Method Validation module
The method validation criteria that can be evaluated with the EffiValidation software are:
- Repeatability
- Accuracy
- Linearity
- Reproducibility
- Selectivity/Specificity
- Limit of Detection/Quantification
- Robustness/Ruggedness
- Sensitivity
- Range
- Intermediate precision
Repeatability
Repeatability is the precision obtained under the best possible circumstances (i.e. within the same laboratory, by the same analyst, by the same instrument, and within one day when possible) and measures the size of the random error included in the method results. Statistically, the precision is the dispersion of results around the mean, irrespective of whether the mean is a correct representation of the true value. With EffiValidation software, there are two ways of estimating repeatability:
- Level by level from multiple measurements
- From parallel measurements
Level by level from multiple measurements: results from repeated measurements, obtained for the given concentration level, are entered in rows. Each row represents one concentration level. Please see the example below:
Repeatability values are calculated as the standard deviation and relative
standard deviation for each level (row). The repeatability of the method is
calculated as a polled variance/standard deviation
EffiValidation software additionally automatically checks data for outliers (by means of the Cochran, Grubbs and Dixon test) and for data normality. Results are presented within other tabs of the same Results window, OUTLIERS and NORMALITY:
Accuracy
Accuracy is the closeness of agreement between the test results and the true/accepted/assigned/reference value. If statistical testing shows that the difference between the test results and the reference value is not statistically significant, the validated method is considered accurate. When the difference is found to be statistically significant, the method does not give accurate, but biased results (results include a systematic error). In the EffiValidation software, there are several ways of estimating accuracy depending on the situation:
- Limited Concentration Range – Sample Reconstitution Possible
- Limited Concentration Range – Reference Material Available
- Large Concentration Range – Blank Material Available: t-test
- Large Concentration Range – Blank Material Available: Regression
- Large Concentration Range – Blank Material not Available: t-test
- Large Concentration Range – Blank Material not Available: Regression
- Comparison of Two Methods/Laboratories: t-test Level by Level
- Comparison of Two Methods/Laboratories: t-test on Difference of Results
Linearity
The linearity of a test method is its ability (within a given range) to obtain results directly proportional to the concentration (amount) of analyte in the sample. Statistically, Linearity describes the degree of the linear relationship between the test results and the concentration. To evaluate Linearity, one of the following methods can be used in EffiValidation software:
- Correlation and QC Coefficient
- ANOVA for Lack of Fit
- Significance of the Quadratic term – F-test
- Sign Test
Reproducibility
Reproducibility is the precision obtained under the worst possible circumstances (i.e., in different laboratories, by different analysts, using different instruments, within a long time interval) and measures the size of the random error included in the method results. Statistically, precision is the dispersion of results around the mean, irrespective of whether the mean is a correct representation of the true value. In the EffiValidation software, there are two ways of estimating reproducibility:
- Data Available
- Results Available
Selectivity/Specificity
Selectivity is the ability of the method to provide a signal that is selective with respect to the property of interest (e.g. the concentration of the analyte of interest). If the method is selective, the influence of potential interferences (matrix) on results is negligible. When the influence of interferences is significant, then the method cannot be considered selective. To evaluate selectivity/specificity, the following approaches can be used:
- Comparison of calibration lines
- Comparison of results with a standard
- Comparison of results with measurements without the interference
Limit of Detection/Quantification
The limit of detection is the level above which the signal of a sample can be reliably distinguished from a blank – this means that starting from this level a qualitative determination can be carried out. The limit of quantification is the level above which a quantitative determination can be carried out with sufficient reliability. Both limits are determined as a multiple of a standard deviation of blank, from the calibration line or from baseline noise in chromatography. To evaluate the limits, the following methods can be used:
- 3s – IUPAC
- 3s – Every sample corrected for blank
- 3s – Continuously measured blank
- From calibration line
- From signal of blank in chromatography
Robustness/Ruggedness
Robustness is the resistance of a method to small changes in the settings of experimental conditions (e.g. to changes in the pH of the mobile phase in the case of the HPLC method, to changes in the detector temperature of a GC method, etc.). Robustness is evaluated by means of experimental design. Two methods can be applied using EffiValidation:
- Dongs algorithm
- AOAC with evaluation
Sensitivity
Sensitivity is defined by IUPAC as the change in signal caused by a unit change of concentration/amount (i.e. the slope of the calibration line). Alternatively, sensitivity is evaluated as the minimum difference in concentration between two samples that can be reliably distinguished with the given method. To evaluate sensitivity, two algorithms can be used in EffiValidation:
- From Calibration Line
- Minimal Difference in the Validated Property
Range
Range characterizes the range of concentration within which validation parameters such as repeatability, accuracy, linearity, etc. have been evaluated and verified.
Intermediate precision
The intermediate precision characterizes the spread of test results around the mean value caused by random errors. Data to evaluate intermediate precision will be recorded under measurement conditions that are between Repeatability and Reproducibility conditions – this means e.g. different analysts, the same instrument, within an intermediate time period.
Chromatographic system suitability
Chromatographic system suitability parameters are used to verify that chromatographic characteristics are sufficient, a good separation, peak shaping, and the repeatability of injection can be obtained. In the EffiValidation software, there are two types of guidelines used to evaluate chromatographic system suitability.
Ph.Eur.:
- Theoretical Plate Number – PhEur
- Resolution – PhEur
- Symmetry Factor – PhEur
FDA:
- Capacity Factor
- Resolution
- Tailing
- Theoretical Plate Number
A complete case study is available here: GC headspace determination of
residual solvents in a drug substance.
