early as possible.
That solicitation, which Regulatory Focus wrote about at the time, later morphed into a more thorough (and slightly different) document package, complete with several responses to industry questions by FDA.
The gist of the program, as explained by FDA, is to adopt earlier recommendations by its Science Board to make it easier to mine data and recognize adverse event safety signals automatically.
"An important goal is to allow the development of prospective hypotheses to provide to Office of Surveillance and Epidemiology (OSE) and Office of New Drugs (OND) to inform their evaluation of potential safety signals with the goal of reducing time and effort spent on false positive signals," FDA wrote in the notice.
The intended tool would automatically search the MEDLINE database's massive store of literature reports-a monumental task that might ordinarily take a large team of dedicated and highly educated FDA staff to accomplish.
But the project never made it to fruition. In September 2013, FDA canceled the funding announcement citing a "lack of funding."
It was unclear at the time if the cancellation was the result of budget woes, or simply a shift in internal priorities at the agency.
The project had been slated to begin by focusing on fluoroquinolone products and lipid-lowering drugs, both products with known safety issues.
… Then Brought Back from the Dead
On 31 March 2014, FDA resurrected a project of the same name and with similar-though not the same-aims.
The FBO notice, Data Mining Tool for Quantitative Adverse-Event Signal Detection In Literature Reports, seems to be more similar to an April 2013 FBO notice entitled "Pharmacological Mechanism-Based Safety Prediction-Data Mining Software/Support."
As Focus wrote about at the time, that notice had been borne out of a recommendation by FDA's Science Board that it create a new safety program within CDER's Office of Clinical Pharmacology (OCP) called the "Pharmacological Mechanism-Based Drug Safety Prediction" (PMDSP) program.
That program, the notice said, is intended to mine data and recognize patterns through the use of technology in order to develop predictions for as-yet unrecognized safety signals.
Under FDA's latest FBO notice, the PMDSP program would be given additional amounts of support to mine data and recognize patterns to help "Develop predictions for safety signals that have not yet been recognized for drug candidates that have come to the FDA as New Drug Applications (NDAs).
"An important goal is to allow the development of prospective hypotheses to provide to Office of Surveillance and Epidemiology (OSE) and Office of New Drugs (OND) to inform their evaluation of potential safety signals with the goal of reducing time and effort spent on false positive signals," FDA wrote.
The contract-a one-year contract with four one-year options-calls for an outside contractor to create an IT system that can data mine "a very broad literature base" including PubMed, the US Patent and Trade Office, and "other relevant databases."
Such a system would also be able to detect signals that have "not been observed," FDA said.
For now, such a system only exists in the minds of FDA, but the agency is hoping for interest from potential outside contractors by 15 April 2014 to make it a reality.
FDA Announcement on Data Mining Project