Selecting System Precision Criteria - System Suitability

[v9991]

Before analysing the samples on the instrument, it is recommended to evaluate the repeatability/sensitivity specific to the method-parameters.

this requirement could be more relevant&specific for pharmaceutical industry as it handles little more complex samples. (this is in addition to the routine calibration - performance verification steps )

( quick ref... (broken link removed)...(broken link removed)...(broken link removed)...article...summary)

Question :-
in order to ensure repeatability following methods are currently in use.
1. calculate RSD of consecutive-multiple-measurements (typically 3) of a known standard. (typically <2%) (Yes, I agree that the no. of measurements are proportional to the sensitivity)
2. ratio of two consecutive measurements. (typically the ratio of two measurements to be 0.98 - 1.02 )

which one of above is superior or preferred.

Here's my quick assessment.,
* for RSD (<2%) , the max range possible is +/- 2%
* however for ratio (+/- 0.02), the max range possible is +/- 1%)

is there a way to assess these scenarios of controls.

 

[v9991]

assessment of precision :- RSD vs similarity factor

I think is more precise and crisp question.,
does the enclosed approach comparison criteria hold any water for preferring similarity factor over %rsd

on left part of the image, is a simulation of possible 3-consecutive values, and the assessment on the range of values (min-max) allowed for given 10% RSD; and similar assessment on the range for possible combinations of values with acceptable similarity factor range.
and point is .... similarity factor allows a 10% difference between 1st and 2nd value; wherein 10% (3 values) allow ~17% range of values.

( yes, there could be other parameters to be considered., viz., no. of replicates for given RSD etc.,)

but, for given RSD vs similarity factor, does the approach sounds an fair step, and are there any other comparison/assessment criteria.


thank you

 

[v9991]

Re: assessment of precision :- RSD vs similarity factor

are there any more-formal ways to compare the 'control' achieved through "Relative Standard Deviation" versus "ratio of injections" route?

thank you in advance, for your thoughts and feedback.