Beyond Accuracy and Precision: Total Analytical Error and Linearity in ICH Q2 Revision 2

Key Learning Outcomes

• Understand how linearity is addressed in ICH Q2 revision 2, no longer as a standalone performance characteristic but as a property assessed within the working range section, and how this applies across linear, non-linear, and multivariate assay types.
• Learn why total analytical error, as a combined approach for accuracy and precision, is the most suitable framework for demonstrating that future reportable values will stay within acceptable limits, and how it relates to the prediction interval, tolerance interval, and confidence interval concepts.
• Discover how the Analytical Target Profile (ATP) defines target measurement uncertainty in relation to product specifications, and why setting individual accuracy and precision criteria separately can systematically underestimate the true combined measurement error.
• Explore how total analytical error validation designs can be efficiently structured, covering bias, intermediate precision, and linearity in a single integrated experiment, without significantly increasing the number of experiments compared to traditional separate approaches.

Event Overview

This session by Dr Pierre Le Brun offers a thorough yet practical approach to two of the most technically complex topics in analytical method validation: linearity and total analytical error. Referencing ICH Q2 revision 2 and a wide range of statistical literature dating back to the 1970s, the presentation clarifies common misconceptions and provides analytical scientists and statisticians with the conceptual and practical tools needed to validate methods that are truly fit for purpose.

The session begins by identifying linearity within the revised ICH Q2 structure, now integrated into the working range section instead of standing alone, and elucidating what linearity actually signifies for various assay types. For linear assays, the emphasis is on the results’ linearity profile: a plot of observed versus expected values compared against the identity line, with confidence intervals on the slope tested for equivalence to one. For non-linear and multivariate assays, the principle of proportionality between validation standard results and true sample values applies regardless of the calibration model, and residual behaviour across the calibration range becomes the key diagnostic.

The session then examines total analytical error, tracing its intellectual origins from Westgard (1974) through Boulanger and colleagues to the current ICH Q2 revision 2 language on combined approaches for accuracy and precision. A key theme is the prediction interval as the statistically appropriate tool for indicating where a single future reportable value will fall, in contrast to confidence intervals, which address parameter uncertainty, and tolerance intervals, which describe population proportions but do not reflect the reality of single-result reporting on a certificate of analysis.

The comprehensive analytical error profile, including prediction intervals calculated across the operational range and compared to the target measurement uncertainty, is presented as a unified framework that addresses linearity, accuracy, and precision simultaneously. The session explains why setting separate acceptance criteria for bias and CV at, say, 15% each, does not ensure a combined total error of 15%, and how the total analytical error approach clarifies this relationship and links it directly to product specifications through the Analytical Target Profile.

The presentation concludes with a practical comparison of validation experiment designs, demonstrating that an integrated total analytical error approach covering five concentration levels over several days in duplicate can achieve the same or higher statistical confidence as traditional separate designs — in as few as 30 to 32 experiments, well within the range of conventional methods.

Who Should Attend?

Anyone responsible for designing, executing, or reviewing analytical method validation who wants to confidently apply the ICH Q2 revision 2 framework — and move beyond checklist-driven validation towards a truly fit-for-purpose, statistically robust approach.

• Analytical Scientists and Method Development Chemists
• Validation and Compliance Scientists
• Statisticians and Data Scientists supporting analytical development
• Quality Control and Quality Assurance Professionals
• Regulatory Affairs and CMC Submission Specialists
• Biopharmaceutical and Small Molecule Development Teams
• R&D and Technical Operations Managers overseeing analytical functions

Unlock Additional Educational Resources

Register today for Pierre’s presentation and gain access to exclusive bonus content, such as the insightful panel discussions on “Understanding the ICH Q2 Guideline” and “Key Challenges and Solutions in Adapting to ICH Q2(R2)”.

Brought to you by: