EVIDENCE AND EXTRAPOLATION: MECHANISMS FOR REGULATING OFF-LABEL USES OF DRUGS AND DEVICES

A recurring, foundational issue for evidence-based regulation is deciding whether to extend governmental approval from an existing use with sufficient current evidence of safety and efficacy to a novel use for which such evidence is currently lacking. This "extrapolation" issue arises in the medicines context when an approved drug or device that is already being marketed is being considered (1) for new conditions (such as off-label diagnostic categories), (2) for new patients (such as new subpopulations), (3) for new dosages or durations, or (4) as the basis for approving a related drug or device (such as a generic or biosimilar drug). Although the logic of preapproval testing and the precautionary principle-first, do no harm-would counsel in favor of prohibiting extrapolation approvals until after traditional safety and efficacy evidence exists, such delays would unreasonably sacrifice beneficial uses. The harm of accessing unsafe products must be balanced against the harm of restricting access to effective products. In fact, the Food and Drug Administration's (FDA's) current regulations in many ways reject the precautionary principle because they largely permit individual physicians to prescribe medications for off-label uses before any testing tailored to those uses has been done. The FDA's approach empowers physicians, but overshoots the mark by allowing enduring use of drugs and devices with insubstantial support of safety and efficacy. This Article instead proposes a more dynamic and evolving evidence-based regime that charts a course between the Scylla and Charybdis of the overly conservative precautionary principle on one hand, and the overly liberal FDA regime on the other. Our approach calls for improvements in reporting, testing, and enforcement regulations to provide a more layered and nuanced system of regulatory incentives. First, we propose a more thoroughgoing reporting of off-label use (via the disclosure of diagnostic codes and "detailing" data) in manufacturers' annual reports to the FDA, in the adverse event reports to the FDA, in Medicare/Medicaid reimbursement requests, and, for a subset of FDA-designated drugs, in prescriptions themselves. Second, we would substantially expand the agency's utilization of postmarket testing, and we provide a novel framework for evaluating the need for postmarket testing. Finally, our approach calls for a tiered labeling system that would allow regulators and courts to draw finer reimbursement and liability distinctions among various drug uses, and would provide the agency both. the regulatory teeth and the flexibility it presently lacks. Together, these reforms would improve the role of the FDA in the informational marketplace underlying physicians' prescribing decisions. This evolutionary extrapolation framework could also be applied to other contexts.