Post-Calibration Checks in Mapping Validations: Should you perform them?

Good distribution practice revisions
Paul Daniel, Vaisala
Senior GxP Regulatory Compliance Expert
Published:
Industrial Manufacturing and Processes
Industrial Measurements
Life Science

During one of our recent webinars, we received a question regarding pre-and post-calibration checks:

"Will an FDA inspector expect to see pre- and post-calibration checks on data loggers?"

 

BLOG-IMAGE-vaccine-storage-cooler

Paul Daniel's Answer:

An FDA inspector will expect to see assurance that the sensors used in mapping are accurate throughout the duration of the study. Pre-calibration checks are the standard practice to show that sensors are accurate at the start of a study. Over the years, inspectors have come to expect post-calibration checks because most people used thermocouple-based systems.

The nature of thermocouples is that normal use, such as bending, will cause the accuracy of the thermocouple to drift. If the drift has been significant, it will only be revealed by a post-calibration check. So there is a historically good reason for post-calibration checks.

Another reason to do a post-calibration check: it's always best practice to show an inspector exactly what they expect to see during an inspection! If you are using nice, stable thermistor-equipped data logger, the accuracy issues simply are not as central to the use of that measurement device. But this doesn't change the legacy left by thermocouples in GxP environments.

Here is a potential solution...

Do the post-cal check. Not to be too self-promoting, but this is REALLY easy with Vaisala data loggers. You can easily create a stable point-of-use environment. For probe-based temperature loggers, simply zip-tie the probes around your reference RTD. Then place the loggers in a small insulated container and then put the container in a controlled environment.

You can even place water bottles in the container to dampen temperature changes. If you already have this "post-cal box" stored inside the environment you are mapping, it will already be equilibrated.

This technique should easily provide an environment that is suitable for a ±0.5C single point verification. The same process could be adapted for temperature/RH loggers. If you choose to use a procedure like this, you should develop an SOP and validate the process.

  1. Once the study is complete, put the loggers in this box overnight, leaving it in the area you just mapped (or a more stable environment if you have access to one). When you download the data for your study analysis, you have already got a data set included to use for your post-cal reports.
     

  2. Compile some data to prove that your data loggers have a minimal drift to provide you and the inspector with confidence that the data is good post-study. Our own internal statistics show that over 95% of Vaisala data loggers (over 98% for temperature channels and over 95% for humidity channels) sent to us for recalibration are still within tolerances after 1 year. (Below is a peer-reviewed paper we published on some of this data.) If you can show that you have compiled data on your own tools and present that to an inspector, it will generally make them happy.
     
  3. Make sure to include acceptance criteria that allows for a 5 - 10% failure rate on post-calibration results.

We do have customers who do not perform any post-calibration checks on loggers. Considering the data we have gathered from loggers sent to us for recalibration, we feel confident that this is a safe practice.

However, it remains to you to assess your own regulatory risk appetite. This is why we won't openly recommend skipping the post-cal. I would only recommend a customer do this if they are comfortable explaining their rationale to an auditor. It may be simpler to just show do the post-cal…

Please leave your questions/comments to me directly. I am happy to offer any clarification or guidance I can!

Respectfully submitted,

Paul Daniel
Senior Regulatory Compliance Expert

Vaisala´s calibration services

Standard calibration

Vaisala Standard calibration is equal to the calibration performed for new Vaisala instruments before those leave the factory. It is a perfect choice when you need to verify that the instrument accuracy matches the original specifications.
On a wall of a laboratory: Temperature and humidity data logger DL2000 – superior accuracy and stable measurement data loggers for temperature and humidity.

Accredited calibration

Accredited calibration is a right option if your instrument is used in regulated industries, used as a reference instrument or your quality standards require a calibration complying with strictest regulations.
Accredited ISO/IEC 17025 calibration by Vaisala.

Custom calibration

Custom calibration is the correct choice if specific points within your instrument’s measurement range are needed.
viewLinc with AP10 and RFL100 in a laboratory setting

Standard ISO 9001 On-site Calibration

Vaisala provides Standard ISO 9001 On-site Calibration Service for selected data loggers and instruments used in Continuous monitoring systems. The service is a convenient alternative when the system downtime is not an option. On-site calibrations are always performed by the Vaisala experts and complemented with an official certificate and documentation package.
LIFT_CareAgreement_mountains.jpg

Calibration Care Agreement

A Calibration Care agreement is the most economical and convenient way to maintain the accuracy of your Vaisala measurement instrument over its lifespan.

Paul Daniel, Vaisala

Senior GxP Regulatory Compliance Expert

Paul Daniel has worked in the GMP-regulated industries for over 25 years helping manufacturers apply good manufacturing practices in a wide range of qualification projects.  His specialties include mapping, monitoring, and computerized systems.  At Vaisala, Paul oversees and guides the validation program for the Vaisala viewLinc environmental monitoring system.  He serves as a customer advocate to ensure the viewLinc environmental monitoring system matches the demanding requirements of life science and regulated applications.  Paul is a graduate of University of California, Berkeley, with a bachelor’s degree in biology.

Comment

Thomascrive

Jul 25, 2018
This is my very first time visiting your site and I'm very interested. Many thanks for sharing and keep up ;)

Janice Bennett

Jul 25, 2018
Thank you very much for visiting our site! We are happy to provide information and know-how!

K subbanna

Nov 3, 2019
Dear Sir, I have performed Autoclave air filter sterilization study with out of 22 calibrated thermocouples 01, 02 were used,during post calibration 15th thermocople failed, what should I do? Whether can I justify with used thermocople (01,02) result are within acceptance or should I do anything more than that..pls reply..

Paul Daniel

Nov 6, 2019
Hello,

Thanks for your comment. I’m sorry to hear of your difficulties! In my past work validating autoclaves, I have seen issues like this before. Despite being inexpensive and tolerant of high temperatures, thermocouples have high drift and that can often lead to a failed post-calibration. To counteract this, I used to write into the protocol document itself a clause that would allow a small percentage of the thermocouples to fail post-cal, yet allow the study to pass. That works well for empty chambers, but is not as robust for loaded chambers, as there may have only been one thermocouple in the affected piece of equipment (filer in this case). Another approach is to analyze the post-calibration failure itself. If we assume the failure was linear and in only one direction, then a thermocouple that appears to be registering lower than expected temperatures during post-calibration would likely have under-reported temperature values, and thus under reported the accumulated lethality. In which case, you might be able to convince QA that the measurements would only have provided an additional challenge. There is no guarantee, because this is based on the assumption that the calibration failure would be in one direction only. If the failed thermocouple was reading too hot, then this won’t work at all. Your best course of action might be to simply repeat the study.

Paul Daniel

kathy shuras

Mar 22, 2022
i am doing a warehouse validation using 26 valprobes one passed pre calibration but failed to read data i was told that u are allowed 5 -10% of your thermocouples to fail is that accurate

Paul Daniel

May 27, 2022
Hello, thank you for this question! Sorry to hear about your failed probe. You need to refer to your mapping protocol and see what failure rate of sensors was allowed. Typically, I write an allowance of 10% into my protocols to deal with cases exactly like this. But if you your protocol doesn’t include an acceptable failure, you will likely have to treat this as a deviation, and see what explanation you can get QA to accept. Warm regards, Paul Daniel

Add new comment