Confidence Intervals for Validation of Analytical Procedures Under ICH Q2 ( R2 )

The International Conference on Harmonisation (ICH) adopted revision 2 (R2) of its Quality Guidance Q2 Validation of Analytical Procedures in 2023. The revision includes a new statement that confidence interval limits for validation performance characteristics (i.e., accuracy and precision) should be compatible with acceptance criteria, unless otherwise justified. It also allows for sponsors to use prior knowledge (e.g., from development or previous studies) to support the validation study conclusion. These two new aspects of ICH Q2(R2) present both opportunities and challenges for the validation of analytical procedures. In this paper, we provide comprehensive examples of how to compute confidence intervals for key validation performance characteristics under different modeling approaches. We also describe methodologies for combining prior knowledge with the validation study results, both from frequentist and Bayesian perspectives. The paper is written in tutorial style and is aimed at statisticians and analytical scientists responsible for validating analytical procedures in compliance with ICH Q2(R2).