logo-curiaFirst: my apologies for the collapsed frequency of articles on this blog – I’m very very busy, literally day and night, with questions about any and all aspects of the MDR and the IVDR. The summer is a bit more quiet so you can expect more content again.

Intermezzo on software under MDR and GDPR

While everybody should be beavering away on the implementation of the MDR and IVDR or overcoming their inertia do so ASAP, still some interesting things happen in the mean time. If your company is in software as medical device, there is even more reason to do so because of the MDR changes the classification rules for software as medical device dramatically, expands the scope of the concept of accessory (which enlarges the scope of software regulation) and includes a whole chapter on software design requirements.

And then there is the new General Data Protection Regulation that your software needs to be compliant with by 25 May 2018 – less than a year, that’s right. How long is your company’s software development cycle? See my presentation at the Personal Connected Health Alliance in May this year below for a very brief overview of MDR and GDPR compliance requirements coming your way very fast:

If you have class I software on the market now, and it is up-classified under the new classification rules, you need an MDR certificate from a notified body by 25 May 2020.

New EU court case law on software qualification as a medical device

Back to the topic at hand: this opinion of the AG at the European Court of 28 June may have gone somewhat unnoticed because it’s not available in English yet (and not in German, Swedish and Gaelic either). I linked to the Spanish version because it’s the language in which it was written, but you can toggle the language of the document as preferred on the EU Court’s site.

Decision support software

The Advocate General at the EU Court issued the opinion in the SNITEM case on 28 June, a preliminary reference from the French Conseil d’ Etat (Administrative High Court) about the question whether software that supports physicians with information regarding contra indications, dosis and interactions between medicinal products when prescribing medicinal products for reanimation and anaesthesia.

French additional software certification requirement for prescription support software

France has a rule that requires a national certification of software that is used for support in prescription of medicinal products. SNITEM, the French medical devices branch association, and one of its members, started proceedings against the French state based on the argument that CE marking as a medical device precludes additional national certifications for the same software. The French Administrative High Court decided to check in a preliminary reference if the software concerned really was a medical device, because this was crucial for resolving the dispute. The French government argued that the software was not a medical device because support of prescription of medicines did not constitute support of an intended purpose within the definition of medical device, while SNITEM and the manufacturer of the software (that had CE marked the software as medical device already) argued that it was a medical devices.

As per normal procedure the Advocate General at the EU Courtopines on the matter, after which the Court renders judgment, usually in line with the AG’s opinion. While the AG’s opinion is not the Court’s judgment, it’s a good indication of where that will go. It would be news if the Court would diverge significantly. So let’s see what the AG has to say.

Clearly a medical device

The AG’s reasoning is very straightforward: both for anaesthesia as for intensive care purposes the software has functionality with which prescribers can determine what medicines to prescribe, map possible allergies and calculate the duration of treatment. Given that finding, the AG says it can hardly be denied that the software is specifically intended for diagnostic or therapeutic purposes, as required under the definition of medical device after the change in 2007 by means of Directive 2007/47 (see paras 48 and 49 of the opinion). The AG remarks that it is important in this regard that the software is not used after the physician has already determined the right treatment, but rather is used by the physician to do exactly that (para 52). I would argue that even in the first case the software would still be a medical device if it was intended to provide a double check  on the physician’s decision.

Some discussion of the software MEDDEV

There is a discussion of the modules section of the software MEDDEV, which is less interesting because the discussion is mainly about whether the preliminary request for interpretation to the EU Court concerned all modules of the software or just one. It is however interesting that the AG also refers to the national guidances on qualification of software ‘that all lead to the same conclusion’. That’s a hopeful statement that I do not necessarily share, as I have demonstrated in the past how different these national guidances are. I personally like the MHRA interactive pdf document the best – practical and comprehensive. If the Brexit goes south I’ll miss the MHRA and it’s contributions to guidance and implementation. Oddly the AG does not refer to any of the qualification statements in the Borderline and Classification Manual, which has a software qualification section (chapter 9) in it for some time now.

Not surprising

I would be surprised if the EU Court rules differently from the AG in its judgment – it’s obvious that the software is a medical device. The only reason why the EU court had to be involved in the first place is because of the French government pigheadedly holding on to the position that it was not as best argument for defending a national rule that in my view is clearly pre-empted by the Medical Devices Directive. It’s a demonstration of why we now have a regulation for medical devices: to give member states less room for this kind of stuff. In sum: this case adds but does not add anything surprising to the acquis on qualification of software as a medical device in the EU.