Software in Medical Devices, by MD101 Consulting

Its application was planned for May 2020, in the middle of the lockdown. The 2020/561 regulation had to be voted in April 2020 in emergency by the European Parliament to postpone this application by one year. As a result, the MDR came into force in May 2021, but the respite was short-lived. At the beginning of 2022, Russia invaded Ukraine, which added new difficulties to the disorganization of the supply chains that had already been disrupted during the lockdown.

Manufacturers are now struggling with shortages of raw materials, components, and skyrocketing prices, while simultaneously investing in a costly transition to MDR that involves R&D, quality assurance, and regulatory affairs teams. It could not be worse in terms of timing. Manufacturers, caught between a regulatory rock and economic and geopolitical hard environment, are weakened, or even, for the weakest, will have to close down.

But this unfavorable situation is in addition to the structural defects that have been present since the beginning of the MDR. The list of missed appointments and unmet milestones is long. If we go into the details of the regulation, we can quickly identify several major problems:

1. The delay in notification of candidate certification bodies. This is certainly the most acute and well-known bottleneck of the MDR, not to mention the IVDR regulation where these problems are multiplied tenfold. Notified Bodies (NBs) are overloaded and take on new cases sparingly. They are not even able to respond adequately to the demands of their current clients. Without trained and qualified auditors, they will not be able to handle the coming wave of certifications before 2024.
2. The delay in publishing the list of Medical Device Coordination Group (MDCG) guides. These guidelines are being released sparingly, and the timetable set by the MDCG is totally incompatible with the transition dates of the MDR. For example, there is no guidance on clinical evaluation for manufacturers. As a result, the MEDDEV 2.17/1 rev.4 guide, applicable under the old directive 93/42/EC, is still applied by manufacturers.
3. The delay in establishing the list of harmonized standards. Manufacturers and NBs are in a total fog, each defining their own rules concerning the definition of the state of the art in standards. This is an important issue that has been dragging on for many years, allegedly due to lack of budget. In defence of the MDR, these problems existed even in the days of the directives. But the MDR is supposed to be there to solve the problems encountered with the directives.
4. The delay in the implementation of the Eudamed database. It is true that Eudamed is being phased in, and existing procedures continue to operate in the absence of a fully functional Eudamed. But, in this case, the timetable was initially defined in spite of common sense; obviously by a working group not very aware of the implementation time of such an information system.

If these delays were all manageable, the situation would be tolerable. But given the level of complexity of the transition, manufacturers are facing, with new and increasingly stringent regulatory requirements, this situation is becoming unacceptable.
Meeting imposed transition deadlines, allocating enough resources, documenting evidence and demonstrating compliance with requirements. The entire burden falls on manufacturers, who are short on resources, short on time, and without available NBs. The MDR is literally "do as I say, not as I do".

The consequences of this situation are poisonous: manufacturers have to survive economically and are abandoning their unprofitable product lines under this new regime. Unprofitable but with a clinical benefit confirmed by several years of presence on the market. Or they simply turn to other markets. Starting with the USA, where the FDA manages the regulatory provisions in a meticulous manner. A regularly updated database of recognized standards, new guidelines published every month (which some people will even cite in a CE marking file for lack of an equivalent on the MDCG side!) and updated every five years. And of course, a submission will be answered by the FDA within the time limit defined by the law: 15 days, 60 days and 90 days for the final decision. We are far, very far from the disorganization of the NBs, which, for some, do not even respond to the manufacturers' requests for quotations.


Another consequence even more intolerable is the overbidding on prices imposed by the NBs. 8000 euros a day for some NBs. This is taking manufacturers for cash cows!



So, what should we do? Let's not reject the MDR in bulk. It clarifies a lot of issues that were little or not addressed by the old directives. To give just a few examples: clinical evaluation, clinical investigation, post-marketing surveillance, content of technical files, obligations of manufacturers and economic operators, and more generally a step forward in the harmonization of the legislation of EU member countries. An update of the MDR to make the task less difficult is therefore the preferred solution.
The voice of manufacturers' associations, such as Medtech Europe, are being heard throughout Europe to extend the transition period to 2026, or even 2028 for low-risk devices. Why not, but what about CE certificates of conformity under the directives that expire before 2024? A new transitional provision would have to be found, which could involve the NBs or the Competent Authorities in extending existing CE certificates.

One solution is to ease the transition of CE-marked devices under the directives to the regulation. A blocking point for many of these devices is the lack of clinical data for a clinical evaluation brought up to the requirements of the MDR. The possibility of a simplified procedure for a device, already CE marked under the directive, based on the absence of changes impacting the safety and performance of the device version to be MDR CE marked, would be a way to ease the workload of manufacturers and NBs.
This may make some people gnash their teeth, but one can agree that the clinical evaluation required by the MDR for low class devices generates a lot of paperwork for little value. If we look at the bigger picture, extending the equivalence procedure, in the manner of the FDA's substantial equivalence, possible for all devices other than class III and IIb subject to article 54, would lighten everyone's workload even more. It works in the US, why wouldn't it work in the EU?

A third solution is to review the classification rules in Annex VIII of the MDR. Indeed, the most striking result of the MDR is seen in this annex. Many devices change class with the MDR, from Class I to Class IIa or higher. Class I is really reduced to a bare minimum. As a result, all those devices that were previously Class I, which did not require NB intervention, will require NB intervention upon MDR CE certification as Class IIa or higher.

This escalation of classification is a major contributor to the difficulties we are experiencing. More devices to be certified and fewer NBs lead to a huge bottleneck in the certification of medical devices. Acting on the transitional period is not enough. Revising the classification system with more class I and less class IIa is also an effective way to reduce the workload of NBs.
Rule 11 should be especially amended for software as medical devices (SaMD). As it stands, all software is at least class IIa, as soon as it has a function to aid in diagnosis, treatment or monitoring of the patient. Even if this function is indirectly related to diagnostic or treatment, even if the monitoring is related to low-risk clinical cases. Rule 11 could be amended by allowing these low-risk software applications to be positioned in Class I. This involves many manufacturers, from start-ups to big corporates, many mobile applications and many cloud applications. This change in Rule 11 would also have the virtue of aligning the European classification policy for low-risk software with that of the FDA, which does not hesitate to position low-risk software in class I or even exempt it from "510(k)", according to its "law enforcement discretion".

Because we have to talk about escalation of classification. An orthopedic surgical instrument, product code LXH is in class I in the USA, 510(k) exempt. The same device must be certified by a NB for the reusable part. Hospital prescription assistance software (CPOE as it is called in the US) is not even a medical device in the US. The same software is class IIb and must be certified by a NB. A remote patient monitoring software, product code QNY, has a clearance exemption procedure with the FDA. The same software will be in class IIa and must be certified by an NB. A non-active mediastinal drain used for more than 30 days, product code GBS, is in Class I 510(k) exempt with the FDA. The same drain is Class III (yes, you read that right!) under Rule 7 of the MDR.

This quick comparison with what the FDA practices makes it clear that the escalation of classification is not an unfounded idea. Some may criticize this comparison as being too pro-American. But on the one hand, it is healthy to look at what works elsewhere and on the other hand, this type of discussion has probably already taken place between the drafters of the MDR in the mid-2010s. It is not shocking to revisit it.

In conclusion, let's advocate not only to extend the transition period, but also to revise the MDR classification system downwards and to facilitate the transition of legacy devices with relaxed rules. It is not one or the other, but all of these actions - and perhaps others - that should be promoted to the authorities, in order to reopen the horizon for manufacturers who are currently in the fog, and to bring stability and serenity to the European medical device market.