# Using Complaint Rate to Update FMEA P.O.C

J

#### Jason3312003

I am looking for some information and guidance on how to properly analyze a need to update an FMEA POC based on the rate of complaints received.

My specific situation at the moment is that I have a complaint lodged that relates to an identified risk. The risk was assigned a severity of 2 with a POC of 2. A 2 POC equates to a failure rate >.01% but <=.1%. With this complaint we now have 2 related complaints our of 695 recorded uses of our device. This equates to a 0.29% rate of failure(POC of 3).

I am currently debating internally whether 695 uses is a significant sample size and a change to our FMEA should be implemented. The 695 uses are single use devices, and the lowest out of 4 total product lines. The arguement against an update has been that even 1 complaint for this product line is over many of our POCs as most are a 2 or 1 estimation.

Is there enough evidence and trend to justify revising the FMEA and ultimately generating a CAPA for Risk mitigation based on 2 complaints?

#### Bev D

##### Heretical Statistician
Super Moderator
you should never let data over rule your optimistic guesses...

seriously, 695 is more than sufficient. It is certainly more accurate than the pro forma guess at occurrence.

human beings are hopelessly inept at guessing the occurrence rate. An example I use in class is the hydraulic loss of United Airlines flight 232 (the Sioux City crash). The odds of an event causing the loss of all redundant hydraulic systems was estimated at 1 in a billion by engineers who are way smarter than me. Prior to flight 232 there were THREE similar events of multiple hydraulic system loss and the 'fmea' was NEVER updated, because, well it was a 1 in a billion chance. I dont' know what the 'billion' referred to, planes, flights, etc. but there were certainly not a billion of any of those things between the 3 occurences... In fact when the pilots tried to get technical assistance on how to fly the plane they were repeatedly told that it was impossible for them to have lost all hydraulics!

you now have data - use it.

#### Ronen E

##### Problem Solver
Moderator
I have nothing important to say about the statistics, but why do you (automatically?) conclude that a revision of the POC necessitates intiating a CA? Is it required in your SOP? Is the risk now unacceptable?

Last edited:

#### Bev D

##### Heretical Statistician
Super Moderator
You make a good point. I was addressing the question from a general perspective. At a severity of 2 I would like have two questions: if the severity really is 2 why would you need a capa or any other mitigation? A 2 severity is very low. Which begs the second question: why would you even get 2 complaints in 695 uses for a severity 2 failure??? Are you sure you have the severity graded correctly?

J

#### Jason3312003

You make a good point. I was addressing the question from a general perspective. At a severity of 2 I would like have two questions: if the severity really is 2 why would you need a capa or any other mitigation? A 2 severity is very low. Which begs the second question: why would you even get 2 complaints in 695 uses for a severity 2 failure??? Are you sure you have the severity graded correctly?

A level 2 severity per our SOP is described as "Low - Results in temporary injury or impairment not requiring professional medical intervention / no injury or impairment but malfunction may require additional surgery time". Both of these related complaints fit this description and have resulted in additional surgery time.

A revision from POC 2 to 3 based on complaint history will maintain our overall moderate risk level; however, a Sev 2 and POC 2 was originally mitigated to try and decrease the risk. This resulted in a Sev 2 POC 2 being allocated but determined the risk could not be decreased further. Now we are increasing the POC to 3 and there for need to generate a CAPA to address the heightened moderate risk level. (Reason for going right to CAPA from FMEA revision).