MSc Health Data Science / Course details

This unit aims to introduce the application of statistical ideas and methodology to randomised controlled trials.

This unit aims to introduce the application of statistical ideas and methodology to randomised controlled trials.

• Introduction to the application of statistics to randomised controlled trials (RCTs)
• RCTs: historical background and ethical issues concerning randomised experimentation on human participants.
• Design and organisation of RCTs. Types of bias and methods for controlling bias in RCTs, including blinding and placebo treatments.
• Critical reviewing publications of RCT – key issues.
• Sample size estimation for continuous and binary outcome measures.
• Methods of treatment allocation including simple randomisation, random permuted blocks, stratification and minimisation.
• Statistical methods for the analysis of parallel group trials including methods for the adjustment for baseline data.
• Implications of protocol deviations and the motivation for the intention-to-treat principle.
• Multiplicity issues: sub-group analysis and multiple outcomes.

A blended learning approach will be used throughout the 11 weeks delivering the new content.

Self-directed learning:

Student review of approximately 1.25 hours of online lecture materials per week. Additionally, students are expected to review lecture slides, including attempting accompanying exercises/ examples (including reading journal papers set for review), and reviewing supplementary notes.

Face-to-face contact:

Review session, complementing online lecture material (additional examples, group discussions); group tutorial (10-20 students per group), supporting students working through solutions to an exercise sheet.

There will additionally be one revision week (with two hours of face-to-face contact).

Recognise, choose and apply statistical methods appropriate to a trial, based on its design, objectives and outcome measures.

Critically appraise the design, analysis and interpretation of results of randomised controlled trials (RCTs), with specific reference to minimising bias.

Analyse data from RCTs of various designs (including: parallel, crossover or cluster; superiority, non-inferiority or equivalence), including the meta-analysis of several RCTs, and interpret the findings in the relevant context

Choose, with justification, the design, including calculation of the sample size, for a RCT (parallel, crossover or cluster; superiority, non-inferiority or equivalence) with a continuous or binary outcome.

Primary Text

John Matthews. An Introduction to Randomised Controlled Trials. Taylor and Francis London (2nd edition, 2006; 1st edition, 2001).

Supplementary Texts

Michael Campbell, David Machin, Stephen Walters. Medical Statistics. John Wiley London (5th edition, 2021).

Brian Everitt, Andrew Pickles. Statistical Aspects of the Design and Analysis of Clinical Trials. Imperial College Press (Revised edition, 2004).

Examples of journal reports of RCTs that may be used for critical appraisal exercises:

Quist-Paulsen P, Gallefoss F. Randomised controlled trial of smoking cessation intervention after admission for coronary heart disease BMJ 2003; 327:1254.

Melchart D, Streng A et al. Acupuncture in patients with tension-type headache: randomised controlled trial.  BMJ 2005; doi:10.1136/bmj.38512.405440.8F.