Robustness characteristics measurement

Validation is the process of measuring the performance of a previously developed quantitative method. The characteristics measured to assess performance are defined and explained in detail later in this chapter. Typically, for a quantitative method the characteristics of specificity, accuracy, precision, and linearity are measured. The range of the method can be defined, or verified, by the region where the aforementioned characteristics are acceptable. Although their determination is more appropriately a development task, the characteristics of robustness, limit of detection, and limit of quantitation can also be verified during validation. 

Table 6.30 lists the main characteristics of quadrupole mass spectrometers. QMS is a relatively simple and robust analyser which does not need such a high vacuum as a sector instrument. The maximum admissible pressure at the source of the spectrometer is 10 6mbar in continuous regime and 10-5—10 4 mbar during short time intervals. Quadrupole technology assures reproducible and accurate molecular weight measurements day in and day out. For figures of merit, see Table 6.27. 

As a measuring science, analytical chemistry has to guarantee the quality of its results. Each kind of measurement is objectively affected by uncertainties which can be composed of random scattering and systematic deviations. Therefore, the measured results have to be characterized with regard to their quality, namely both the precision and accuracy and - if relevant - their information content (see Sect. 9.1). Also analytical procedures need characteristics that express their potential power regarding precision, accuracy, sensitivity, selectivity, specificity, robustness, and detection limit. 

The other obvious extension, which is more useful for the case where you may still have to measure samples with the same characteristics as the old ones, is to simply keep adding to and expanding the calibration set as new samples become available. The new samples then not only allow you to test for robustness, but inclusion of such samples will actually make the calibration more robust. I think we all know this intuitively, but I have also been able to prove this mathematically. 

In what follows we will discuss systems with internal surfaces, ordered surfaces, topological transformations, and dynamical scaling. In Section II we shall show specific examples of mesoscopic systems with special attention devoted to the surfaces in the system—that is, periodic surfaces in surfactant systems, periodic surfaces in diblock copolymers, bicontinuous disordered interfaces in spinodally decomposing blends, ordered charge density wave patterns in electron liquids, and dissipative structures in reaction-diffusion systems. In Section III we will present the detailed theory of morphological measures the Euler characteristic, the Gaussian and mean curvatures, and so on. In fact, Sections II and III can be read independently because Section II shows specific models while Section III is devoted to the numerical and analytical computations of the surface characteristics. In a sense, Section III is robust that is, the methods presented in Section III apply to a variety of systems, not only the systems shown as examples in Section II. Brief conclusions are presented in Section IV. 

If we try to refine the adequacy between the measurement procedures and the practical needs for wastewater quality monitoring, different metrological (analytical) characteristics have to be considered, such as detection limit, reliability and robustness (Table 3). Even if it is very difficult to compare the analytical methods carried out in the laboratory with on-site measurements (with on-line or tests kits), this presentation points out the main features of the measurement required for different needs. These characteristics define the quality of the available information [3], which constitutes one of the major problems that... 

Current tests are generally destructive (i.e., sample is altered or destroyed) and robust estimates of measurement system variability (all aspects of the procedure including the operators) are difficult to obtain without using methods such as Gauge R R-reproducibility and repeatability (14). Suitability of current methods then is based on calibration using a calibrator system that has its own built-in variability and other assumptions (e.g., in physical testing such characteristics as size, shape, density can alter aerodynamic and/or hydrodynamic behavior of materials in a test system and contribute to systems variability). 

The purpose of an analytical method is the deliverance of a qualitative and/or quantitative result with an acceptable uncertainty level. Therefore, theoretically, validation boils down to measuring uncertainty . In practice, method validation is done by evaluating a series of method performance characteristics, such as precision, trueness, selectivity/specificity, linearity, operating range, recovery, LOD, limit of quantification (LOQ), sensitivity, ruggedness/robustness, and applicability. Calibration and traceability have been mentioned also as performance characteristics of a method [2, 4]. To these performance parameters, MU can be added, although MU is a key indicator for both fitness for purpose of a method and constant reliability of analytical results achieved in a laboratory (IQC). MU is a comprehensive parameter covering all sources of error and thus more than method validation alone. 

The phrase related to implies that the relationship is known and valid. This will only be realized if the relationship at every step of the process is clearly defined and valid. Hence the requirement for an unbroken chain of comparisons. The parallel between these issues and those addressed by method validation is worth noting. Validation is the process of establishing that a method is capable of measuring the desired measurand (analyte), with appropriate performance characteristics, such as level of uncertainty, robustness, etc. It should also address systematic effects, such as incomplete recovery of the analyte, interferences, etc. These latter issues can be dealt with by designing a method to eliminate any bias, at a given level of uncertainty, or if that is not possible, to provide a means of correcting for the bias. This may be done at the method level, by applying a correction factor to all results, or at the individual measurement level. 

To illustrate MCD-regression, we analyze a data set obtained from Shell s polymer laboratory in Ottignies, Belgium, by courtesy of Prof. Christian Ritter. The data set consists of n = 217 observations, with p = 4 predictor variables and q = 3 response variables. The predictor variables describe the chemical characteristics of a piece of foam, whereas the response variables measure its physical properties such as tensile strength. The physical properties of foam are determined by the chemical composition used in the production process. Therefore, multivariate regression is used to establish a relationship between the chemical inputs and the resulting physical properties of foam. After an initial exploratory study of the variables, a robust multivariate MCD-regression was used. The breakdown value was set equal to 25%. 

---