Complaint Return Sample Size Question

chaos

Registered
#1
Greetings, first time poster here.

I had a question about complaints with respect to physical returns.

We have a number of low-risk, high-frequency complaints related to the service of our devices each year. If you looked at the failure modes in our FMEAs and hazards established in our system risk files, they are well documented and complete.

Most of these complaints involve failure modes related to wear/tear of subcomponents where assignable cause is fairly straightforward; however, actual root cause is not well understood. We trend these failure modes, and when they breach a certain threshold, CAPA is triggered and root cause investigation will occur at that point.

Getting to the point, my question is: Is there any sort of guidance on sampling for physical complaint returns, or any way to justify "we've received and investigated x number of complaints related to this particular failure mode, so getting an additional y physical returns adds no value", either based on statistical techniques, sampling plans, and/or other?

Thank you for looking.
 
Elsmar Forum Sponsor

planB

Super Moderator
#2
Welcome to the cove!

One thing that strikes me:
How to re-concile "actual root cause is not well understood." with "CAPA is triggered and root cause investigation will occur", specifically:
how would you close your CAPAs without having identified a root cause? Could you still demonstrate a functioning CAPA process, preventing re-occurrence of the same issue and first occurrence of similar issues?

With regards to the wear & tear you observe: would limiting the life time of your device/component to a period that most parts "survive" be an option? In case you had such a defined life time at hand, broken devices having exceeded this life time, might be excluded from having to be dealt with in your complaints / CAPA system.

HTH,
 

chaos

Registered
#3
One thing that strikes me:
How to re-concile "actual root cause is not well understood." with "CAPA is triggered and root cause investigation will occur", specifically:
how would you close your CAPAs without having identified a root cause? Could you still demonstrate a functioning CAPA process, preventing re-occurrence of the same issue and first occurrence of similar issues?
I should clarify that with the extreme number of subcomponents in our devices (dozens upon dozens), understanding root cause for a single subcomponent still leaves numerous others in play where root cause is not well understood. If these subcomponents do not breach the post market surveillance control limit, there is no urgency/trigger to get to root cause.

In a nutshell, we cannot address root cause for every subcomponent from a resourcing perspective. There has to be prioritization based on trending and risk/criticality, impact to customer, etc.

Minor subcomponents with low risk failures don't tend to get promoted into the purview of CAPA.

With regards to the wear & tear you observe: would limiting the life time of your device/component to a period that most parts "survive" be an option? In case you had such a defined life time at hand, broken devices having exceeded this life time, might be excluded from having to be dealt with in your complaints / CAPA system.
Unfortunately we don't have an extensive library of reliability studies for each subcomponent. There are subcomponents that are replaced during preventative maintenance that we can likely justify not returning based on expected wear.
 

Tidge

Trusted Information Resource
#4
Some snipping and bolding follows
[...]
We have a number of low-risk, high-frequency complaints related to the service of our devices each year. If you looked at the failure modes in our FMEAs and hazards established in our system risk files, they are well documented and complete.

[...]

Getting to the point, my question is: Is there any sort of guidance on sampling for physical complaint returns, or any way to justify "we've received and investigated x number of complaints related to this particular failure mode, so getting an additional y physical returns adds no value", either based on statistical techniques, sampling plans, and/or other?
The first part of the quote I included leads me to believe that if I look at the risk management file (RMF), I am going to see some sort of engineering study (i.e. not a WAG) that established the known "wear rate" for the diverse components, along with a confidence interval and power (of the study).

The second part leads me to believe that field data is being collected of sufficient quantity such that it should be possible to construct a hypothesis test to determine (with appropriate confidence and power) that the observed rate of failures in the field (for the specific, pre-attributed components) is NOT exceeding the established failure rates in the RMF. This assumes that the sole point of value is "did we get the rates right in the original RMF?" This could be something like a non-inferiority test.

If I may be frank: I have no doubt that an original design team understands (i.e. can predict) some number of potential failure modes due to device "aging". I have much less personal experience with design teams doing this in a thorough manner that justifies specific, reduced (occurrence) ratings in the RMF. Pre-market, I have observed appeals to "the bathtub curve", "MTBF", and "shake and bake" (on finished devices), all of which can tell you something about the gross reliability properties of a finished device but haven't (in my experience) provided that much information about specific failure modes... unless those modes actually appear!

My recommendation is: don't try to rely on "appropriate statistical methodologies" that could be (mis)interpreted as trying to avoid learning about potential failure modes that were not considered. If nothing else, the analysis of failures in the field can be used to solidify the assessment of (post control) ratings in FMEA.
 

Steve Prevette

Deming Disciple
Staff member
Super Moderator
#5
A primary concern is economics. How much money do the returns cost you, and how much do they damage your reputation? If it is an "oh never mind" situation, then likely not worth a lot of effort. However, from a risk perspective, if the costs are high, it's worth the return on investment to put some effort into this.
 
Thread starter Similar threads Forum Replies Date
Melissa Complaint about Adhesive ISO 13485:2016 - Medical Device Quality Management Systems 9
K Do Contract Manufacturers need a complaint procedure? Medical Device and FDA Regulations and Standards News 9
A Complaint review as part of the complaint handling process? ISO 13485:2016 - Medical Device Quality Management Systems 3
M Reduce occurrence rating based on the PMS data and customer complaint data ISO 14971 - Medical Device Risk Management 2
B Complaint Records - Accessing records on Easy Track Software Records and Data - Quality, Legal and Other Evidence 3
J Customer Complaint & SCAR, false data Nonconformance and Corrective Action 14
M ISO 13485:2016 Complaint Definition Clarity Customer Complaints 2
J Customer Complaint Response 21 CFR Part 820 - US FDA Quality System Regulations (QSR) 3
S How to treat a customer complaint ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 7
A Requirements for Complaint Files - Distributed Goods Customer Complaints 3
S Is any dissatisfaction over a Medical Device considered as a complaint? 21 CFR Part 820 - US FDA Quality System Regulations (QSR) 8
A Complaint alert levels - Industry standards pertaining to alert levels for complaints Customer Complaints 5
B Procedures for Complaint Handling and Post Market Surveillance EU Medical Device Regulations 10
J KPIs or Metrics to Measure a New Complaint Handling Process 21 CFR Part 820 - US FDA Quality System Regulations (QSR) 9
D Class II medical device - When should a complaint be closed? Customer Complaints 6
PhilM Nonconformance report, Customer complaint investigations and RMAs ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 8
G Problem Resolution Report Monitoring - Customer complaint or PRR as general motors use Customer Complaints 12
R Does an RMA = a Customer Complaint Customer Complaints 4
G Customer complaint over a Reference Dimension Manufacturing and Related Processes 9
Z Security for Approvals - Cloud based Complaint, NC, and CAPA systems Qualification and Validation (including 21 CFR Part 11) 8
S How to collect complaint data from customers Customer Complaints 24
E Is there an Automotive Customer Complaint Index? IATF 16949 - Automotive Quality Systems Standard 3
R In a software development company: Is every bug reported by the customer a complaint? ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 27
E Corrective Action or Customer Complaint ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 16
T Decontamination responsibility on customer complaint investigation Customer Complaints 4
N Possibly worlds first customer service complaint Supply Chain Security Management Systems 0
A What is a Customer Complaint? (ISO 9001:2015) ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 9
S What is the difference between a service request and a complaint? ISO 13485:2016 - Medical Device Quality Management Systems 2
M Should Potential Customer Complaint Outcome Define Registrar NC Rating? ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 8
J FDA definition of a complaint - Customer orders spare parts 21 CFR Part 820 - US FDA Quality System Regulations (QSR) 5
R 510(k) - Finished Test Reports - RTA - Bit of a Complaint 21 CFR Part 820 - US FDA Quality System Regulations (QSR) 6
P Customer Complaint Definition - Food Distributor Customer Complaints 5
J Using Complaint Rate to Update FMEA P.O.C ISO 14971 - Medical Device Risk Management 4
C Complaint Training - FDA - How to recognize complaints and what actions to take 21 CFR Part 820 - US FDA Quality System Regulations (QSR) 8
M Complaint Handling Responsibilities for a Design Partner - ISO 13485 ISO 13485:2016 - Medical Device Quality Management Systems 9
S Merging Post Market Surveillance and Complaint Handling ISO 13485:2016 - Medical Device Quality Management Systems 3
R Customer Complaint - Samples vs. Removal - FDA Requirements 21 CFR Part 820 - US FDA Quality System Regulations (QSR) 2
J Retention Requirements of Complaint Medical Devices Customer Complaints 2
C Web based Complaint Software which interfaces with our Enovia system Document Control Systems, Procedures, Forms and Templates 0
Moncia Complaint / Non-Conforming Product / CAPA Procedures - Please share your experience ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 3
Moncia Customer Complaint Procedure example needed Document Control Systems, Procedures, Forms and Templates 1
C Medical Device Malfunction - Customer Complaint - Response vs. Correction Other US Medical Device Regulations 3
J ISO 13485 Complaints (what is a complaint) Customer Complaints 5
Q Internal Audit of Product Quality Complaint System (21 CFR Part 820) Customer Complaints 9
T PPAP - Complaint from a Tier One customer concerning Product Dimension IATF 16949 - Automotive Quality Systems Standard 26
M Using a Complaint Management Software (Medical Device companies) 21 CFR Part 820 - US FDA Quality System Regulations (QSR) 2
A Production and Post-Production and Complaint Handling ISO 14971 - Medical Device Risk Management 2
M Process for closing complaint records after a period of inactivity? Quality Manager and Management Related Issues 2
Q Customer Complaint Vs. Nonconformance - CAPA SYSTEM Customer Complaints 2
N Survey to Customers Regarding Our Complaint Response Time Customer Complaints 7

Similar threads

Top Bottom