Intermediate precision, statistical validation

Precision components are defined at three levels reproducibility, intermediate precision, and repeatability. Reproducibility is the variability of the method between laboratories or facilities. However, as a laboratory is not randomly selected from a large population of facilities, laboratory is a fixed effect. Consequently, the assessment of reproducibility is a question of comparing the average results between laboratories. Additionally, the variation observed within laboratory should be compared to ensure that laboratory does not have an effect either on the average result of the method or on the variability of the method. To assess reproducibility, conduct the same set of validation experiments within each laboratory and compare both the accuracy results and the precision results. If the differences are meaningful, analysis of variance (ANOVA) tests can be conducted to determine whether there is a statistically significant laboratory effect on the mean or on the variance of the method. For simplicity, the validation discussed within this chapter will not consider reproducibility and only one laboratory is considered. 

In a statistical model, fixed effects have an influence on the mean value or average of the method s response while random effects have an influence on the variability of the method. Fixed effects are assessed in the context of accuracy. Random effects are assessed in the context of precision and become the intermediate precision components. In designing the validation design matrix the validation assays need to be balanced over both the fixed effects and the random effects. A mixed effects model (or design) occurs when both fixed effects and random effects are present (6). 

The experimental design selected, as well as the type of factors in the design, dictates the statistical model to be used for data analysis. As mentioned previously, fixed effects influence the mean value of a response, while random effects influence the variance. In this validation, the model has at least one fixed effect of the overall average response and the intermediate precision components are random effects. When a statistical model has both fixed effects and random effects it is called a mixed effects model. 

Method validation for NIR or Raman spectroscopic methods using chemometrics is outlined in USP chapter <1119>. The criteria for method validation are the same as other quantitative analytical methods accuracy, precision, intermediate precision, linearity, specificity, robustness. However, because these methods are statistical in nature and are based on a previously validated analytical method, the validation of MVA methods is somewhat different than traditional analytical methods. 

Principles and Characteristics Whereas parameters most relevant to method development are considered to be accuracy, system precision, linearity, range, LOD, LOQ, sensitivity and robustness, method validation parameters are mainly bias, specificity, recovery (and stability of the analyte), repeatability, intermediate precision, reproducibility and ruggedness. However, method development and validation are highly related. Also, validation characteristics are not independent they influence each other. Acceptance criteria for validation parameters should be based on the specification limits of the test procedure. Quantitation and detection limits need a statement of the precision at their concentration levels. Procedures used for validation of qualitative methods are generally less involved than those for quantitative analytical methods. According to Riley [82], who has discussed the various parameters for validation of quantitative analytical methods, the primary statistical parameters that validate an analytical method are accuracy and precision. 

---