Are yearly MDL calculations required?

[anionman]

The lab where I work does not report detect/non-detect when the concentration of an analyte is below the reporting limit. My understanding is that if a lab does not report below the LOQ or reporting limit then yearly MDL studies are not necessary. The TNI auditors have still been requiring us to determine the MDL every year using the last 2 years of data. That means I may have to go hunt down data packets from 2 years ago to see if the data is acceptable. It's a waste of time.

[vmu]

You have to confirm that your LOQ is always not higher than the reporting limit. For instance, the methods used for the determination of impurities in pharmaceutical products usually include the S/N check (with the limit "not less than 10") at the LOQ level as a part the system suitability test.

[anionman]

We check the LOQ which is our reporting limit in every run. Our real LOQs could be quite a bit lower but we run the low calibration standard as our reporting limit standard every day to make sure we are seeing it accurately. We also run blanks and make sure they are less than half the reporting limit. They fail about once a year. It's almost always because of some contamination issue like working with a 100 ppm Fluoride solution and not cleaning up properly.

Sometimes a new analyst will accidentally run a reporting limit standard rather than a blank and export that into the system. It shows up as a bad blank and then I have to find the data packet from 1-2 year earlier and see if I should use than in the calculation. But a lot of times they don't bother to examine why it failed by rerunning the vial and writing in the report that it was mislabelled. My understanding is that data points can be excluded if they result from documented instances of "gross failures," such as instrument malfunctions, mislabeled samples, or cracked vials.