ISO 17025 Para. 5.4.5 - Method (Calibration Procedure) Validation Requirements

[LClaeys]

My company asked me to put together a quality manual according to ISO17025.

I'm trying to understand the validation of methods and what a method is.

Example: Customer send us a voltmeter. I will calibrate this voltmeter using manufacturer's specifications. I write a procedure in Met/Cal that connect a calibrator to the voltmeter and test a certain amount of test points.
At this point I just use my procedure and calibrate the unit, any errors of the use of standards or the procedure I fix along the way since I need to use the unit to see if my standards and procedure is correct.

According to 5.4.4 you need to validate a method before use?

So is my method: calibration according to manufacturer's specifications?

And how do I validate my method?

 

[dwperron]

This is one of those 17025 paragraphs that you can conduct a seminar on.
First, a "method" is the procedure you will use to calibrate the instrument. The earlier sections of 5.4.X talk about choosing an appropriate method, or procedure, that is adequate for your needs. There is an emphasis on using "accepted" procedures, but the lab can develop their own procedures. The notes after section 5.4.4 give guidance on what should be included in the procedure. And yes, you must have a documented and controlled procedure so that the calibration can be duplicated. Your Met/Cal procedure will satisfy this requirement, but you must document that this is the procedure to be used, exercise revision control, etc.

Concerning validation, 5.4.5.2 says you shall (must) validate "non-standard methods, laboratory-designed / developed methods, standard methods used outside their intended scope, and amplifications and modifications of standard methods". That means that if you use one of the procedures that is taken from a reputable organization (manufacturer procedure, GIDEP, ASTM, AMSE, etc.) and use it as written you do not need to validate it. If you deviate from the procedure (delete test points, add test points, substitute equipment, etc.) you must validate that your changes still will provide valid results for your intended uses. For instance, if you delete tests from the original you must show that this will still be valid for your uses. If you substitute equipment you must show the equipment you are using is adequate for the job. If you add or modify tests you must show how they are adequate to meet your purposes. In your case, since you are creating your own Met/Cal procedure, it must be validated.

The validation of the procedure is "the provision of objective evidence" that the method works, and as such must be documented and retained as a record. The Note 2 of 5.4.5.2 spells out the validation techniques to be used. In your case here the simple way would be to perform the same calibration steps manually and compare your results to the Met/Cal results. In other cases you can use different standards to see if you get the same results, show how this same method works for similar instruments, or some other method that works for you to show that you get consistent and adequate results using your method. Be sure to record your validation results as evidence that you performed the validation.

This isn't really as bad as it looks: the last section of 5.4.5.2 Note 2 says that your validation technique can include "scientific understanding" and "practical experience", so if you can demonstrate that you know what you are doing, that you understand the process and the risks to the measurements involved, know your measurement requirements, and that you have documented your validation you should be all set.

 

[Hershal]

My company asked me to put together a quality manual according to ISO17025.

I'm trying to understand the validation of methods and what a method is.

Example: Customer send us a voltmeter. I will calibrate this voltmeter using manufacturer's specifications. I write a procedure in Met/Cal that connect a calibrator to the voltmeter and test a certain amount of test points.
At this point I just use my procedure and calibrate the unit, any errors of the use of standards or the procedure I fix along the way since I need to use the unit to see if my standards and procedure is correct.

According to 5.4.4 you need to validate a method before use?

So is my method: calibration according to manufacturer's specifications?

And how do I validate my method?

Very simple. Manufacturer states certain things. Now, in your internal procedure, refer to the manufacturer procedure. That takes care of the validation part.

There is an exception, where you may need different data points. Those specific points you must validate. Submit a plan to management listing points or methods and how results may be verified (e.g., another lab). Provide results to management, they sign off, validated