Statistical methods for design and analysis of precision medicine trials

Precision medicine is about going beyond assessing whether a new treatment works on average to predicting which subgroups of patients receive benefit and to what extent. When subgroups, often defined by biomarkers, genetic, phenotypic or psychosocial characteristics, are associated with a treatment’s efficacy or toxicity, precision medicine offers substantial advantages to patients, trial sponsors, and the wider healthcare system. However, a barrier to realising the promise of precision medicine is the inappropriate use of traditional clinical trial design and analysis, which rely on estimates of population-averaged effects.

In this course we introduce the concept of precision medicine and cover innovative approaches. This includes novel designs (basket, umbrella, adaptive signature and adaptive enrichment designs) as well as efficient analysis approaches (Bayesian hierarchical modelling and using instrumental variable methods to help estimate sub-group specific treatment effects in the presence of intercurrent events). These approaches have all been developed to improve the power, quality of information, and patient benefit provided by clinical trials. Examples from a wide variety of therapeutic areas will be discussed, with implementation in R software. Perspectives will be given on the future development of design, conduct and analysis of clinical trials in the field.

Course Tutors

Dr Haiyan Zheng, Newcastle University

Professor James Wason, MRC Biostatistics Unit and Newcastle University

Professor Christina Yap, Institute of Cancer Research

Professor Jack Bowden, University of Exeter

Intended Audience

Students and professionals with some knowledge of clinical trials and statistics are welcome. Statisticians will benefit the most from this course. Clinicians and trialists with good knowledge of statistics and computing will follow much of the course and will be exposed to a range of potential new methods. With the practical, participants will have a chance to implement the novel methods in concrete examples. Some of the course will be difficult to follow for individuals without a good knowledge of statistical theory or R software.

Prerequisites

Some basic knowledge of R (although full support given), knowledge of clinical trials and basic statistics. For some parts of the course, knowledge of Bayesian statistics is required.

Computing Practicals

Each lecture will have a computer practical in R. Most of the practicals will be possible to do without advanced knowledge of R as guidance will be given.

Course Objectives

1. To learn how traditional clinical trials are not always appropriate in the context of precision medicine.
2. To learn about new quantitative approaches for clinical trials where the aim is to find which types of patients a treatment works well for.
3. To explore how to design and analyse these trials using R.
4. To hear about some future directions of the area.

Software Download

Pre-course installation instructions will be sent out prior to the beginning of the course.

Registration

Registration link when available

This course will run in the same week as our 2 day workshop on Clinical trials in the era of precision medicine: new developments

Course Enquiries

For answers to course queries, please email courseadmin@mrc-bsu.cam.ac.uk