Ok, so I have met some people that LOVE to calibrate as often as possible which I think is an absurd thing to do. Not only does that cost a lot of unnecessary money, it is introducing more variables into your testing since every calibration is prone to human and instrumentation errors.
I treat calibration as a ritual and ensure that every step is taken to ensure the most accurate calibration and when it is done, I will run tests to ensure the previous curves results and the results of the new curve are within an acceptable variance. Once this is done, the calibration will easily last one month. If the continuing calibration verification samples are looking perfect, why recalibrate?? Sometimes regulations mandate recalibration every X amount of time but oftentimes they do not.
What are your thoughts on this? I want to hear both sides of this argument.
I would say that you need to run a new calibration curve every time you run samples for potency/assay testing. Since working conditions may vary from analysis to analysis, day to day, and analyst to analyst… Every time assay samples are introduced for quantification, you should run a new calibration curve.
Think of it this way… The last time you performed a calibration curve, the chromatographic column was in different condition, lamp energy of the detector was at different intensity, the mobile phase was prepared by a different technician, and so on and so forth.
Generally, a unique calibration set is performed with each running sequence, just to eliminate all of the above mentioned variables.
It may be time and money consuming, but in my opinion it definitely eliminates lots of potential variables.
There are always System Suitability Test Runs to check the current condition of the instrument, and also acceptance criteria such as factor of correlation (R-squared) from the linear regression of the calibration curve, to exclude human errors in preparation of calibration standards.