Discussing the FDA’s latest draft guidances on 510k submissions with M. Jason Brooke, Managing Member at Brooke & Associates

M. Jason Brooke is an Attorney and Managing Member at Brooke & Associates, where he advises clients of all stages, sizes, and clinical, therapeutic and technological domains to address issues across a broad spectrum - from counselling on compliance with legal and regulatory requirements, providing practical guidance on product-specific market authorization strategies, and serving as an advocate for the development of smart policy. He is also a Board Member and General Counsel at Vasoptic Medical, a medical device company that innovates, develops, and commercializes high-quality, cost-effective products for non-invasive, dynamic assessment of blood flow and vascular status. Previously, he led Regulatory and Quality efforts at AmalgamRx and Navigant (a Guidehouse company).

‍

Jason was kind enough to stop by our virtual fireplace to discuss the FDA’s latest draft guidances on 510k submissions and share tips on how medical device companies could best navigate them.

‍

Thanks so much for joining us to discuss the FDA's latest guidance on 510(k) submissions, Jason. Can you give us a high-level take on these guidances and what they mean for medical device companies?

Absolutely. The FDA's new recommendations focus on what they expect to see in upcoming 510k submissions. They anticipate manufacturers of the subject device (also referred to as a "sponsor" in FDA lingo) to detail how they've adhered to the guidance, specifically when selecting a predicate device. This involves opting for a predicate with well-established testing methods, minimal adverse events, and no device-related recalls. If such a predicate doesn't exist, the FDA expects a comprehensive explanation of risk mitigation linked to the chosen predicate device. 

‍

Essentially, the FDA suggests that you need to consider additional factors during your predicate analysis that are not required currently. For instance, are there any adverse events or design-related recalls linked to the predicate? If so, how are you addressing those risks in your product? 

‍

Keep in mind that information about predicate devices is often not easily ascertainable and adverse event/recall information might not even be relevant to the predicate analysis. Furthermore, sponsors often hide such information to make the predicate comparison difficult despite the legal obligation to make basic technical information publicly available. Specifically, when companies file a 510k, they can choose between submitting a 510k summary, which provides a general product description and its substantial equivalence analysis, and a 510k statement. This choice is significant because the 510k statement approach allows sponsors to hide critical information. Whatsmore, it can negate some of the FDA's recommendations in this guidance, placing the burden on sponsors of the follow-on devices and the FDA to hold sponsors accountable to ensure that required information is available and to facilitate a meaningful predicate analysis. Unfortunately, the FDA often lacks the resources to manage requests for information and has little time for enforcing disputes about whether required information disclosures have been completed. If the Agency can’t enforce compliance with disclosure requirements, much of what they recommend becomes less impactful. 

‍

From your perspective, are there areas where the new guidance doesn’t go far enough? And what would you recommend to improve it?

From a legal standpoint, I find the FDA's "best practices for selecting the predicate(s)" guidance somewhat excessive.

People need to understand that while these guidance documents claim to be non-binding, they’re de facto-binding. The FDA's recommendations can come across as actual requirements, which can be problematic. Particularly with the best practices guidance, there's a potential overreach concerning the information the FDA expects in a 510K submission.

Specifically, the FDA is obligated to adhere to a "least burdensome approach," meaning they should only request the minimum information required to determine a device's substantial equivalence. However, this guidance goes beyond that by asking for details on devices not chosen as predicates. The primary focus should be assessing substantial equivalence against a selected predicate device, not diving into devices that weren't selected. This divergence from the "least burdensome" principle is one of the areas where I believe the FDA is overstepping, and a correction is needed.

‍

Do you know of anyone who's challenged the FDA on this yet?‍

No client has faced this from the FDA in my dealings. I'm working on comments for the public comment period to highlight the FDA's potential overreach, particularly regarding their "least burdensome" mandate. In specific product discussions, the FDA hasn’t made such requests yet. Typically the Agency waits until such a guidance is final before imposing new policies like these on sponsors.

‍

Got it. Switching gears a bit. In the “recommendations for the use of clinical data” guidance, there are four instances where clinical data might be required for substantial equivalence. Can you outline the main changes from past guidance?

The key change is the fourth scenario. The initial three scenarios are standard - differences in indications for use, technological characteristics, and inability to demonstrate substantial equivalence with non-clinical data - the fourth scenario is new and somewhat concerning. It addresses when there are new or increased risks with the predicate device. If the FDA identifies new risks through various sources like adverse events or recalls, they may ask for clinical data for the 510k submission of the subject device. The logic has flaws. Using software as an example, a minor code change can alter the predicate device's functionality. But this doesn't imply that the subject device will have the same issue. Hence, increasing the subject device manufacturer’s burden doesn't make sense based on a predicate device's unrelated issue.

‍

However, in hardware device cases, it might be logical. If a common component across devices is problematic, the FDA might ask manufacturers using that component to address the issue. But in software, this seems like an unnecessary burden, based merely on risks that might not be directly related to a predicate device's issue.

‍

Moreover, the FDA sometimes possesses information that is not publically accessible and can't share it due to confidentiality obligations. So manufacturers could be in a challenging position where the FDA suggests there’s a risk, but the manufacturers can't ascertain the specifics. I don't think that's a fair expectation of the Agency. 

‍

Developing a product or building a medical device company comes with enough unknowns, and you don’t want the FDA holding you to account for more than you already have to deal with.  Let's dive into a scenario where non-clinical testing cannot determine substantial equivalence, and you need to run a clinical trial. Are there any frameworks you give medical device companies to help them think about what kind of clinical study they need to set up to demonstrate substantial equivalence?‍

That’s a tough question because the type of data and statistical analysis needed to show substantial equivalence or safety and effectiveness is product-specific. For instance, AI-enabled products have different FDA expectations than deterministic rule-based algorithms. Both must establish a ground truth and offer a sensitivity-specificity analysis to demonstrate the product's safety and effectiveness. Historically, medical device clinical studies had smaller sample sizes, typically ranging from 50 to a few hundred, to have sufficient power for substantial equivalence and sensitivity-specificity analyses. And I should say, as a footnote, I'm not a statistician, but I know enough to talk somewhat intelligently here. This changed with the introduction of AI-based products, which now often require sample sizes in the thousands due to their probabilistic nature. And so when you're designing a clinical study, it's very important to think about the type of product you're developing and what are the FDA's expectations or concerns that you're going to have to account for when it comes to your clinical study design. 

‍

Makes sense, thanks Jason. One of the interesting aspects of ‘the use of clinical data’ guidance is the FDA’s recommendations around how RWE could constitute “valid scientific evidence sufficient to support the 510(k) submission”. How could medical device companies use RWE to support a submission package?

While I see RWE as vital for pre-market reviews, the FDA's stance is somewhat mixed. Their guidance documents are often driven by policymakers who differ from the technical or device reviewers. Although the guidance might lean towards using RWE to support 510k submissions, the frontline reviewers, with their academic backgrounds in science and engineering, often struggle to shift from a purely academic view to accepting RWE. Despite what the guidance may state, the practical application by reviewers might not align. The FDA should invest in more internal training to broaden its perspective on RWE. Some within the FDA are open to it, but overall, there's much work to be done.

‍

Given the new guidance, any suggestions for medical device companies planning their pivotal clinical trial, particularly for a 510K?

It's crucial to conduct a comprehensive risk assessment, considering clinical safety, user, scenario, and cybersecurity risks. This should shape the trial design, with the FDA emphasizing risk understanding and mitigation. Early risk assessment is key, and any risks found in pilot studies should be addressed in pivotal ones. I can't emphasize enough the importance of consulting with the FDA early. Too often, clients approach me with their clinical data and only then seek FDA feedback, only to find that the FDA takes issue with the generated data and their approach. I recommend the Q-sub process to minimize trial and submission risks, ensuring FDA feedback is integrated into designs.

‍

When do you expect this guidance to be finalized by the FDA? 

There's a public comment period that ends in early December. I strongly encourage everyone to contribute their insights. The more consistent and numerous the comments, the more likely the FDA will incorporate changes. Depending on the volume and nature of feedback, this will determine the timeline for the final guidance versions. Due to its less controversial nature, the clinical data guidance might be released within a year or a year and a half. However, the best practices guidance might take two to three years because of the inherent controversy and necessary internal deliberations.

‍

Understood. Any final thoughts or reflections you’d like to leave our readers with?

Fundamentally, what the FDA has proposed, and in the best practices guidance in particular, reflects a good strategy for, for example, determining a strong predicate candidate. It’s what everyone should be doing and what I do with my clients. Except it goes too far in expecting you to submit all that back-end deliberative information to the FDA for review. I would say companies should follow the best practices laid out in the way they pick which predicate to use. But I don’t think the way the FDA laid out their expectations for what to submit to them is necessarily supported by the law. Companies should be cognizant of that and should make their own decision as to whether or not and to what extent they want to provide that information to the Agency.

I would also say in that context that the law requires demonstrating substantial equivalence to a legally marketed device. It doesn’t require the best or most modern device to be the predicate. The FDA, of course, tries to push this narrative that the most modern, high-tech device on the market should be the predicate. In reality, there are many, many reasons why it’s completely acceptable to choose a product from 1980 as your predicate.

Many people don’t realise that surgical instruments haven’t changed in 100 years. There may be economic reasons why not using the most expensive, modern, high-tech tech product as your predicate is a good strategy for a company, and the FDA shouldn’t be in a position of demanding that every product on the market be a Lamborghini. So, the FDA shouldn’t burden sponsors to choose the most modern product as their predicate. This is an area that companies should be pushing back on the Agency.

‍

That was very well put. Lastly, how can interested companies contact you for further discussions?

Companies can reach out to me on LinkedIn or through my website, devicecounsel.com. I'm always available for consultations and to address questions.

‍

Excellent. Thanks so much for your time, Jason. Take care, and we’ll chat soon!‍

Likewise. Appreciate the opportunity to chat with you guys!

‍