Failure Mode and Effects Analysis (FMEA): Risk Acceptance Matrix
This document defines under which circumstances risk are acceptable. It defines probabilities and severities and each probability-severity combination is either acceptable or not, as shown in the risk acceptance matrix at the bottom.
Generally describe what your company’s risk policy is. Most likely, that’s something along the lines of “our products are generally safe and shouldn’t cause more harm than the next-best alternative”. So, if you’re in the business of predicting Covid infections, maybe you’d compare yourself to people not using your software and doing a non-professional google research about their symptoms. Obviously, that’s a worse alternative. In your clinical evaluation you should collect data on those alternatives to be able to compare yourself to them :)
Define what can go wrong with your product here. Make the examples specific - chances are, your product can’t cause skin lacerations. In all likelihood it also doesn’t cause death. So, feel free to remove severity rows here. But most importantly, customize the definitions and examples so that they resemble the harms in your product.
|Severity||Definition and Examples|
|S1: Negligible||Minor, reversible damage, e.g. superficial skin irritation, delay of non-critical treatment|
|S2: Marginal||Minor, reversible damage with required medical intervention, e.g. skin laceration requiring stitches|
|S3: Critical||Major, irreversible damage with required medical intervention, e.g. irreversible deterioration of disease|
Define your probabilities. You can probably just use these definitions. The idea is that each probability row is 10^2 apart from adjacent ones.
Also, change the “Estimated Maximum Event Count”. That’s the usage number you estimate for your (not yet released product) during its entire lifecycle (which you need to define). So, if you assume that your product will be on the market for 4 years and that it’ll be used 100 times per day, that results in 146.100 usages (100 usages/day * 365.25 days/year * 4 years). The numbers in the lower columns of “Estimated Maximum Event Count” are simply the total usage number multiplied with the upper limit probability of the same row, e.g. you want to know “how often can probability P3 occur if the product is being used 100 times per day?”
|Probability||Upper Limit||Lower Limit||Estimated Maximum Event Count|
|P5: Certain||1||10^-2||1000000 (change this)|
Risk Acceptance Matrix
The most important part. You assess each severity-probability combination whether it’s acceptable for you as a company. There are no definitive rules on what’s deemed acceptable. It depends on your company’s risk policy and, more importantly, the benefits of your product which you show in your clinical evaluation. So, for example, if you product saves 10 lives per day, it might be acceptable to cause one death per day. If your product doesn’t save any lives, it might not be acceptable to cause any deaths. You get the idea, I hope.
|Probability||S1: Negligible||S2: Marginal||S3: Critical||S4: Catastrophic||Estimated Maximum Event Count|
Template Copyright openregulatory.com. See template license.
Please don’t remove this notice even if you’ve modified contents of this template.