MDR - Rule 11 Software class I

TomQA

Involved In Discussions
Hello,

This is more of a survey/curiosity but are medical device softwares currently being classified as class I under the MDR ? Or are notified body really against it?
For instance, we have a sensor that sends data to a mobile application to present/visualise the data.
At the therapist uses the data is considered as " information which is used to take decisions with diagnosis or therapeutic purposes" then the mobile app should be a class IIa correct?

It just seems like a very regulatory burden.
 

yodon

Leader
Super Moderator
If you're just providing a means to aggregate and present the data, you're (your software is) not making any diagnoses or providing clinical direction.

That said, what do you find to be so burdensome? As I've suggested before in other posts, most of the "burden" is just good software engineering and enables better support throughout the life of the software.
 

TomQA

Involved In Discussions
Yes but it's providing information which is used to take decisions with diagnosis or therapeutic purposes no..? So class IIa?

Well if it's class I the device can be placed on the market much faster and at a lesser cost. It was more burdensome on that side.
 

eskimomike

A&D Req. Engineer & Quality Expert (30+ YOE)
Yep software that's used for making diagnostic or therapeutic decisions i.e. web app that visualizes sensor data for therapists, would likely be classified as class IIa.

most of the "burden" is just good software
@yodon – What exactly are the controls that are applied at the application code layer under MDR? e.g. SOC2 describes controls at the infrastructure layer so if you're deploying your software on a cloud computing server providers like AWS you're automatiacally compliant with some, not all, of those controls. I'm not a software guy though so I could be wrong.
 

eskimomike

A&D Req. Engineer & Quality Expert (30+ YOE)
the device can be placed on the market much faster and at a lesser cost

Again not a software guy but why would you go to market without testing on a focus group first, thus alleviating the burden of complying with any regulation?
 

yodon

Leader
Super Moderator
@eskimomike - what I typically hear as being burdensome is the need to write design documentation (architecture, detailed design).

@TomQA - in theory, it's possible to separate device class from software safety class (62304 --- although FDA has gone a different direction with categorizing by documentation level --- which I understand is also in the works for a 62304 update --- talk about shifting sand!). In practice, that's been challenging. And while not doing good software engineering may get you to market faster, you typically pay out the other end with maintenance difficulties.

Still not sure about the IIa designation if the software isn't making any clinical decisions and just displaying data. I would assume your sensor to be IIa since it's providing the data. I'll defer that to a regulatory expert.
 

dgrainger

Trusted Information Resource
It likely to be class 2a but I would ask for more information about the use case.
  • What data is the sensor gathering?
    • How important is it for the decision making?
  • Is the sensor a device in its own right?
    • Might the software be an accessory?
    • Does the software drive/influence the use of the sensor?
  • How is the software present/visualise the data, is it just reproducing the data from the sensor?
  • Is the situation high risk?
    • Might class 2b or 3 be more appropriate?
  • Who is the user? patient or HCP?
  • Might other rules apply? Rule 10 if for diagnosis.
MDCG 2019-11 is the guidance to look at.:)
 

mjamesl

Registered
The challenge we are running into is the meaning of the term "influence" in the rule. We have a software device that accepts video inputs and routes them to monitors in the OR. We don't control the sources or process the video. The reviewer is questioning the risks associated with influencing the use of these sources (all of which have their own dedicated monitors also) though. Has anyone else had to deal with the "influence" question? Any advice or examples are appreciated. Thanks.
 

DanMann

Quite Involved in Discussions
Again not a software guy but why would you go to market without testing on a focus group first, thus alleviating the burden of complying with any regulation?
I think Tom is saying that Class I is much less burdensome than Class IIa, because there is no NB approval needed for Class I, but there is for Class IIa. I'd also be interested to hear if anyone has found a sensible example of a Class I Software Medical Device.
 

yodon

Leader
Super Moderator
The reviewer is questioning the risks associated with influencing the use of these sources (all of which have their own dedicated monitors also) though.
@mjamesl oops... somehow missed this reply.

You bring up a very interesting point.

In your case, it sounds like a software failure could lead to complete loss of video or possibly a lag in display. I could see either of those introducing significant risk.
 
Top Bottom