Statistical Issues in Drug Development

A Two Day Course

(Sample outline and timetable)

A Course for Life Scientists & Statisticians Working in Drug Development
Course Leader
Professor Stephen Senn, Consultant Statistician,
Edinburgh, UK

Led by Stephen Senn, and based on his critically acclaimed Statistical Issues in Drug Development, this course is particularly suitable for life scientists who wish to deepen their appreciation of important statistical concepts and issues. It is also suitable for statisticians who wish to obtain a concise overview of controversies they may encounter.

The emphasis of the course will be on "why" rather than "how". Mathematical technicalities will be avoided but problems and issues will be frankly explored.

Suitable for: Research and Development Life Scientists; Clinicians; Clinical Trial and Project Managers; Clinical Research Associates; Statisticians; Health Economists; Clinical Data Processors, Managers and Quality Assurance; Administrators and Managers; Regulatory Affairs Personnel; Marketing Managers.

About the course leader. Stephen Senn, PhD, CStat, FRSE was formerly head of The Competence Center for Methodology and Statistic at the Luxembourg Institute of Health. He was previously a professor at the University of Glasgow and before that at University College London. He worked for many years in the Swiss pharmaceutical industry and consults regularly in Europe and America on statistical issues in drug development. He is the author of Cross-over Trials in Clinical Research (1993, 2nd edition 2002), Statistical Issues in Drug Development, (1997, 2nd edition 2007, 3rd edition 2021) and Dicing with Death (2003, 2nd edition 2022). He is the winner of the George C Challis award for biostatistics of the University of Florida (2001) and the PSI award for the most interesting speaker in 25 years of conferences (2002) and of the Bradford Hill Medal of the Royal Statistical Society (2008). In 2017 he gave the Fisher Memorial Lecture.

FOUR VIEWS OF STATISTICS IN DRUG DEVELOPMENT

9.00 Registration

9.30 Introduction

Professor Stephen Senn
Consultant Statistician, Edinburgh

9.40 HISTORICAL

Statistics and probability have intrigued mankind since the 17th century. Who were the key people and what were the problems they grappled with? This presentation will look at the development of statistics through time and explain how the work of Pascal, Fermat, Bayes, Fisher and Neyman have shaped modern day thinking in statistics and probability. - Pascal and Fermat get the dice rolling - The never ending story. Bernoulli and the lore of large numbers - An epistle from Dr Arbuthnot. Sex, statistics and the birth rate. - The Reverend Bayes and his irreverent followers - Square roots and diced vegetables: Fisher and agricultural experimentation - Pole position: Neyman and hypothesis testing - The neo-Bayesian movement - The rise of medical statistics

10:30 Quiz

10.45 Coffee

11.15 PHILOSOPHY

Within the process of Drug Development, Clinical Trials represent one of the most important stages simply because human subjects are involved. What is not always appreciated is that there are two general aims for clinical trials. First, to assess the effects of treatment in the patients actually studied and second to predict future effects of the treatments on patients NOT studied. This presentation will discuss and explain: - Causality: subsequence and consequence - The man who woke up Immanuel Kant. Hume s problem of induction. - Controlled inference - Regression to the mean: things aint what they seem. - Randomisation. Ration roulette - Blinding. The misunderstood placebo - Measuring effects: The dangers of over-interpretation

12.15 Discussion

12.30 Lunch

13.30 METHODOLOGY

This presentation will review some fundamental statistical concepts, and will illustrate the nature of an important divide in statistics between Bayesian or Subjectivist statistics on the one hand and between Classical or Frequentist on the other and how the different schools of thought impact on Drug Development. The following will be discussed: - A simple problem. Heavy metal meets statistical inference. - The Bayesian solution. Belief and logic. - Hypotheses tests. Lemmas and dilemmas - Significance tests. Fisher s legacy. - Confidence intervals. Estimates and guesstimates - Problems and solutions

14:30 Questions and answers followed by WORKSHOP

15.00 Tea

15.30 PROFESSIONAL

Regulatory agencies often expect to see formal overviews or meta-analyses of the results of drug development programs. These overviews are statistical in nature and so it follows that the time to start consulting the statistician is not when the first trial plan is being drafted, but even earlier and they should be closely involved in designing the development program from the beginning. - The role of the medical statistician - Five purposes of a protocol - Design issues - Sample size determination - Research and the pharmaceutical statistician

16:30 Question and answer session

17.00 Close

CONTROVERSIAL STATISTICAL ISSUES

09:00 ALLOCATING PATIENTS TO TREATMENT

- Randomisation - Placebo run-ins - Minimisation - Ethical problems - Bandit designs - Cut-off designs - Sequential designs

10:00 Questions and answers

10:15Coffee

10:45 BASELINES AND COVARIATE INFORMATION

- Baseline balance - Adjusting for baselines - Multiple baselines - Analysis of covariance - Late baselines

12:00 Lunch

13:15 MEASUREMENT OF TREATMENT EFFECTS

- Additivity - Individual response - Trend effects - Dichotomising continuous variables - Surrogate endpoints

14:15 MULTI-CENTRE TRIALS

- Fixed versus random effects - How to combine information - Effect reversals - Planning issues

15:00 CASE STUDY: A case study involving a multi-centre trial

15:30 Tea

16:00 DOSE-FINDING

- Regression versus factor approaches - Accompanying placebo designs - Designs in phase I in cancer - Parallel assay

17:00 Close

This course is also available in one day and three day versions.
Return to StatWise Courses