VPAQ  Guided tour
Introduction
In this page we propose a quick presentation of the validation steps of an analytical procedure with the software SOSstatVPAQ ®.
The steps of the analysis
The validation of an analytical procedure is sequenced in 4 steps that can be followed using the navigation panel located on the left of the window.
 Parameterization of the validation protocol by defining the experimental design (number of concentration levels and number of repetitions of measurements) and definition of the validation criteria
 Calibration and calculation of the response function
 Validation of the analytical procedure with calculation of the accuracy profile
 Export the validation report
1  Settings of the analysis
The identification panel is used to enter informative data on validation. These administrative data make it possible to clearly identify the purpose of the study and those responsible for it.
In a second step, we can specify the experimentation protocol.
 Unit of Concentration Levels and Response
 Calibration with or without a matrix
 Choice of standard or custom validation protocols by specifying Number of series, number of levels and number of measurement repetitions
Finally, the acceptance criteria are determined:
 The $
\alpha
$ risk (used to define the confidence level $\beta = 1\alpha
$ )  The limit of acceptance. The most commonly accepted limit values are:
 2% for bulk,
 5% for pharmaceutical measures
 15% for biological analyzes and 20% near the LOQ
2  Calibration
The aim of the calibration step is to compute the response function which establishes a relationship between the reading quantity and the study quantity. This function is then used during validation to predict the value of the samples presented. Several mathematical models can be used to identify the response function, this model can be changed at the validation stage to obtain the best accuracy profile. For more details, see the page recalling the principle of validation par l'erreur totale (or accuracy profile).
The calibration panel is organized in two parts:
 The left side shows datagrids to enter calibration measurements (with or without matrix).
 The right panel contains the graphical and numerical results of the calibration
 Calibration curve
 Résiduals
 Models of the response functions for each series.
If the option to detect outliers has been checked, VPAQ checks the homogeneity of the samples.
3  Validation
The validation panel has the same organization as the calibration panel: with the validation design of experiment datagrid on the left and tabs to view the results on the right.
Validation consists of evaluating the different measurement uncertainties from a new set of experiments. The data grid has the same columns as for calibration.
The user can select to do a validation with and without a matrix to evaluate the matrix effect. Validation of an analytical procedure may require testing of several response function models. VPAQ offers the possibility to test these models one by one or to determine the most appropriate based on performance indicators.
At the end of the validation, if a bias is highlighted (matrix effect for example), it is possible to apply a correction on the measurements.
One can quote among the graphical results:
 The absolute and Relative Accuracy of measurements based on concentrations (The accuracy of a result expresses the narrowness of the agreement between the test result and the accepted reference value).
 The Error Profile is used to display the magnitude of the measurement error according to the concentration values. The total error of an analytical procedure evaluates its ability to produce accurate results.
 The Accuracy Profile is the validation summary graph. It allows to validate the analytical method but also to define the quantification limits (calculated automatically).
In the numerical results, we find:

Criteria of Trueness of the measure:

Absolute bias: Average difference between the concentration estimate and the actual concentration
 Relative Bias: Bias expressed as a percentage of actual concentration
 Recovery: Percentage of actual concentration found in the estimate

Precision criteria for the measure:

Standard deviation and coefficient of variation of repeatability (intraseries)
 Standard deviation and interseries coefficient of variation
 Standard deviation and coefficient of variation of intermediate precision

Summary indices :

Total error
 Relative error
 Limits of the tolerance interval
 Limits of quantification
 If the outlier detection option has been checked for validation, VPAQ checks the homogeneity of the samples after realigning the points.
 Table of performance indicators of the method to select the best response function.
4  Report
VPAQ offers several export techniques for archiving results:
 A VPAQintegrated word processor lets you design a customized validation report.
 Exporting a preformatted PDF report allows you to quickly present the results of your analysis, including results tables, graphs and definitions. The PDF document can be passwordprotected.