Data exclusivity for drugs and biological products gets all the attention.  (In fact, recently I read a law review article asserting that medical devices are not entitled to any sort of exclusivity period after approval.  But this is wrong!)  It is apparently not as well known, but sponsors of premarket approval applications (PMAs) enjoy six years of data exclusivity.  Folks interested in FDA and innovation policy should know about the device scheme because it has a unique history (with a novel and clever – though unworkable – approach in place for seven years) and because at a high level it is still analogous to drug and biologic exclusivity even though the regulatory paradigms are different.

Congress did not enact premarket approval requirements for medical devices until the Medical Device Amendments (MDA) of 1976.  This legislation included a provision – section 520(h)(3) of the FDCA – precluding FDA from using the data in one PMA to establish the safety or effectiveness of any device other than the one for which the data had been submitted.

The First Data Exclusivity Provision (really, the first “use” provision)

Fourteen years later, in 1990, Congress authorized FDA to use the data contained in an approved PMA – for the benefit of its competitors – one year after FDA had approved the fourth device of the kind.  Under the four-of-a-kind provision, release of the data in a PMA was not authorized, but the agency could use the data in reviewing other PMAs, in reclassifying (really, down-classifying) medical devices, or in developing special controls for Class II devices (not subject to the PMA requirement). A key Senate report at the time explained that the “original purpose” of the “absolute prohibition on the use of PMA data” in 1976 was “to support a relatively youthful industry and create incentives to foster innovation.”  This need had been “particularly important” because the 1976 statute required disclosure of “detailed summaries of safety and effectiveness data.”  Fourteen years later, though, the “absolute protection of safety and effectiveness data” was “no longer justified” because the device industry was “mature and strong.”

Pegging the Length of Exclusivity to the Investment Made

The four-of-a-kind approach was fundamentally different from the approach the legislature had taken for human drugs in 1984 and animal drugs in 1988.  The most interesting implication is that the length of the exclusivity period was inherently variable.  It might be two years, five years, ten years, or longer before FDA had approved the fourth of a particular type of device.  The variability was intentional.  The key Senate report explained that the new approach would “provide a degree of protection to device firms roughly proportionate to the size of a firm’s investment.”  Specifically, “since the disclosure of data is dependent upon the approval of a fourth device of a kind, the time it will take to achieve four approvals will largely relate to the novelty and complexity of a device.”  In other words, “newer and more complex technologies will likely involve greater commitments of time and money,” and the Committee believed “substantially longer periods of data protection will result for the early developers.”  Conversely, “where product development is less ambitious, and costs less, lower barriers of entry to the marketplace will exist,” which will lead to “more PMA approvals and shorter periods of data protection.”

These assumptions strike me as overly simplistic.  The time it would take to achieve four approvals could well reflect the strength and number of the various patents protecting the technology in question, for instance, as well as the attractiveness of the market for the type of device in question.  Each of the first three applicants would have a different exclusivity period, and assuming that the length for each corresponded in some fashion to the size of its investment (rather than other factors that might affect ordering) seems naive.  And the rule creates some very strange incentives.  The fourth company to file a PMA could make as large an investment in its own clinical trials as the first company, but immediately upon its own approval the data in all four applications would be fair game for use by competitors.  No one would want to be fourth, and the first three could have a strong interest in ensuring there was no fourth.

Abandoning the Four-of-a-Kind Rule

The four-of-a-kind rule was generally viewed as unworkable, and in fact FDA never used it.  And it may be worth noting that it had been something of an afterthought, to begin with.  In the summer of 1990, the draft legislation that would become the Medical Device Amendments had proposed ten years of exclusivity, which FDA endorsed.  When Senator Kennedy introduced Senate Bill 3006 in August 1990, he proposed a fixed term of four years.   The four-of-a-kind approach was substituted during a markup in September 1990.

In any case, Congress replaced it with a fixed exclusivity term seven years later.

The 6-year provision in today’s law was part of the 1997 user fee statute.  There is nothing in the legislative history to explain why “six” was the number chosen.  Indeed, as FDA has noted in a guidance document on the data exclusivity provision, there is “little explanation” of the provision at all, in the legislative history.  The key Senate Report simply summarizes the provision:

The Secretary is authorized to use data from a premarket approval application six years after an application is approved for the purpose of facilitating reclassification and/or approval of applications submitted by other device sponsors for the same kind of device.

Current Law

Today, section 520(h)(4) of the FDCA provides six years of data exclusivity for medical devices approved pursuant to a PMA.  (This appears in 21 U.S.C. § 360j.)  FDA may release a detailed summary of safety and effectiveness of a medical device at the time a PMA is approved.  Generally, these summaries cannot “be used to establish the safety or effectiveness of another device . . . by any person other than the person who submitted the information.”  Six years following the date of approval of a PMA, however, FDA may make available “[a]ny information contained in [that] application for premarket approval . . . that demonstrate the safety and effectiveness of [that] device.”  These data may then be used by FDA in approving another device, establishing performance standards, reclassifying the device, or for other purposes.

Use of the Provision

Although FDA did not use PMA data when the four-of-a-kind rule was in place, it has used PMA data under the new law.  In response to an interpretive dispute and citizen petition filed in 1999, FDA determined that it would apply the new language only to data in PMAs approved after November 28, 1990, the data of the provision’s enactment.  But, with this interpretation in place, the agency has reclassified numerous medical devices using data in PMAs that had been approved six or more years before the date of the reclassification decision.  (For an example, see here, reclassifying extracorporeal shock wave lithotripters from class III to class II.)

FDA has also explained in guidance how PMA applicants should reference the data from another company’s approved PMA and how PMA holders may learn that the data in their files have been used.  And it has suggested “Pre-Submission” — a written request for feedback — when a company wants input on the extent to which existing data can be leveraged in its PMA submission.

As far as I can tell, there have not been any other citizen petitions relating to interpretation or use of the provision.  Someone well versed in citizen petitions filed with FDA’s drug center might imagine a petition from one company objecting to the use of its data to support another company’s PMA, but this simply hasn’t happened.  Citizen petitions filed by one device manufacturer regarding the approvability of another manufacturer’s device are rare at the device center, in any case.  (For a rare example, see this petition from Wright Medical Technology, regarding another company’s hip resurfacing system.)

It’s not clear to me how often PMA data are used to support a single PMA filed by another company, but it is quite clear that the agency has used PMA data repeatedly to reclassify medical devices.

Any Relevance to Drugs and Biologics?

The regulatory paradigm for medical devices is very different from the regulatory paradigm for biological and non-biological drugs, most notably in the absence of a “copy” provision – the equivalent of the biosimilar pathway and the generic drug pathway.  Put another way, the paradigm for drugs and biologics involves full applications by pioneers and — some number of years later — abbreviated applications by their competitors.  The competitors submit comparative data sufficient to justify reliance on the pioneer’s data application.  This bifurcated approach — a single full application followed some time later by abbreviated applications — has no parallel in the device world.

The size and scope of medical device premarket submissions depends on their classification, which in turn is a function of their risk.  There is no such thing as a generic device application.  It’s not unreasonable to characterize the premarket notification — 510(k) submission — as abbreviated, in the sense that it does not purport to demonstrate safety and effectiveness, but rather makes a comparative showing relative to another device.  But the predicate device was not itself “approved” on the basis of a full application showing safety and effectiveness, and the 510(k) does not rely on previously submitted safety and effectiveness data.

The device data exclusivity provision is thus drafted very differently.

At a more general level, though, I think there is a useful analogy.  It relates to reclassification.  Consider this example.  A device type might be in Class III for a decade or more, with all sponsors filing PMAs to support premarket approval.  A new company might file a petition for reclassification into Class II, asking FDA to use the data in the previously approved PMAs to support the reclassification.  The agency might grant the petition using the data and develop “special controls” for the devices also using the data.  After FDA grants the petition, subsequent applicants will be permitted to reach the market using the 510(k) procedure, which is much faster and less expensive.  In this example, section 520(h)(4) allows FDA to use the data in one company’s PMA to reduce the barrier to entry for subsequent applicants seeking to market the same type of device.  This is at least roughly analogous to the role that data exclusivity plays in the drug and biological product setting.