On August 4, 2010
U.S. FDA issued two comprehensive evaluations addressing changes
underway that will impact the manner in which the agency will
regulate the premarket notification process for medical devices,
i.e. the 510(k) process. This recent announcement is essentially a
progress report on the work of two committees formed in September
2009 to address long-standing concerns about the 510(k) regulatory
track affecting the great majority of medical devices. The workings of these two committees has
generated keen interest among device suppliers, and their reported
work-in-progress has already created a firestorm of concern, much
of it unwarranted in my view.
The August
news release can be found here, and that
document contains links to additional detail: Highlights of
recommendations made to date and detailed reports from each of the
two committees. Your
regulatory staff and/or advisors should have already seen these
documents and digested their contents; this is important because
the time for comment on these preliminary recommendations is
now, before they are converted to specific guidance and
regulations.
My purpose
here is to recount the highlights of the ten recommendations
published in August by the Director of CDRH, Dr. Jeffrey Shuren,
and to offer my views on these recommendations. To fully understand any of these
topics I suggest you consult the full published detail. But here's the 10,000-foot
view:
1. Streamline the premarket pathway
for lower-risk novel devices.
This
recommendation recognizes the current deficiencies of the tortuous
de novo pathway
available to devices that cannot use the 510(k) pathway for lack of
a predicate device but whose risks don't warrant a PMA
pathway. Reform of the de novo pathway is long
overdue. Let's see if FDA
can really come up with an improved protocol. Kudos for
trying.
2. Enhance science-based
professional development for CDRH staff.
Given FDA's
statutory responsibilities, this is a no-brainer. It's all about personnel
qualifications and money.
FDA funding has been given short shrift before. Is there any real hope
now?
3. Establish a network of external
experts to better inform the review of cutting-edge
technologies.
This seems to be a
worthy idea. What could be
wrong with building a bigger tent of expertise? Well, are these experts going to
be compensated? How can we
be assured that they will be objective and not influenced by
industry relationships? FDA
says that these folks would not serve in an "advisory
capacity". What does that
mean? I really like this
idea but I think implementation will be fraught with
difficulties. Good
luck.
4. Increase the predictability of
510(k) data needs by establishing a new "class IIb".
This
recommendation has received the most attention so far; many call it
PMA-lite. It actually makes
great sense as the goal is to make certain 510(k) requirements and
processes more predictable by developing guidance for certain
subsets of class II devices "for which clinical or manufacturing
information would typically be necessary to support a substantial
equivalence determination".
Of course the devil's in the details, especially defining the
subsets and adequately defining the additional data
requirements. If this works
it will at least reduce the occurrence of nasty surprises, but I
don't think it will measurably reduce review
times.
5. Create a new "Notice to
Industry" tool to more rapidly communicate changes in premarket
expectations.
Of course this can
be nothing but good for suppliers, and it's probably long
overdue. But which industry
sectors will receive these notices? Will all registered suppliers
receive all the notices? If
not, how will the industry sectors be defined, and who decides
which suppliers are in which sectors? How does one deal with a
supplier who is contemplating entering a new sector? These details can be worked
out. This is a good
idea.
6. Clarify the meaning of key terms
in the 510(k) "substantial equivalence" review standard to improve
the consistency, transparency, and timeliness of the review
process.
I've never been a
fan of "substantial equivalence", and it's becoming ever more
meaningless. There are
better ways to skin this cat. But I guess we're stuck with it
for now. FDA has had many
opportunities to clarify its definitions over many years. Why should we believe it can be
successful now? I'm
pessimistic that this recommendation can be implemented
satisfactorily.
7. Establish a Center Science
Council as a new governance model to assure quality and consistency
in CDRH's science-based decision making.
Fancy,
bureaucratic language.
Worthy goal. Hard to know if it could really make a
difference. I'm
skeptical.
8. Require the up-front submission
of more complete safety and effectiveness information to support
the review of 510(k) devices.
This one has
suppliers the most worried.
CDRH would revise existing regulations to "explicitly require
510(k) submitters to provide in their 510(k)s a summary of all
scientific information known or that should be reasonably known to
the submitter regarding the safety and/or effectiveness of the
device under review". On
the surface this makes great sense as it efficiently informs FDA
reviewers of available data. But some of that data may be
proprietary to the submitter. Will that data later be made
public? If so, that's a big
problem.
I have a better,
albeit much more radical, solution. Forget about demonstrating
device effectiveness. Focus
exclusively on demonstrating safety. Let clinicians determine device
effectiveness as evidenced by postmarket refereed publications over
time. And, if a device is
ultimately proven ineffective in its intended use, FDA could
require its removal from the U.S. market. Exceptions to this solution
would have to be made for devices used exclusively by patients
without clinician involvement.
9. Create a searchable online
public database to provide more detailed, up-to-date medical device
information to industry, the health care community, and
patients.
I really like this
idea. The resulting
database would be helpful to all. Can proprietary information be
appropriately omitted? Are
the FDA's IT resources up to the job, and how long would it
take?
10. Clarify CDRH's 510(k) rescission authority
and the circumstances under which a device should not be used as a
predicate.
Makes sense to
me. For example, previously
cleared devices that are proven to be unsafe should not be cited as
predicates. It should be
possible to define the agency's rescission authority without
stepping on industry toes too much.