I have got a question regarding the calculations of analytical validation parameters. Maybe one of you can help me?
In some guidelines it is stated to calculate the within run and between run accuracy. Searching the internet I have found publications were the between and the within run mean value and the corresponding within and between run accuracy/trueness and precision has been calculated. I have found several guidelines to calculate within run and between run precision but not the mean values and the accuracy.
Is it simple mathematics and I do not get it? Does anyone know how to calculate this? Are there any equations I can use`?
I have got a question regarding the calculations of analytical validation parameters. Maybe one of you can help me?
In some guidelines it is stated to calculate the within run and between run accuracy. Searching the internet I have found publications were the between and the within run mean value and the corresponding within and between run accuracy/trueness and precision has been calculated. I have found several guidelines to calculate within run and between run precision but not the mean values and the accuracy.
Is it simple mathematics and I do not get it? Does anyone know how to calculate this? Are there any equations I can use`?
Thank you![/quote]
Have you tried the ICH guidelines or USP <1225>? These will give you guidance on what is needed - depending on the governing body who will be assessing this Method Val. ICH would be pretty standard to follow.
ICH Guidelines for Accuracy pasted below:
Chapter4. ACCURACY
Accuracy should be established across the specified range of the analytical procedure.
4.1. Assay
4.1.1 Drug Substance
Several methods of determining accuracy are available:
a) application of an analytical procedure to an analyte of known purity (e.g. reference
material);
b) comparison of the results of the proposed analytical procedure with those of a
second well-characterized procedure, the accuracy of which is stated and/or
defined (independent procedure, see 1.2.);
c) accuracy may be inferred once precision, linearity and specificity have been
established.
4.1.2 Drug Product
Several methods for determining accuracy are available:
a) application of the analytical procedure to synthetic mixtures of the drug product
components to which known quantities of the drug substance to be analysed have
been added;
b) in cases where it is impossible to obtain samples of all drug product components ,
it may be acceptable either to add known quantities of the analyte to the drug
product or to compare the results obtained from a second, well characterized
procedure, the accuracy of which is stated and/or defined (independent procedure,
see 1.2.);
c) accuracy may be inferred once precision, linearity and specificity have been
established.
1 see chapters “Reporting Impurity Content of Batches” of the corresponding ICH-Guidelines:
“Impurities in New Drug Substances” and “Impurities in New Drug Products”
9
Validation of Analytical Procedures: Methodology
4.2. Impurities (Quantitation)
Accuracy should be assessed on samples (drug substance/drug product) spiked with
known amounts of impurities.
In cases where it is impossible to obtain samples of certain impurities and/or
degradation products, it is considered acceptable to compare results obtained by an
independent procedure (see 1.2.). The response factor of the drug substance can be
used.
It should be clear how the individual or total impurities are to be determined e.g.,
weight/weight or area percent, in all cases with respect to the major analyte. 4.3. Recommended Data
Accuracy should be assessed using a minimum of 9 determinations over a minimum of
3 concentration levels covering the specified range (e.g., 3 concentrations/3 replicates
each of the total analytical procedure).
Accuracy should be reported as percent recovery by the assay of known added amount
of analyte in the sample or as the difference bet
I am very sorry for the delayed answer, but I am still struggling with the definitions/calculations.
Is it possible to calculate within and between run bias for standards if you use new freshly prepared standards for every validation run and mwasure them on every validation run (in triplicate for example)?
or in other words:
Is it possible to calculate within and between run bias for QC samples if you prepare them on one day and measure them in >2 aliquots on several days?
For a method validation where the validation/QC samples are analyzed in multiple replicates (e.g., n = 3 or 5) on more than one batch you would calculate the same statistics for both within run (intra-run) and between run (inter-run). Those stats are the Mean, Standard Deviation, Coefficient of Variation (CV) and Accuracy. The accuracy is the mean expressed as a percent of the target (e.g., T=100, Mean = 95, Accuracy = 95%). If you want the bias that is usually the difference between the target and the mean expressed as a percent (e.g., T=100, Mean = 95, Bias = -5%).
In a method validation let’s say you ran 3 levels of QC samples each in reps of 5 for each run over a total of 3 runs to determine Accuracy and Precision. You would calculate the within run stats (Mean, SD, CV, Accuracy) for Run 1 with a N=5. For between run stats you would use all of the values for N=15…
Here’s an example from one of our method validations.
[quote][=sbonin;24290]For a method validation where the validation/QC samples are analyzed in multiple replicates (e.g., n = 3 or 5) on more than one batch you would calculate the same statistics for both within run (intra-run) and between run (inter-run). Those stats are the Mean, Standard Deviation, Coefficient of Variation (CV) and Accuracy. The accuracy is the mean expressed as a percent of the target (e.g., T=100, Mean = 95, Accuracy = 95%). If you want the bias that is usually the difference between the target and the mean expressed as a percent (e.g., T=100, Mean = 95, Bias = -5%).
In a method validation let’s say you ran 3 levels of QC samples each in reps of 5 for each run over a total of 3 runs to determine Accuracy and Precision. You would calculate the within run stats (Mean, SD, CV, Accuracy) for Run 1 with a N=5. For between run stats you would use all of the values for N=15…
Here’s an example from one of our method validations.
[/quote]
So if I am interested in within run statistics I calculate for each run the mean,sd so on three times,right? I have assumed, within run statistics are reported for every Run (in this example 3x) but everywhere I look only one mean/bias/accuracy is calculated. Should I then calculate the mean of the three means?
