Paladin Medical’s Elaine Duncan explains the strategic requirements that go into a regulatory assessment for product clearance in the US, the determination of a predicate and how it’s an important process for premarket application.
The US FDA premarket application process for Class 2 devices has always seemed complex when compared with the EU’s Medical Devices directive. In contrast to European regulations, the rules in MED DEV, the US Food and Drug Regulations, use predicates. Although the concept of comparison with a prior and well-established product is certainly part of the MED DEV Clinical Evaluation Report, the practice of establishing a substantially equivalent argument under US regulations can be a difficult problem to solve. Recent interpretations by FDA can be confusing: sometimes a predicate isn’t a predicate.
The first step is to determine the device classification; we start by applying a keyword search in the database to determine how the applicant device may be regulated. Typing ‘catheter’ in the device field, for example, will yield 185 different, three-digit device codes. These are called ‘pro codes’ and this category is important in order to match your device to a predicate. Using the pro codes and the 510(k) database, it is usually possible to find product clearance and resolve the proper classification for the applicant device by comparative analysis.
Pro codes relate to the US Food and Drug Regulation and are separated into Class 1, 2 and so on. A urethral catheter, for example, could be Class 1, or an ablation catheter could be Class 3. Thus, we see the regulatory impact of ‘intended use’. The classification codes are intended to reflect the risks associated with the intended use of the device as understood at that time, and have changed and expanded along with medical technology. As a guiding principle, something that is commonly used in a general procedure where the risk is understood by the user is most likely to be considered Class 1.
If it carries an inherent risk, such as intravascular penetration, however, the designation may be Class 2 or 3.
Even during the classification process, requirements for premarket notification do not always follow the classification level. The majority of Class 1 devices are exempt from 510(k); Class 2 devices need a 510(k) – but not always. The determination process became more complicated when many more Class 1 devices received an additional designation as exempt from the 510(k) premarket notification process. Subsequent revisions to the regulations attempted to exempt most Class 1 devices from the 510(k) regulations and FDA has begun to include more Class 2 devices in the exemption subcategory.
The 510(k) exemption provision has always had an exception: FDA regulations explain that the 510(k) exemption “is only to the extent that the device has existing or reasonably foreseeable characteristics of a commercially distributed device within that generic type”. The regulations detail that, despite the exemption, a 510(k) is required when “(a) the device is intended for a use different from the intended use of a legally marketed device in that type”, or “(b) the modified device operates using a different fundamental scientific technology than a legally marketed device in that type”. Following the various new guidance documents from FDA, we can better appreciate that a new risk from a new technology; even for a nominally Class 1 product, could cause even a 510(k) application to be denied.
It is difficult to be precise about the term ‘predicate’ in practice with FDA because, in guidance documents, the term almost invariably denotes ‘legally marketed device’, though not all devices are legally marketed, and not all legally marketed devices are predicate to the applicant device. To be a predicate, the legally marketed product must also be ‘substantially equivalent’ (SE) to the applicant device. So a predicate device for an applicant device is legally marketed in the US and substantially equivalent to the applicant device. Hence, the paradox – and we must be mindful of the methods FDA will use to determine equivalence.
A new medical device still may not qualify for 510(k) application if FDA has already designated the type of product as either requiring a humanitarian device exemption or premarket approval (PMA). Sometimes, FDA will call for a PMA for an existing product that has been previously cleared under the 510(k) rules – most small companies seeking a first-time US market toehold hope to avoid the PMA process. There is an intermediate process for a new device not allowed under a 510(k) clearance known as ‘de novo’.
Now that the applicant has one or more candidate predicates, the next step is to conduct a detailed characteristic comparison. Reviewing predicate submission summaries from the FDA database can be informative and it is possible to acquire full 510(k) documents through the Freedom of Information Act or commercial purveyors. Understanding how the candidate predicates compare with the applicant’s own input requirements (or essential requirements) is the best foundation for the comparison matrix.
FDA Guidance, ‘The 510(k) Program: Evaluating Substantial Equivalence in Premarket Notifications’, redefines how FDA will establish substantial equivalence. Although the introduction states that the guidance is not intended to implement significant policy changes, the reality is that FDA did just that. The document, and others issued between now and 2014, combine to have a substantial impact on how we prepare and submit premarket notifications.
The first task is to determine that an applicant device is intended for the same use as a predicate device. Next, the evaluation addresses the question of ‘same technological characteristics’. Often, an applicant’s device will have some new feature, or a new material, pushing the device into the ‘yes’ category for ‘new technology’. Does this new technology raise new questions concerning safety and effectiveness?
Most companies developing a new product take such modern improvements for granted as part of the process, or may not appreciate how antique the predicate devices truly are. However, failing to assess this very important question adequately – in accordance with current FDA practice – stops many devices in their premarket-submission tracks.
Paladin Medical recommends a detailed comparative analysis to a proposed predicate (assuming confirmation that the candidate predicates are legally marketed in the US); it is de rigueur, especially for medical device manufacturers approaching FDA with an improved device seeking substantial equivalence.
Along with the comparative analysis of the predicate features compared with the applicant device, submissions should provide a hazard analysis focusing on the risks associated with the differences. Conducting a focused analysis on how the differences could introduce a new hazard to the patient or user is extremely important for the 510(k) strategy. A focused assessment of the differences between the predicate and applicant device should begin with the known risks of the predicate device, and determine if the new technological differences introduce truly new risks. The new technology could introduce a different failure mode or probabilities, but how do these differences change the risk?
Frequently, 510(k) applications discuss differences between predicate and applicant devices, and, generally, the differences don’t introduce new risks, but such conclusions may prove insufficient. Using the example of material change, the applicant may see that the same risks were already present with the predicate, thus there is no new risk. A deeper hazard analysis, however, would reveal that the material change could introduce a new failure mode and the different failure mode should be analysed for any different potential harm. If the applicant does not conduct this analysis and demonstrate the mitigation process for the different potential failure mode, FDA could perceive the risk as new and the premarket application could be rejected.
During a device-characteristic analysis, it might be observed that a new technology is already in use in a predicate product but not in the primary predicate. Recall that to select the applicant device predicate, the primary criteria was to have the same intended use. FDA does not accept a lot of predicates: one for the material, another one for colourants and another for accessories. Prolific use of multiple predicates has been viewed as the source of unanticipated consequences due to device differences. In some situations, there is insufficient analysis of how these multiple differences could combine in a different use to create a perfect storm.
Following the 2014 guidance, the revised 510(k) decision flow cart asks if there are “different questions of safety and effectiveness”. If there is a new or different risk, the applicant device is not substantially equivalent and the application fails.
The 510(k) flow-chart requires a review of the “scientific methods for evaluating the characteristic’s effects on safety and effectiveness”. The FDA standards database details which standards, sorted by pro code, that FDA recognises and any exceptions to that recognition that it has published.
However, standardised tests are seldom conducted only on the applicant product because if the predicate hasn’t also been tested under the same standard, using the same protocol and at the same laboratory, it may be impossible to demonstrate that the applicant device is equivalent to the predicate. Sometimes, an applicant may find published information about the predicate device performance. Early planning to conduct side-by-side laboratory testing is the ideal method for obtaining the comparative evidence required for an application.
What if the applicant device does have a different risk and no standardised testing or reference predicate is able to buffer the determination of new risk? FDA has the option to consider an alternative regulatory clearance short of requiring a PMA: if the risks can be mitigated through special controls and if the type of device to which the applicant product would be comparable is not already designated as Class 3. FDA issued in 2014 the de novo classification process (evaluation of Automatic Class 3 Designation) that provides a pathway to Class 1 or Class 2 for devices for which there is no legally marketed predicate device. Under these circumstances, the applicant product is known as a new device type, which means that there is no existing pro code. The FDA Modernization Act allowed for the alternative to an automatic Class 3 designation, creating the de novo process for a lower-risk device. A subsequent amendment to the Federal Food, Drug & Cosmetic Act allowed a direct path to de novo, rather than requiring that a device fail the 510(k) application first. FDA encourages a pre-sub meeting prior to a de novo application.
A successful de novo application is the summary of known and potential threats to health, and this is greatly enhanced when these can be put in the context of known risks from other alternative devices or therapies. This is why it could be critical to the success of either a 510(k) device application or a de novo application to have done the basic differences risk assessment homework first. A risk/benefit assessment is important for a de novo application.
This helps the applicant understand how important it is to present the extent of probable risks in the application. Where a new technological difference has introduced the new potential risk, it should be straightforward for development engineers to assess the differences in the hazard resulting from new technology.
Understanding how a new technology can affect the risk/benefit profile of an applicant product helps to prepare strong premarket documentation, suitable for either pathway. Where bench-testing cannot demonstrate satisfactory risk mitigation, applicant companies should be prepared to conduct usability studies or clinical trials.
Applicants to FDA must become experts on the risks associated with their product, considering every aspect of its intended use, including how alternative technologies or differences from a predicate device could present a different risk profile.