Validation for unknow amount of compounds

[johncroker]

Hello Community.

I used to work on the petrochemical field, which have very solid methods and standards.

Now I was hired by a company to develop some methods and I am having some fun to solve this question for myself. I am used to work more in developing but I never worked with validation before.

How to validate a method if the sample matrix are very diversified and has different amounts of the target compound ?

Most of the matrices are plants and the target compound range can be from 0% to 10%.
How can I choose the range to calculate Accuracy for example if I cannot decide what is the % of nominal substance because for each plant the compound% is different.

Should I prepare the spike solutions from the maximum expected amount ? But a method validated for he maximum expected amount will work well when we have traces of the target compound ? Should I do range validations, like 3 or 4 validations for different max expected amounts ?

And what about contaminants ? Compounds with unknown amounts ? What should be the spike amount to use and validated the method ?

Thank you all.

[DR]

S-Matrix Fusion, Empower's Method Validation Manager.

Things like this use Design of Experiments theory to assist with such situations and complications. Either that, or get really good with Minitab or similar software.

[johncroker]

Thank you Dr.

I used to work with Minitab, but that was more than 15 years ago

I will take a look into those softwares, thank you.

[johncroker]

For anyone else with the same doubt, I am posting what I decided to do.

As the range of concentrations can be widely spread between target compounds, I decided to validate all substances in a same concentration range in the calibration curve.

If a high concentrated sample overpass the signal on a curve, I will say it overpass the range for the validated method and I will just dilute it to try to fit in on the validation curve.

That will eliminate the need for multiples ranges validated methods and calibrations.