On 15 May 2014, the Royal Statistical Society's Young Statistician's section (YSS) and the Society's Primary Health Care special interest group held a joint meeting at Leeds University titled 'Beyond the two-arm randomised controlled trial workshop'.
After an introductory talk by the YSS, Sandra Eldridge, professor of biostatistics at the Centre for Primary Care and Public Health, the Blizard Institute, Barts and the London School of Medicine and Dentistry, gave the first talk summarising the planning of cluster RCTs, sample size and intracluster correlation coefficients (ICCs). Sandra was followed by Richard Hooper, a senior lecturer in medical statistic also from the Blizard Institute, who gave a presentation on sample size by simulation.
These were followed by four case studies examining sample size calculations in four unusual trial designs. Duncan Wilson, from the Leeds Institute of Clinical Trials Research (LICTR) gave a troublesome example of an optimal phase II Bryant & Day trial design, while Isabelle Smith, also of LICTR, introduced the ALPHA trial, an adaptive RCT which conducted interim analysis to review sample size. Neil Corrigan of LICTR looked at the 'surgeon effect' and sample size considerations in surgical trials, while finally, Dan Green of Keele University looked at the interim analysis of POST, a musculoskeletal randomised cluster trial. A short discussion followed, prior to breaking for lunch.
The afternoon session consisted of three talks; the first by Amanda Farrin, professor of clinical trials and evaluation of complex interventions at LICTR, who gave a talk on cluster RCTs with balanced incomplete blocks. She was followed by Chris Roberts, professor of biostatistics at Manchester University, who gave a presentation on the Analysis of Trials with Partial Nesting. The final talk of the day was on design and analysis of non-inferiority trials by David Gillespie, a research associate in statistics at Cardiff University's School of Medicine.
A final discussion took place after the talks.