How to handle Out of Tolerance during Calibrations and exceeded Alert Action Limits

[Phoebe]

We have a system where deviations, discrepancies and nonconforming product are all routed to our CAPA system. How are others handling out of tolerance results and exceeded environmental monitoring results? Are they treated as a separate category or is the nonconformance expanded to include more than a nonconforming product per release specifications? We are ISO 9001 and ISO 13485.

 

[Duke Okes]

Re: How to handle Out of Tolerance during Calibrations and exceeded Alert Action Limi

That's your organization's decision, but putting them into the CAPA system makes sense to me. However, when you assess the risk you may decide that containment & remedial action are all that are required, rather than going after physical and system causes.

 

[Jim Wynne]

Re: How to handle Out of Tolerance during Calibrations and exceeded Alert Action Limi

We have a system where deviations, discrepancies and nonconforming product are all routed to our CAPA system. How are others handling out of tolerance results and exceeded environmental monitoring results? Are they treated as a separate category or is the nonconformance expanded to include more than a nonconforming product per release specifications? We are ISO 9001 and ISO 13485.

If you use the ISO 9000 definition of "nonconformity" which is "non-fulfillment of a requirement," you can fit anything you want into the CA category. Worry less about the box you put something into and more about the contents. Do what helps to prevent recurrence.

 

[Phoebe]

Re: How to handle Out of Tolerance during Calibrations and exceeded Alert Action Limi

Would someone be willing to share a sample report for an OOT finding?

 

[BradM]

Re: How to handle Out of Tolerance during Calibrations and exceeded Alert Action Limi

Hello Phoebe!!

OK, this is just me kind of talking about the processes and such. Duke and Jim have provided some excellent advice. So this is more food for thought more than anything.

1. Have your acceptance tolerances clearly defined up front, and in writing. So for example, say your adjusment criteria is 2, but the process tolerance which would impact the process is 4, then only initiate a CAPA (if that is your use) for what would impact the process. So, only involve others, when the error found could impact their process. Again, this should be identified up front.

2. If you have an out of tolerance that might affect the process, then enter that information with the CAPA. I would include the magnitude of error, and some previous history on the device. The calibration entity should provide the sufficient information surrounding the instance, so people can make better informed decisions.

Your CAPA should follow a general process:

CORRECTION: The unit was adjusted for maximum accuracy, replaced, taken out of service, sent to mfg. etc.

CORRECTIVE ACTION: Maybe replace the unit, investigate the calibration process to see where the error may have come from during calibration, etc. May be some training issues. Did you identify the cause for the failure? Do you have a solution for the cause?

PREVENTIVE ACTION: Whatever your organization deems to assure this does not happen again. A fair amount of time there is too tight of a tolerance somewhere, or it's a random event (assuming that the calibration process is stable and somewhat mature).

However you decide to assess these occurences (and yes, they will happen), have things identified up-front to minimize knee-jerk reactions and such (and should be a risk-based approach), and complete it; from start to finish. Too... If you find you are having failures on the same instruments, address your corrective/preventive actions, as they are most probably not sufficient.

Oh... try to avoid taking the easy paths- shortening calibration interval, sending to mfg. for service, or replacing the instrument. They have their place, but so many times are used as Band-aids for a problem, without figuring out what is really going on.