Laboratory, Medical and Device Performance and Validation following Regulatory and ICH Statistical Guidelines
This course is designed to introduce to individuals the understanding and interpretation of the statistical concepts one uses when investigating quantitative ICH Guidelines such as analytical methods validation, procedures and acceptance criteria in calibration limits, and process and quality control. One also considers ICH Q8 and Q9. These techniques covers both clinical and laboratory applications. This applies to many areas such as stability testing, outlier analysis and risk management. It is not a course in statistics but introduces the participant to an applied approach to the statistical techniques one uses, how they are reasonably interpreted. One will address the various challenges facing pharmaceutical and biotechnology companies when it comes to quantifying results in a meaningful interpretable manner through tabulations and graphical presentations.
In this two day workshop seminar one will learn the different regulatory agencies expectations of the quantification and development of a sound statistical monitoring of process control that are utilized, effective, and efficient. Participants will become familiar with the important aspects of the statistical methods and learn how these guidelines are applied in practice.
Evaluate linear quantitative measurement procedures and sources of error.
Distinguish the difference between confidence and tolerance intervals
Evaluate the appropriateness of the effect of sample size in given procedures.
Evaluate laboratory/clinical quality control based on risk management
Interpret statistical process control
Distinguish between FDA requirements and ICH guidelines
Day 1 Schedule
Overview of ICH Methodology
Introduction to the simple regression model
Interpreting the slope and intercept in validation procedures
Residual analysis and error detection
Stability and Potency issues
Outlier strategies using the linear model in calibration methods
Imputation techniques for missing data
Outlier strategies for non normal or ranked data
Consequences of outlier elimination/substitution
Sample size and analysis issues
Confidence and tolerance bounds on risk models
Parametric and non parametric (non normal data) procedures
Exact computational techniques
Day 2 Schedule
Discussion of risk management in general
Traditional risk management strategies in clinical settings
Predictive models in risk assessment
Discussion of the Design Space
Risk Management in pre-analytical phase of laboratory testing
Introduction to validation of models in hazard assessment and risk management
Demonstration of laboratory Validation procedures
Bivariate models and confusion matrices and derived statistics
Statistical process laboratory control and capability
Normal and non normal data procedures
Evolutionary Operations Process
Confidence and tolerance bounds on limits of risk
Al Bartolucci, Ph.D.
Dr. Al Bartolucci is Emeritus Professor of Biostatistics at the University of Alabama where he also serves as a Senior Scientist at the Center for Metabolic Bone Diseases, AIDS Research Center and Cancer Center.