Design and Analysis of Clinical Trials

A ten-week graduate course

Welcome

This is the online version of Design and Analysis of Clinical Trials: A ten-week graduate course by The rgtlab Curriculum Project.

The book covers the design, conduct, monitoring, analysis, and reporting of randomized clinical trials, written for first- and second-year graduate students in biostatistics, epidemiology, and public health, and for medical students with an interest in research. The core is structured for a single ten-week academic quarter: one chapter per week, each roughly two hours of lecture and four to six hours of assigned work. Eighteen further chapters treat advanced topics in depth for readers who continue past the quarter.

The organizing claim of the book is that a trial is an experiment whose design decisions are made once, in advance, under uncertainty, and cannot be revised after the first patient is randomized. Everything the book teaches follows from that constraint: why the estimand is specified before the sample size, why the randomization procedure dictates the analysis model, why interim looks must be priced into the type I error before the first look, and why a pre-specified analysis of imperfect data is worth more than a perfect analysis chosen after seeing the data.

The ten weeks are:

1. The trial as an experiment: history, ethics, regulation.
2. The protocol, the estimand, and the endpoint.
3. Randomization and allocation concealment.
4. Blinding, bias, and trial conduct.
5. Sample size and power.
6. Phase I: dose-finding designs.
7. Phase II: screening designs.
8. Phase III: the primary analysis.
9. Interim monitoring, group-sequential, and adaptive designs.
10. Missing data, sensitivity analysis, and reporting.

The book sits in the rgtlab graduate sequence:

• R for Biostatistics: A One-Week Boot Camp, pre-program preparation in the R language.
• Biostatistics Practicum, workflow and reproducibility infrastructure.
• Statistical Computing in the Age of AI, introductory methods and computing.
• Advanced Statistical Computing in the Age of AI, advanced numerical and Bayesian computation.
• Applied Statistical Methods for Health Sciences Research, the applied-methods axis, whose two clinical-trial chapters this volume expands into a full course.
• Applied Generative AI for Health Sciences Research, generative AI as the orthogonal axis.
• Design and Analysis of Clinical Trials (this volume), the trials axis.

See the Preface for motivation and the Conventions page for visual cues.

License

This book is licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International.

Code samples are licensed under Creative Commons CC0 1.0 Universal, i.e. public domain.