We’ve been hearing for several years about an FDA proposal to overhaul its device regulatory framework with ISO 13485, potentially the most ambitious FDA undertaking in a quarter century. Those who don’t follow these things might find the subject terminally boring, but such a change could be a massive headache for industry, although it doesn’t have to be if the FDA can get the temperature of this regulatory porridge just right.
To recap, ISO 13485 is a med tech-specific quality management system standard assembled by the International Standard Organization (ISO), and the FDA parallel is the Quality System Regulation, otherwise known as Part 820. To date, much of the thinking is that the FDA will propose a more drastic overhaul than industry is willing to tolerate when it comes to aligning Part 820 with ISO 13458, particularly given the ongoing overhaul of med-tech regulation in the European Union. I will confess that I’ve expressed skepticism about all this in chatter on LinkedIn, but there may be a way for the agency to finesse this in a way that keeps everyone happy, including any FDA audiences that are pressuring the agency to do a little harmonizing.
ISO 13458 is an offshoot of ISO 9001, and as is true of all ISO standards, is subject to revision every five years. The current edition of 13485 was published in 2016 whereas Part 820 has stood in its current form for about 25 years. Obviously, ISO had a chance to revise this standard last year and passed on the opportunity, but there’s an odd twist to that part of the story.
In 2019, the ISO technical committee charged with keeping tabs on 13485 noted that the standard had been designated not just a management system standard (MSS), but a sector-specific MSS. This means that roughly 10 pages of text would have to be added to 13485 to make it align with ISO high-level structures (HLS) upon the next revision. This requirement had not been imposed on MSS standards until after the 2016 version of 13485 was already out the door, however.
The problem now is that any further revisions would have to come with that extra 10 pages, and the members of that ISO technical committee seemed to believe this requirement might make a crippling mess of 13485. This is made clear by the title of the ISO blog, which is: “Will ISO 13485 remain useful for its regulatory use purpose?”
The committee said ISO does have some exit ramps, such as removing 13485 from the MSS list, but the standard is still listed as an MSS as of February 2022. The committee had also mentioned standing pat with 13485’s existing content as a means of sidestepping the HLS problem, which appears to be the option they invoked last year. Whether this was the principal driver of the decision to leave 13485 as is, however, is not at all clear.
But here is the question for companies doing business in the U.S: Does the MSS headache even matter?
One way of looking at the FDA’s interest in harmonization is that the agency’s staff really believes there is a need for an extensive rewrite of Part 820. The problem with that approach is that there are tons of guidances that would have to be revised. Beyond the question of guidance alignment, it’s tough to imagine anyone at the agency is interested in all the snorting and head-butting that would accompany a drastic rewrite. Plus, any such rewrite would have to be a completely de novo exercise because all ISO standards are copyrighted, and consequently the FDA cannot simply copy and paste from 13485.
Rather than jump into a complete overhaul, the FDA could just make a few tweaks here and there to clean up some the more awkward structural elements in Part 820 without making any substantive changes. Corrective and preventive action (CAPA) would be a great place to start, and even Kim Trautman, who authored the Quality System Regulation, has said CAPA could be improved by splitting it into corrective action and preventive action. This is a change the FDA could make without exciting a ton of pushback from industry, partly because these two functions are addressed separately in 13485, and partly because device makers are sick of hearing about CAPA in FDA warning letters.
No doubt industry has a list of regulatory pet peeves they’d like to see addressed in a revised Part 820, but it’s not clear just how much conversation has passed between the FDA and device makers about an industrial wish list. And don’t get me wrong, a radical overhaul would give people in my line of work a lot to write about without having to learn a completely new skillset, but entertaining the media with a slow-motion train wreck really should not be the endgame here.
Obviously, I’m speculating, but it’s tough to imagine that I’m the only one who’s idly ruminating about this. Incidentally, the FDA is still dragging itself out of a controversy-laden pandemic and has yet to extract itself from a device user fee quagmire, so maybe there’s already a limited appetite at the agency for yet another ennui-inducing slog. That doesn’t mean the FDA won’t try something impracticably ambitious, but the better move might be to float a few targeted fixes to 820 that avoid a bunch of controversy, and just let it go at that.
If everyone is happy with a modestly edited Part 820, all the parties to this deal could ignore the next do-over of 13485, particularly if the HLS nightmare is part of that package. Either way, the FDA’s tweaks to Part 820 would largely satiate any industrial demand for harmonization, especially given that harmonization of artificial intelligence regulation might make for a more compelling narrative going forward. While there may be a good reason to scrap Part 820 and start all over again, a modest and not-too-burdensome tweak would allow the FDA to serve up a crowd-pleasing bowl of harmonization that’s just right without having to visit those three agitated bears again for a very, very long time.