Posts
On August 5th, the Center for Devices and Radiological Health (CDRH) announced its preliminary findings on the need to update the Premarket Notification ("510(k)") process. An overview of the evaluation process can be found at http://www.fda.gov/AboutFDA/CentersOffices/CDRH/CDRHReports/ucm221069.htm, while the preliminary (read: almost certain to dramatically change) findings themselves are at http://www.fda.gov/downloads/AboutFDA/CentersOffices/CDRH/CDRHReports/UCM220784.pdf. Not linked, but available on FDA's website, is a second part of the report dealing with internal recommendations on the burden of scientific evidence that should be required by FDA's Office of Device Evaluation.
First, a recap of the current situation. In late 1976, Congress passed the Medical Device Amendments, and followed up a few years later with the Safe Medical Device Act. These laws, which for the first time brought medical devices under FDA's direct pre-market authority, designated two types of FDA authorizations: the Premarket Approval (PMA) process, and the Premarket Clearance ("510(k)"), named after the corresponding section of the original Act. PMAs are the medical device equivalent to the drug world's NDA, and correspond to the European Union's Design Dossier submission (in other words: big, complicated, and expensive), while 510(k)s are a somewhat similar idea to a generic drug's ANDA, and somewhat similar to the E.U.'s Technical File (simpler, quicker, and consisting mostly of documentation that the manufacturer would have anyway). There is also a "Product Development Protocol" route that parallels the PMA process, but it is almost never used.
In theory, a PMA is always required for a Class III device, and a Class III device is always the most serious type of medical device: complex, new and novel (i.e. unproven), can easily injure or kill someone. And also in theory, a 510(k) is for a Class II device (and sometimes for certain Class I devices) and a Class II device is one that has a markedly lower risk to public health, especially with regards to familiarity: substantially equivalent to a device that is already on the market, the idea being that a lot is already known about that type of device.
Makes sense, right? The idea is that if you have a relatively simple device or a device whose design has already been proven through extensive experience with similar devices, you don't have to file eighteen binders of a thousand pages each. There is also a somewhat esoteric legal distinction: technically speaking, FDA only "approves" a Class III device; in the case of a Class II device, FDA merely agrees that the device is in fact Class II and that basic safety and design information is properly documented.
But unfortunately, Congress wrote a requirement into the original law that, while it may have made sense at the time, is a major burden to FDA and industry (and drives European and Japanese companies crazy): for a device to be considered Class II, and therefore be eligible for the simpler submission, it must be equivalent to a device that was already on the market when the law was written. In 1979 that may have made sense, but now more than a quarter of a century has gone by. Look at it this way: in 1979, you could sell your medical device by showing it was equivalent to something marketed in 1975 (four years earlier), but today, you cannot sell your medical device by showing it is equivalent to something marketed in 2000 (ten years earlier). Obviously this is a logical contradiction, but FDA cannot do anything about it directly because it is the law.
In response, companies routinely come up with crazy-quilt arguments that their device is similar to device A marketed in 2000, which is similar to device B marketed in 1990, which is kinda sorta like device C marketed in 1975. I should add that this doublespeak logic also can go in the opposite direction: I have been on projects developing devices that absolutely, clearly were Class II (injectable wrinkle remover), but due to the thread of prior "predicate" devices—marketed for much more serious indications—this Class II device became by default Class III.
It is in reference to this last situation (which happens quite often, actually) that FDA invented the "De Novo" submission process. This is an attempt by FDA to clone the Class IIa/Class IIb Notified Body review process used by the Europeans. The idea is that some (actually, make that "many") medical devices that cannot trace back to a pre-1977 predicate device simply do not pose a serious health risk (but still are not so obvious as to sidestep risk management processes such as Design Controls, which would have classified the device as Class I like a tongue depressor). But the De Novo process is convoluted and rarely utilized, and even then, it only addresses part of the overall problem just described. CDRH's analysis linked above is FDA's latest attempt to come up with a science-based approach to intelligently classify and approve or clear medical devices, rather than ODE and firms coming up with ad hoc and frankly often silly daisy-chains of pre-1977 pedigree.
As is often the case (look at the recent debacle over electronic cigarettes!), FDA tries to do the right thing, but finds itself hamstrung by the constraints Congress wrote into a law. For example, in a recent post on this blog I expressed dismay over FDA launching new initiatives regarding software as a medical device because I feel it is driving FDA outside of what the law allows…and opening the door for endless lawsuits and reversals.; and this latest initiative still cannot ultimately address the need to base the level of device scrutiny on the actual risk it poses, and not on some arbitrary date (although similarity to an existing device can certainly be used as a mitigating consideration, and indeed that is precisely the European regulatory model).
Again in this case, the predicate law (which FDA has no choice but to obey, even if that means finding technical loopholes and workarounds) is hopelessly out of date and needs to be revised to place the U.S. device industry on a firm footing with the rest of the world and base device review and approval on science (what is the potential risk of this device?) and not bureaucracy (did Jimmy Carter use this device?).
In the FDA Modernization Act, Congress directed FDA to harmonize regulations with those of other countries. The result was the Quality System Regulation, universally recognized as the best GMP regulation in the world (and one that the Center for Drug Evaluation and Research is inching towards adopting in their own fashion). Perhaps it is time that Congress do a repeat performance; this time, to allow—or order—FDA to bring the device approval process into the next millennium.
First, a recap of the current situation. In late 1976, Congress passed the Medical Device Amendments, and followed up a few years later with the Safe Medical Device Act. These laws, which for the first time brought medical devices under FDA's direct pre-market authority, designated two types of FDA authorizations: the Premarket Approval (PMA) process, and the Premarket Clearance ("510(k)"), named after the corresponding section of the original Act. PMAs are the medical device equivalent to the drug world's NDA, and correspond to the European Union's Design Dossier submission (in other words: big, complicated, and expensive), while 510(k)s are a somewhat similar idea to a generic drug's ANDA, and somewhat similar to the E.U.'s Technical File (simpler, quicker, and consisting mostly of documentation that the manufacturer would have anyway). There is also a "Product Development Protocol" route that parallels the PMA process, but it is almost never used.
In theory, a PMA is always required for a Class III device, and a Class III device is always the most serious type of medical device: complex, new and novel (i.e. unproven), can easily injure or kill someone. And also in theory, a 510(k) is for a Class II device (and sometimes for certain Class I devices) and a Class II device is one that has a markedly lower risk to public health, especially with regards to familiarity: substantially equivalent to a device that is already on the market, the idea being that a lot is already known about that type of device.
Makes sense, right? The idea is that if you have a relatively simple device or a device whose design has already been proven through extensive experience with similar devices, you don't have to file eighteen binders of a thousand pages each. There is also a somewhat esoteric legal distinction: technically speaking, FDA only "approves" a Class III device; in the case of a Class II device, FDA merely agrees that the device is in fact Class II and that basic safety and design information is properly documented.
But unfortunately, Congress wrote a requirement into the original law that, while it may have made sense at the time, is a major burden to FDA and industry (and drives European and Japanese companies crazy): for a device to be considered Class II, and therefore be eligible for the simpler submission, it must be equivalent to a device that was already on the market when the law was written. In 1979 that may have made sense, but now more than a quarter of a century has gone by. Look at it this way: in 1979, you could sell your medical device by showing it was equivalent to something marketed in 1975 (four years earlier), but today, you cannot sell your medical device by showing it is equivalent to something marketed in 2000 (ten years earlier). Obviously this is a logical contradiction, but FDA cannot do anything about it directly because it is the law.
In response, companies routinely come up with crazy-quilt arguments that their device is similar to device A marketed in 2000, which is similar to device B marketed in 1990, which is kinda sorta like device C marketed in 1975. I should add that this doublespeak logic also can go in the opposite direction: I have been on projects developing devices that absolutely, clearly were Class II (injectable wrinkle remover), but due to the thread of prior "predicate" devices—marketed for much more serious indications—this Class II device became by default Class III.
It is in reference to this last situation (which happens quite often, actually) that FDA invented the "De Novo" submission process. This is an attempt by FDA to clone the Class IIa/Class IIb Notified Body review process used by the Europeans. The idea is that some (actually, make that "many") medical devices that cannot trace back to a pre-1977 predicate device simply do not pose a serious health risk (but still are not so obvious as to sidestep risk management processes such as Design Controls, which would have classified the device as Class I like a tongue depressor). But the De Novo process is convoluted and rarely utilized, and even then, it only addresses part of the overall problem just described. CDRH's analysis linked above is FDA's latest attempt to come up with a science-based approach to intelligently classify and approve or clear medical devices, rather than ODE and firms coming up with ad hoc and frankly often silly daisy-chains of pre-1977 pedigree.
As is often the case (look at the recent debacle over electronic cigarettes!), FDA tries to do the right thing, but finds itself hamstrung by the constraints Congress wrote into a law. For example, in a recent post on this blog I expressed dismay over FDA launching new initiatives regarding software as a medical device because I feel it is driving FDA outside of what the law allows…and opening the door for endless lawsuits and reversals.; and this latest initiative still cannot ultimately address the need to base the level of device scrutiny on the actual risk it poses, and not on some arbitrary date (although similarity to an existing device can certainly be used as a mitigating consideration, and indeed that is precisely the European regulatory model).
Again in this case, the predicate law (which FDA has no choice but to obey, even if that means finding technical loopholes and workarounds) is hopelessly out of date and needs to be revised to place the U.S. device industry on a firm footing with the rest of the world and base device review and approval on science (what is the potential risk of this device?) and not bureaucracy (did Jimmy Carter use this device?).
In the FDA Modernization Act, Congress directed FDA to harmonize regulations with those of other countries. The result was the Quality System Regulation, universally recognized as the best GMP regulation in the world (and one that the Center for Drug Evaluation and Research is inching towards adopting in their own fashion). Perhaps it is time that Congress do a repeat performance; this time, to allow—or order—FDA to bring the device approval process into the next millennium.
No comments:
Post a Comment