Posted by Sten Westgard, MS
Another question coming in from one of our website members:
Could you give me some suggestions on establishing QC ranges for unassayed chem. Controls?
We currently use [Brand X] unassayed chem controls (much less in cost than assayed controls) for some of our chemistry analytes.
For these controls, we are provided a “target mean” and a range.
Occasionally, our established mean (i.e. n=30) for a new lot is outside of the range provided for “target mean” provided by manufacturer.
Could you suggest guidelines for acceptance of lab established means for unassayed controls?
Obviously, I would like to know how far from the “target mean” could the new lab established mean be?
The answer, following the jump...
The general guideline is to obtain a minimum of 20 measurements over a 20 day period to establish a new mean and SD.
If the established mean does not fall within the expected target range, the material could still be used for QC, given that future measurements are evaluated against the established mean and SD, not the target mean. However, if there is a large difference from a manufacturer's target mean, that suggests a possible systematic error or bias. It would then be helpful to have an assayed material or even a PT survey sample that has an established value to assess whether the bias is consistent between materials or unique to the one control material. In the latter case, the manufacturer may have some ideas why this could occur.
The new C24-Ed4 document has several pages on establishing the mean and SD for control materials.
Thanks for the question, and keep them coming.
James O. Westgard, PhD