Confused by UDIs? Check out our step-by-step video course and watch us purchase real UDIs

Articles Questions

November 15, 2022

Should You Do a BfArM Classification Request? The Answer Is No.

Dr. Oliver Eidel


We have no clue how to classify our medical device under the MDR. It could be class I, so we won’t have a notified body. Is there anyone who could give us a final decision? Our company is located in Germany. We heard of filing a so-called “BfArM classification request”. Should we do that? What are our options?

Short Answer

If you think your device is a MDR class I device, you have three options:

In this post, we’ll focus on the BfArM classification request. Should you do it?

The answer is no. It will take too long - one year or even longer. If you’re a startup or any other company which has a legitimate interest in making progress and shipping products, it’s not viable for you to wait so long for a classification decision.

Long Answer

The classification of software as a medical device (SaMD) under the MDR is currently quite unclear. The situation is especially bad when it comes to MDR class I software, because no one on this planet currently has a clear idea what it really looks like. This has lead to different states in Germany coming up with different “tolerance levels” for MDR class I software - Hamburg and Bavaria have multiple MDR class I software devices on the market while Berlin has none at all. Crazy!

But, arguably, this is not their fault because the MDR is vague and there’s no useful guidance around. So they’re just doing their best, I suppose. Wait, so don’t we have a solution to this - an entity or a process which could make final classification decisions for real medical devices?

Luckily we do (or so we think). A so-called BfArM classification request (“Antrag auf Klassifizierung”). The concept itself sounds really cool and like a perfect solution to this problem: You submit some documentation about your product to the BfArM, e.g. your intended use, user manual, (draft) classification, etc., and the BfArM will get back to you with a final decision on what the “real” MDR class is. This decision is final, in other words, notified bodies can’t argue against it.

Okay, so that’s the concept. Now, how does it work? Or rather.. does it work at all?

The answer is no, it doesn’t work, you shouldn’t do it, and the mere idea of thinking that this process is helpful to anyone is wrong. I’m trying to become more tolerant towards state institutions because I feel that they often contain friendly people with good intentions, so I’ll stop the ranting here and show you a timeline of a BfArM classification request we’ve submitted for one of our clients. Feel free to make your own judgements while reading along:

BfArM Classification Request Timeline

Quick joke: How do you identify a programmer? By the fact that they start counting from zero, including bathroom stalls. So let’s start with Day 0.

Quick intermission!

(Imagine a quick break during an Ice Hockey match in which cheerful music starts playing)

So far, so good! We’ve submitted the appropriate documents to the BfArM. We forgot to send the power of attorney stuff and they told us about that. Emails were sent, replies were received, communication was facilitated. All signs are looking good for The Fellowship of the Ring to throw the Ring of Power into Mount Doom in a reasonable time frame. No need to lose time walking through Mordor.

Let’s see how the journey progressed.

Okay, time for the next intermission!

(Tragic music starts playing while the Ice Hockey players curiously look at the speakers, wondering who the hell selected this music)

You’re surely now expecting me to conclude this post with the final BfArM response containing a classification decision, after which we end up having a constructive discussion with them, and everyone congratulated each other on another day’s productive work in the world of medical device regulation, bringing safe medical devices to market and helping patients.

Nope, that didn’t happen. Are you ready for the real outcome?

The process is still ongoing. I haven’t heard back from them. It’s been significantly more than 312 days which have passed now - we’re actually getting close to our one-year anniversary of our BfArM classification submission. I’m not sure whether this is an anniversary we should celebrate. Probably not.

So far, we got nothing except a list of questions and a request for an updated letter for power of attorney.

That’s it.

Before we get all depressed and conclude that our bureaucratic system of medical device regulation is completely broken, let’s try to analyze this from different perspectives and see what we can do better.


First off, are we having the wrong expectations? Are we those crazy startup people who want to rush things which should not be rushed? Is it better if these decisions take at least one year?

Let’s look at this process from the perspective of a startup. Startups are often unsure whether their software is MDR class I or IIa (see: the MDR class I fiasco situation). This is especially tricky if you think you’re class I because you don’t have anyone to confirm this for you. If you made the wrong call for class I while your “real” class was actually class IIa, you might have to take your product off the market until you’ve gone through a notified body audit. So being “maybe class I” is a risky place to be in.

The idea of a BfArM classification request sounds like an elegant solution to this problem - an independent entity which makes final classification decisions. This could be so cool! However, the usefulness of this process is determined by how well it works. And receiving a response in a reasonable time frame is probably one of the most important factors for that.

What would a reasonable time frame be? 1-2 months maybe? I mean, any amount of time a company could survive while literally doing nothing at all and waiting for a reply. Wait, why would the company be doing nothing? That’s because many small companies revolve around one (or very few) products, and a timely classification decision is essential because it determines their entire product and regulatory strategy, e.g. whether to select a notified body for class IIa or bring it to market as class I.

But what if this takes one year or more? Then it’s completely useless for small companies.

If it’s useless for small companies, is it still useful for large companies? I’m not the right person to assess that as we mostly work with small companies. All I can say is that large companies tend to have many more products and more revenue. They also tend to have (many) more meetings, political infighting, and ridiculous overhead (those stories probably deserve another website). Productivity-wise, they are very inefficient - which makes sense as productivity doesn’t scale linearly with employee count for knowledge workers.

But that may actually be very beneficial if you’re making a BfArM classification request. If you have 99 other products out there generating revenue and 50% of your people are stuck in useless meetings most of their time, why not task some of those with a BfArM classification request?

If the response of the classification request is positive, that’s cool, because you can maybe incorporate that into your future product strategy - that is, if your CEO is still the same 5 years from now and your product strategy hasn’t changed in the meantime.

If the response is negative, that’s also cool because it can be used as bargaining chip in internal politics - like “hey, remember Joe who did the BfArM classification request? It seems like he really has experience dealing with authorities. We should make him Head of Regulatory!”

Okay, trying to reduce the level of sarcasm here. I would love to see the BfArM classification process work out. I think that, at its inception (heh), it was created with good intentions and I agree with those! It would be really, really cool to have a working version of this process to clear up the confusion around MDR device classification, especially around software. So, trying to be constructive here, which improvements might we need?

So those things would help. Until then we’re unfortunately left with a process which is rather useless due to being opaque and, worst of all, very slow. Why has everyone been accepting this, and why are we the first people to speak up?

In our Healthcare system, when we get hurt, we expect someone to have our back: We can call an ambulance and it will arrive soon. We can also walk into a hospital and get treated soon. In many ways, Healthcare is about delivering treatment to patients within a reasonable time frame.

This concept of a reasonable time frame does not seem to extend to the regulation of medical devices. Currently, notified bodies decline new customers or don’t even respond to emails, and, as we’ve seen here, BfArM classification requests drag on for almost a year or even longer. This way, everyone loses - not only manufacturers, but also doctors and patients. If we want to continue delivering treatment to patients in a reasonable time frame, this has to change.

On a slighty different note: You want to get your medical software certified under MDR but don't know where to start? No worries! That's why we built the Wizard. It's a self-guided video course which helps you create your documentation yourself. No prior knowledge required. You should check it out.

Or, if you're looking for the most awesome (in our opinion) eQMS software to manage your documentation, look no further. We've built Formwork, and it even has a free version!

If you're looking for human help, did you know that we also provide some limited consulting? It's limited because we are not many people. We guide startups from start to finish in their medical device compliance.

Congratulations! You read this far.

Get notified when we post something new.

Sign up for our free newsletter.

Dr. Oliver Eidel

I'm a medical doctor, software engineer and regulatory dude. I'm also the founder of OpenRegulatory.

Through OpenRegulatory, I've helped 100+ companies with their medical device compliance. While it's also my job that we stay profitable, I try to dedicate a lot of my time towards writing free content like our articles and templates. Maybe that will make consulting unnecessary some day? :)

If you're still lost and have further questions, just send me an email.

Read more about me here.

Digital Health Jobs No Cookie For You Privacy Policy Imprint
No QMS on this planet will save you from creating crappy software.