I don't like the idea of even associateing the word "Grace Period" with a calibration interval. As a metrologist, it strikes me that its introduction may bely a less than adequate understanding of really what a calibration interval is. Some people, I think, may perhaps think the interval is just some magic amount of time between the dates you are supposed to put a new sticker on an instrument. Among those who may believe that, are some that don't believe or understand the purpose of calibration. So it isn't a far stretch to think that some may think it perfectly reasonable to add a grace period.
But when you understand the meaning of the interval, which is essentially the period of time to a defined confidence under which an instrument is likely to yield in specification results. And if you go beyond that interval, it isn't "grace" but risk. So if a company wants something written into a quality manual about some added days allowable, it should be more properly called "increased risk period." Because when you extend the interval of a calibrated instrument, you are by definition increasing the absolute uncertainty of every measurement it makes, and increasing the risk of an out of tolerance measurement. This is why the FDA regulates it so tightly. Those who wrote those instructions understood I think, well, that to go beyond the due date is to incur an unaccep0table statistical risk. Who are we to second guess the OEM of a good instrument (who determined the interval through experimentation).
The only acceptable way to extend an instrument's calibration intercal is through historical data demonstrating that it still meets the defined confidence level of in tolerance measurements (which many instruments do). But it is not a good quality practice to arbitrarily extend intervals with no appropriate data (which is what it appears is the case with grace periods).
A little of a vent, but I believe this is correct.
But when you understand the meaning of the interval, which is essentially the period of time to a defined confidence under which an instrument is likely to yield in specification results. And if you go beyond that interval, it isn't "grace" but risk. So if a company wants something written into a quality manual about some added days allowable, it should be more properly called "increased risk period." Because when you extend the interval of a calibrated instrument, you are by definition increasing the absolute uncertainty of every measurement it makes, and increasing the risk of an out of tolerance measurement. This is why the FDA regulates it so tightly. Those who wrote those instructions understood I think, well, that to go beyond the due date is to incur an unaccep0table statistical risk. Who are we to second guess the OEM of a good instrument (who determined the interval through experimentation).
The only acceptable way to extend an instrument's calibration intercal is through historical data demonstrating that it still meets the defined confidence level of in tolerance measurements (which many instruments do). But it is not a good quality practice to arbitrarily extend intervals with no appropriate data (which is what it appears is the case with grace periods).
A little of a vent, but I believe this is correct.