A GLIMPSE OF THE FUTURE?
A Cycle of 510(k) Whitepapers by Nancy J Stark and Kathleen Johnson with
IVD input from Gail Radcliffe
I have to see something visually to really grasp it. It took me nearly a day to reconstruct the timeline presented here and I doubt that it is comprehensive. It's value, I think, is that it emphasizes how events have been happening at CDRH in rapid succession and that respected public groups, such as Consumer's Union, have been complaining for a long time.
In February 2010, CDRH held a two-day "town hall" meeting called Strengthening the Center in which they presented several of their concerns and invited selected industry and user representatives to present theirs. This was followed by two internal surveys—one from the 510(k) Working Group and the second from the Task Force on Utilization of Science in Regulatory Decision Making—published in August 2010.
The 510(k) Working Group's Preliminary Report—What's it to me?
The 510(k) Working Group preliminary report is 120 pages in length. It took me a good four days to make my way through it in detail. My goal in this first cycle of whitepapers is to reduce the report to a manageable chunk of reading, yet to give you a good sense of FDA's probable direction. And, because you are mostly device clinical and regulatory professionals, I intend to speculate on what it may mean in your daily work-life. I've invited my regulatory colleagues, Kathleen Johnson (medical devices) and Gail Radcliffe (in-vitro diagnostics—IVDs), to keep me out of trouble.
The Working Group had seven findings and made a number of recommendations for improvement within those findings, many of them will make our lives easier. But for some firms there will need to be a re-thinking of regulatory procedures if the recommendations are implemented.
Finding 1: There is insufficient clarity with respect to the definition of “substantial equivalence”
The terms "intended use" and "indications for use" are poorly defined
Substantial equivalence begins with determining that a device and its predicate have the same intended use. For IVDs Gail tells me a statement of intended use is required in the electronic submission and is then incoporated into the IVD Decision Summary. It is defined as a description of what the manufacturer intended to measure with a certain test, and typically consists of the following elements: 1) mechanism of assay, 2) what is being assayed, 3) the population being assayed, 4) whether the assay is quantitative or qualitative, 5) whether the assay is an adjunct or a replacement for the gold standard, and 6) what is the next step (surgery, further diagnosis, and the like.)
For non-IVDs, 21 CFR 801.4 defines intended use as the "objective intent of the persons legally responsible for the labeling of the device" and 21 CFR 814.20(b)(3)(i)] defines indications for use as "a general description of the disease or condition...." Kathleen tells me no overt statement of intended use is actually required in a 510(k) submission; instead 21 CFR Part 807.87(e) approaches the subject obliquely saying a 510(k) submission should contain "proposed labels, labeling, and advertisements sufficient to describe ... its intended use...." and 807.92(a)(5) mixes concepts asking for "a statement of the intended use...including a general description of the diseases or conditions...." In practice, the reviewer turns to the labeling and interprets your intention.
The Successful 510(k)
Listen to Janice Hogan and Yarmela Pavlovic present the whys and hows of crafting a successful 510(k) on CD or OnDemand. Their detailed and thorough presentation covers: 1) the legal framework, 2) finding predicates, 3) chosing the right claim, introducing new features, 4) assessing data requirements, 5) software and combination products, 6) dealing with a not-substantially equivalent (NSE) determination, 7) handling device modifications, and 8) seeing the 510(k) from the reviewers perspective. You'll receive the slide deck, video/audio recording, and links to important guidances. Order here.
The meaning of indications for use is much cleaner: for IVDs it is a description of why a patient would use a certain test; for non-IVDs it is a disease or condition for which a device will be used [21 CFR 814.20(i)]. A statement listing the indications for use must accompany every 510(k) submission.
If you get a 510k for an intended use and then you add a new intended use, you have to file a new 510(k) and demonstrate, newly, that your device is substantially equivalent to a predicate. But if you get a 510k for an intended use and then add a new indication for use, you do not have to demonstrate substantial equivalence newly. So hypothetically you can keep adding new indications for use to an existing device and never file a new 510(k).
But suppose you add indications like "relieves cancer pain", "relieves pain from third-degree burns", or "relieves pain during childbirth"? These statements meet the definition of a new indication, but they also introduce new questions of safety and effectiveness and FDA will expect to see a new 510(k).
Working Group recommendation
The Working Group recommends combining the concepts of "indications for use" and "intended use" and then calling the combined concepts "intended use". They further recommend only certain characteristics be included under the term “intended use” so that modifications that are currently considered to be only changes in “indications for use” are not in the future construed as changes in “intended use”.
CDG comments: Thirty years of device submissions and regulation are built upon the phrases "intended use" and "indications for use". To reshuffle the definitions now will only cause new confusion but could also create a large backlog of additional 510(k)s that would not normally have to be submitted.
An unambiguous definition of "intended use" and exactly how it differs from "indications for use" is what CDG believes is required. For non-IVDs we recommend the term "intended use" be defined to mean the purpose (such as treatment or diagnosis) or function (such as provide support or stabilize) of the device and that the intended use be clearly set forth in the 510(k) rather than rely on a reviewer's interpretation of the labeling. We would like to see an "intended use" statement form like the "indications for use" form included in the 510(k) required documentation.
Currently, a new 510(k) is not required for the addition of a new indication that does not raise new questions of safety or effectiveness. If the terms are combined, will we end with ten times more 510(k) submissions every year? At the present time, we look for a predicate device that has the same intended use and similar indications for use; if the terms are combined, what will constitute a legitimate predicate?
Amend the Act to provide agency with authority to consider off-label use
It has been known to happen that a manufacturer identifies an easy-to-get-cleared intended use for a device, knowing full-well the device will be used for another purpose or even promoting it for that purpose. But it happens we live in a legal society where, although we may contemplate a crime, we aren't guilty until we commit it.
Working Group recommendations
The Working Group recommends seeking an amendment to the Food, Drug, and Cosmetic Act (Act) so that FDA can consider the possibility of off-label use when reviewing a 510(k) and even force the submitter to change the intended use.
CDG Comment: FDA doesn't regulate medical practice (read the Belmont Report), so if a physician uses a device off-label it is between him, his professional society, and his malpractice insurance. And CDRH already has the authority to clear a device as "substantially equivalent with limitations", meaning the labeling has to contain a statement that there is no data for likely unintended uses.
But giving a reviewer the authority to force an intended use on the submitter confuses the roles of reviewer and submitter and pretends the reviewer is a mind-reader. And as a submitter, you may be forced into a regulatory pathway you don't want, or forced into gathering literature or clinical data you can't afford.
Questions of safety or effectiveness
Here are the reviewer's steps toward clearing a 510(k) submission: 1) determine if the device under review has the same intended use as the predicate device, 2) if so, determine the device has the same technological characteristics as the predicate device, 3) if not, determine if the technological differences raise new questions of safety or effectiveness. If there are none, the device is substantially equivalent (SE) to the predicate and the 510(k) is cleared.
The Working Group recommends the agency develop a core list of technological characteristics that, when different, are likely to raise new new questions of safety or effectiveness.
CDG Comment: The recommended core list will be helpful even in the early stages of device design and development. Some of the technologies identified may be addressable in literature reviews, but others may require animal studies or even human studies. The ability of the manufacturer to predict the downstream consequences of a design change will allow them to plan better and design more wisely.
Finding 2: CDRH's current practice allows for the use of some types of predicates that may not be appropriate
One quarter of 510(k) submissions rely on predicates more than five years old
Predicate age is not, in and of itself, a concern. Using an older predicate may be appropriate for mature technologies. But in many cases device performance has improved dramatically and older predicates are widely recognized as substandard. The Working Group felt that submitters sometimes cited poorly performing predicates, including predicates that had been withdrawn from the market.
In 2009, 29% of the predicates cited were devices not registered and listed in FDA's Establishment Registration & Device Listing database. In most cases the predicate itself was cleared through the 510(k) process and was not required to have an independent showing of safety and effectiveness. Because the agency does not have the technological infrastructure to trace a device backwards to arrive at the "original" predicate for the intended use the reviewer is unable to trace the predicate's pedigree.
The Working Group recommends developing a guidance to describe when a device should no longer be available for use as a predicate because of safety and/or effectiveness concerns.
CDG Comment: We disagree that any legally approved device should be unavailable for use as a predicate, even if the device has been recalled or removed from the marketplace. If a manufacturer makes strides in improving the performance of an older device, they should not automatically be restrained from using the original device as a predicate. FDA has the authority to require evidence of safety and effectiveness if that is necessary.
First, FDA should do its part by improving the 510(k) database with features like term explosion, predicate lists, and text search capabilities. For example, yesterday I entered a single keyword into the classification database and got three hits, all for Class III PMA devices, and then I entered same keyword into the 510(k) database and got 16 records. The classification database is clearly flawed. With better research tools at our disposal, we believe the quality of 510(k) submissions will improve on their own.
The Act is silent with respect to rescission or modification of device clearance, although the Working Group argues that agencies have inherent authority to reconsider their decisions in certain circumstances such as fraud or error, or to rectify their mistakes.
The Working Group recommends that CDRH issue a regulation to define the scope, grounds, and appropriate procedures for partially or fully rescinding a 510(k) clearance.
CDG comment: Clearly, events of the past few years have shown that CDRH should have the ability to rescind 510(k) clearance. In situations where new information comes to light, or errors have been made, it is reasonable for the agency to revoke a decision. What the industry wants and needs is a level playing field that promotes innovation. What the public wants and needs is an agency that makes impartial decisions to protect the public health.
Multiple predicates, split predicates and bundling
First, let's get some definitions straight. You can argue substantial equivalence using two or more predicates for, say, a multi-parameter monitoring system which combines the functions of several devices. For example, you might combine an ECG and a blood pressure monitor into a single device that performs both functions. Another example of multiple-predicate devices are IVD multiplex assay systems, where two or more analytes are assayed in one well. The "Guidance on the CDRH Premarket Notification Review Program 6/30/86 (K86-3)", also known as the Mohan Memorandum, has allowed this use of "multiple predicates" since 1986.
A "split predicate", on the other hand, is a completely different animal. Here, the intended use of a device is based on one predicate and the technological characteristics are based on another. It's like using a lawn mower to trim a hedge.
"Bundling" is yet a different situation. Bundling refers to grouping many devices into a single submission, with one predicate for each separate device. This is a logical approach to save on user fees for a family of devices. Bundling is described in the guidance "Bundling Multiple Devices or Multiple Indications in a Single Submission" and is accepted by FDA when the devices or indications may be addressed in one review.
The chart on the left shows that the more predicates you use in your submission, the longer the review time. This is logical because the reviewer is compelled to assess the scientific issues associated with each predicate mentioned.
The agency's problem is that it does not have an adequate database to trace the predicates backwards over a number of submissions. They are forced to manually review the text of each 510(k) submission to asses its predicates.
The Working Group recommends the agency develop a guidance describing when multiple predicates are appropriate, disallowing the use of split predicates, and updating its existing bundling guidance to distinguish between multiplex devices and bundled submissions.
CDG Comment: we agree the guidance is necessary and that using multiple predicates makes sense for two or more products combined into one or for multiplex IVD assays; and we agree that bundling predicates makes clear sense for families of similar devices. Furthermore, CDG believes that using split predicates—one for the intended use and one for technological characteristics—is acceptable when accompanied by new data to support safety and effectiveness.
Updated and expanded guidance documents are welcome because they make they make planning regulatory strategies easier and review processes more efficient.
CDG can help write your next 510(k)
Kathleen Johnson is an independent monitor and regulatory consultant for non-IVD medical devices. Gail Radcliffe is an independent regulatory and clinical expert for IVD devices. Both are part of a family of technical professionals who work with CDG, many with 10-15 years experience in their field. CDG has capabilities in in-vivo diagnostic, in-vitro diagnostic, and non-diagnostic medical devices; regulatory engineers; a toxicologist experienced in device biological safety; statisticians; and other professionals who may contribute to 510(k) preparation or clinical trial implementation. We work collaboratively with a point-person on your side so that you are involved in the process every step of the way.
Phone or email us at 773-489-5721 or firstname.lastname@example.org.
Nancy J Stark, PhD
President, Clinical Device Group Inc