Autorentext

Riccardo Bonfichi is an independent consultant specializing in the application of statistical methods to pharmaceutical manufacturing and control, with over 35 years of experience in GMP-regulated environments across Italy and abroad, including roles at Sandoz Pharma AG in Basel. Over his career he progressed from analytical scientist to senior quality roles - Head of Quality Control, Head of Quality Assurance, and Qualified Person - responsible for the batch release of sterile active ingredients and medicinal products. He now supports pharmaceutical companies in Italy and internationally as a freelance consultant and trainer, focusing on Monte Carlo simulation, resampling methods, and statistical decision-making in QA/QC, and he regularly delivers webinars and courses on these topics for industry associations and regulatory bodies.

Klappentext

In pharmaceutical manufacturing, quality control, and quality assurance, critical GMP decisions are often made with incomplete data, variable processes, and uncertain measurements. Batch release, process validation, continued process verification, analytical data interpretation, sampling decisions, and stability assessment all require more than a deterministic "pass/fail" view. This book introduces Monte Carlo simulation as a practical framework for converting variability, uncertainty, and risk into quantitative, decision-relevant information.

Unlike general texts on simulation or pharmaceutical quality systems, this book focuses specifically on GMP manufacturing and control. It connects statistical modeling with real pharmaceutical problems, showing how stochastic simulation, bootstrap methods, uncertainty propagation, and predictive risk analysis can complement traditional QA/QC tools and support transparent, scientifically defensible decisions.

Key features include:

• Practical Monte Carlo case studies including assay, dissolution, process validation, CPV, stability, sampling plans, microbiological counts, capability, and measurement uncertainty
• Reproducible examples in R, requiring no proprietary statistical software
• Clear distinction between variability and uncertainty, with direct GMP interpretations
• Coverage of OOS probability, confidence intervals, non-normal data, overdispersion, bootstrap capability, empirical OC curves, expected loss, and predictive control
• Quantitative support for comparability assessments, manufacturing changes, and post-approval lifecycle decisions
• Alignment with modern risk-based thinking, including ICH Q9/Q10/Q11/Q14, FDA process validation, and USP / Written for QA/QC professionals, manufacturing and process engineers, regulators, and statisticians in pharmaceutical and chemical-pharmaceutical environments, the book is a practical companion for moving from qualitative risk descriptions to quantitative, reproducible decision support. Its central message is simple: better GMP decisions can be made when uncertainty is explicitly modeled rather than ignored.

Inhalt

Preface 1 Introduction to Monte Carlo in Pharma 2 Random Numbers vs. Random Variates 3 Simple Distributions 4 The Transfer Equation 5 A Complete Monte Carlo Simulation in R 6 Analysis of Simulation Results 7 Case Study 1 - API Assay in Tablets 8 Case Study 2 - Dissolution with Noyes - Whitney Law 9 Case Study 3 - From 3 Batches to Continuous Confidence (Monte Carlo & Bootstrap) 10 Case Study 4 - Capability Indices in Pharma: When Few, Skewed Data Challenge Model Distributions 11 Case Study 5 - Predictive Stability Control with Monte Carlo 12 Case Study 6 - Monte Carlo Sampling Plans and Empirical OC Curves 13 Case Study 7 - Microbiological Counts: Overdispersion via Monte Carlo Simulation 14 Case Study 8 - Monte Carlo Percentile-Based Capability for Non-Normal Data 15 Case Study 9 - Monte Carlo Measurement Uncertainty & Risk of Non-Compliance 16 Understanding Uncertainty Analysis in GMP Models 17 Case Study 10 - Decision Under Uncertainty: Expected Loss and Regulatory Trade-offs via Monte Carlo Simulation 18 Case Study 11 - From CPV Data to Predictive Control 19 Decision and Risk 20 Toward Probabilistic GMP Systems 21 Conclusions References