How to estimate risk for a medical device according to ISO 14971:2019

[@robt3575]

Great, presentation and look forward for more videos on Risk Management and Design controls.

[@KrishnaChaitanyakc9]

Very well done.

[@ricardoparales8611]

Nicely done. Really like the areas to look for team unity at the end.

[@MedicalDeviceHQ]

Thanks!

[@marcmuller355]

nice overview. thanks :)

[@vsilte2]

Thanks for sharing! Very interesting topic!

[@MedicalDeviceHQ]

Glad it was helpful!

[@vsilte2]

@@MedicalDeviceHQ I liked so much that I registered to your course online.

[@timithomas6116]

Nice video

[@SK-ph7sz]

Superb!

[@MedicalDeviceHQ]

Thank you! Cheers!

[@PMpunk]

Any videos/slides on EU MDR/MDD and different phases in them? Great content 👍

[@MedicalDeviceHQ]

Thanks for the feedback. There aren't any "phases" as such defined in EU MDR. You would have to turn to ISO 13485 that says you need to define the stages of development in your design planning. And what they are and what you call them are up to you!

[@AB-qh3zz]

Super

[@vijaykumarb2266]

I am Considering you as my GURU 👌🏻👌🏻

[@MedicalDeviceHQ]

Thank you Vijay! I appreciate that.

[@beluhmabasa4944]

Is Risk Control Measure an integral part of Risk Management Report.? If true. Why has never been a column in risk management report for risk control measure? Thank's

[@MedicalDeviceHQ]

Risk controls should not be part of the risk management report. The minimum risk management report contents are specified in the ISO 14971: the results of the risk management review, the result of the overall residual risk evaluation and the answer to whether you have processes in place for collection and review of production and post-production information.

[@Dan-oq4ui]

Hi, the link for matrix doesn't seem to work, how can I source this template?

[@ivanguzman6462]

Common mistake is taking the probability of failure of product design risk tools (e.g. dFMEA) into hazard analysis as a probablity of the harm.

[@MedicalDeviceHQ]

That is indeed a common mistake.

[@markkotfila6286]

The math at 11:20, 100%*0.1%=0.01% is incorrect.

[@20304374]

You just need to remove the % it is actually 0.01 or 10%

[@MedicalDeviceHQ]

Yes, you are right. THat one slipped through the cracks. It is updated in the online course though. We will be making an update here too.

[@chriskortright8004]

Question: at 12:21 the example shows P1 as a combination of Hazards A, B and C. These probabilities are multiplied together to get P1 and I interpret this to mean all three hazardous events must occur for the Hazardous situation to occur. But if any one of the 3 event could lead to the hazardous situation on it's own, would P1 then be the sum of event probabilities? Thanks for the video.

[@hamdymonem8616]

how can i dowload the materix file you mentioned in that vedio

[@wilmarkjohnatty4924]

YOu cant, the website is desgined just to collect your private information

[@delphinemercier5382]

What is difficult is to evaluate the probability of harm when you don't have data, for example when you want to market a very innovative device. According to ISO 14971, if you don't have data then you should consider a probability of 100%. This could lead to non acceptable individual risk and to non acceptable global risk for device. The consequence is that the device won't be put on the market unless the manufacturer performed large clincal investigation.

[@MedicalDeviceHQ]

Hi Delphine, sorry if I come across as being picky with words, but the standard tells you to work with Po=100% when it cannot be estimated. But not having data is not the same as it being impossible to estimate. "Expert assessment" is one method to use to estimate the probability. In fact, for most new products, it is the method that is used for most risks. There are however cases where you cannot estimate, for example when software is involved in the foreseeable sequence of events or where the clinical consequence of a specific circumstance cannot be explored due to ethics

[@delphinemercier5382]

@@MedicalDeviceHQ thank you for your answer. I agree that "not having data" = "not possible to estimate". The device I have in mind is a software and you "answer" in the second part of your post. 🙂

[@ortonormab126]

Yeah, I know; real late comment: Biggest mistakes is adding numerous of innovative risks and not thinking 1 "are all of these unique risks or can we merge them" 2 "Is this really a probably risk?" Often I see imagninary risks based on really bad behaviour. I mean You can stab someone with a stetoscope. And why add ten different wounds from a sharp edge!