Risk trace matrix vs design trace matrix

[yodon]

Good point. I did neglect to say that in our Hazard Analysis, we do identify standards to which the system will be designed to address the hazards. For example, we will "design the system to comply with IEC 60601-1 for electrical safety."

In your examples (user uses device after expiration, improper disposal), those *are* failures which we would cover in a Use FMEA.

 

[Ed Panek]

Correct. I’m not sure it’s been mentioned but it’s possible you won’t be able to reduce a risk to an acceptable level. In a classic example is a CT. Or x ray device where the fundamental risk of the device is also highly risky. In that case you identify it and justify that risk of radiation to the therapeutic benefit of the device. A risk benefit analysis.

 

[stm55]

Good point. I did neglect to say that in our Hazard Analysis, we do identify standards to which the system will be designed to address the hazards. For example, we will "design the system to comply with IEC 60601-1 for electrical safety."

In your examples (user uses device after expiration, improper disposal), those *are* failures which we would cover in a Use FMEA.

Fair point regarding Use FMEA. I partly misspoke in terms of my examples, but I still have the same general question regarding "normal use risks" which would not appear on FMEA. You originally said that your Hazard Docs would not get into Controls and Risk Scoring, and instead you would rely on your FMEA to do that. But then how do you address the normal use? I quickly looked for examples-- a defibrillator being used exactly as instructed can still potentially have the paddles leave burn marks on the patient-- since there is no fault condition, this theoretically would not appear on an FMEA (by definition, since there is no failure). However, still dealing with a potential for patient harm so still need to put together risk scores and potentially further controls?

(There's obviously other examples of this of these normal faults and there seems to be numerous articles out there arguing that these do not get sufficiently addressed by FMEA... Curious how you would address something like this since you do seem to have a good handle on your 14971 risk docs otherwise)