Click here for PDF
FDA announced in late 2017 a plan to modernize medical device regulation. Their effort was aimed largely at the 510(k) process. Specifically, CDRH has been concerned about the use of predicate devices that are more than 10 years old; it says that 20% of 510(k) submissions use predicates more than 10 years old.
The agency also expressed concern over a perceived lack of improvement in some devices with predicates that are more than 10 years old. The solution, according to FDA, is the encouragement of the use of newer predicate devices. Further, FDA plans to eliminate some predicates that are seen as outdated or pose safety issues. It has already eliminated nearly 2,000 devices from the compendium of predicates.
Additionally, FDA intends to create a different 510(k) pathway. This variation would allow applicants to prove substantial equivalence through the use of performance and safety data. FDA expressed confidence that the alternative avenue of approval will result in a “new baseline for safety and effectiveness-and rely on predicate devices that are sometimes decades old as our point of comparison-our premarket review would be based on a contemporary baseline that looks to the future and a baseline that can be updated as technologies advance.” FDA acknowledged that the new criteria could push many applicants toward the de novo approval process.
What all of this means in practice is yet to be determined. FDA argues that the redesigned process will result in more technologically advanced and safer products entering the market. Others contend that the revised process will require more and potentially redundant or unnecessary work that will slow down the clearance process and contribute little to safety. As for the de novo process, it is no secret that FDA has been gently nudging sponsors toward that route.
It is clear that FDA believes the process requires tweaking and the primary driver is a desire to enhance patient safety in response to criticism that the process allows many devices to slip through that either have to be altered or recalled, or which actually resulted in harm or death to patients.
What all of this fails to address is the growing challenges to expeditious and effective clearance of 510(k) devices. We see this in the context of software as a device and hardware with a software component. FDA has had to contend with an explosion of technological progress. The consequence was that the agency desperately needed software experts. FDA pays much less than industry, and such expertise is in high demand and commands top dollar in the private sector. The shortage of people who can fill these spots does not augur well for FDA’s ability to manage the clearance of products that incorporate ever more sophisticated technologies.
Part of this is the result of the sweeping and rapid pace of innovation that would challenge almost any regulatory body. This is different, in that FDA is scrambling to keep pace already and is facing a talent (and some would say) a lack of funding sufficient to add qualified headcount.
Still, FDA is trying to bridge the gap. It has undertaken several other programs in the past several years that are designed to assist FDA in its efforts to both streamline the approval and enhance the safety of more technologically advanced programs.
The Software Precertification (Pre-Cert) Pilot Program, as outlined in the FDA’s Digital Health Innovation Action Plan [PDF], will help inform the development of a future regulatory model that will provide more streamlined and efficient regulatory oversight of software-based medical devices developed by manufacturers who have demonstrated a robust culture of quality and organizational excellence, and who are committed to monitoring real-world performance of their products once they reach the U.S. market.
This proposed approach aims to look first at the software developer and/or digital health technology developer, rather than primarily at the product, which is what FDA currently does for traditional medical devices. Because software products can be adapted to respond to glitches, adverse events, and other safety concerns quickly, the FDA wants to establish a regulatory framework that is equally responsive when issues arise to help ensure consumers continue to have access to safe and effective products. In the new Pre-Cert program, the FDA is proposing that software products from precertified companies would continue to meet the same safety and effectiveness standard that the agency expects for products that have followed the traditional path to market.
FDA last year rolled out the Pre-Cert Program. As FDA described it, Pre-Cert 1.0 is the first version of the program that will be available for pilot testing. It will be limited to Pre-Cert pilot participants’ software as a medical device (SaMD). FDA is offering the voluntary program to manufacturers of software in a medical device (SiMD), and other software that could be considered accessories to hardware medical devices. At last check, nine companies were enrolled in Pre-Cert.
The 2019 test plan addresses how the FDA will test the model with certain SaMD De Novo Requests from Pre-Cert pilot participants and internally using selected previously-reviewed SaMD submissions. During this first phase, FDA will assess whether the Excellence Appraisal and Streamlined Review components together produce an equivalent basis for determining reasonable assurance of safety and effectiveness for a software product prior to its introduction to the market, as compared to the traditional paradigm.
FDA says that the pilot program will allow FDA to:
- benefit a participating organization based on its “precertified” status by offering the ability to participate in a streamlined premarket review and opportunities to collect and leverage real-world postmarket data;
- leverage and use information and data from all available sources allowing FDA and SaMD manufacturers to be more efficient and streamlined without compromising safety and effectiveness of SaMD products, and
- enable a modern and tailored approach that allows software iterations and changes to occur in a timely fashion.
Another program that has received attention is the Proposed Regulatory Framework for Modifications to Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD).
FDA’s existing paradigm of medical device regulation was not designed for adaptive artificial intelligence and machine learning technologies. Under FDA’s current approach to software modifications, the agency takes the position that many of these artificial intelligence and machine learning-driven software changes to a device may need a premarket review.
The ideas described in the discussion paper published by FDA (comments are due by June 3, 2019) leverage practices from FDA’s current premarket programs and rely on the FDA’s benefit-risk framework, risk management principles described in the software modifications guidance, and the organization-based total product lifecycle approach (also envisioned in the Pre-Cert program).
In its framework, FDA introduces a “predetermined change control plan” in premarket submissions. This plan would include the types of anticipated modifications—referred to as the “Software as a Medical Device Pre-Specifications”—and the associated methodology being used to implement those changes in a controlled manner that manages risks to patients —referred to as the “Algorithm Change Protocol.”
FDA will expect a commitment from manufacturers on transparency and real-world performance monitoring for artificial intelligence and machine learning-based software as a medical device, as well as periodic updates to FDA on what changes were implemented as part of the approved pre-specifications and the algorithm change protocol.
The agency says that the proposed regulatory framework could enable FDA and manufacturers to evaluate and monitor a software product from its premarket development to postmarket performance. According to CDRH, this potential framework would allow for FDA’s regulatory oversight process to embrace the iterative improvement power of artificial intelligence and machine learning-based software as a medical device, while assuring patient safety.
Industry has cautiously welcomed these initiatives, but some industry believe the changes will be time and cost intensive, with what is perceived as a questionable concomitant benefit on the safety side. Another concern is that FDA is using a precautionary approach, which will result in more and more responsibility passed along to sponsors.
Clearly, FDA grasps the urgency engendered by the rapid pace of change. The concern here is that much of this may be aspirational without an infusion of funding and talent. Congress has always managed to find a way to take care of FDA’s funding needs. That willingness may be tested by the increased needs of the agency just to keep up, let alone remain ahead of the curve. The other concern is that FDA does not have the authority to do some of the things it would like to do. If that proves to be the case, then statutory changes will be needed.
The reality is that this process will depend on a greater degree of industry-agency cooperation than we have seen in years past. That process will bring tensions, but there is no escaping the need for a cooperative approach under the circumstances. It is going to be industry’s task to help educate the agency about new technologies. But that brings us right back to the underlying problem of a dearth of talent. FDA has taken the position that it has the resources to keep up, and that the programs described above will make the entire clearance process more transparent, more effective and safer, as well as adaptable to changing technology. While FDA has made a real effort to keep up with advancements, it remains hamstrung by the realities of government service. Lack of competitive compensation will lead to difficulties recruiting talent, which threatens to stymie FDA’s efforts to modernize the device approval process.
With the pace of change continuing to accelerate, we are likely to know sooner rather than later whether the agency will be able to keep up. It must, or else there will be a slowdown, rather than a streamlining of the clearance process. The question remains whether there is enough congressional support for a sweeping talent infusion. That will cost a lot of money, but that impediment must be addressed. The cost could be a slowdown of clearances of critical products that will save lives, and improve the quality of life for millions.