Spring term seminars 2011 Ives Ntambwe (WMS) Sample size re-estimation in clinical trials with multiple endpoints The choice of an appropriate sample size is a main concern in the design of any clinical trial. In the planning stage of a trial one is often quite uncertain about the sizes of parameters or assumptions needed for sample size calculations. The idea of this project is to explore the use of designs that allow checking of these assumptions and adjustment of the sample size if necessary. Tuesday 25 January 12:30 13:30 (Room B026) Wittes and Brittain (1990) proposed use of designs with sample size re-estimation, also called designs with internal pilot study (IPS) which look at assumption regarding the nuisance parameters at an interim analysis. Multiple endpoints are not uncommon in clinical research. One example is the use of a test battery in schizophrenia. The analysis is complicated in the presence of multiple endpoints and special techniques are needed to control the Type I error rate because hypotheses are tested for various endpoints. This project aims to bring together the concept of designs with sample size re-estimation and the methodology for dealing with multiple endpoints. This will provide an extension of the current methodology for single endpoint sample size re-estimation to the multiple outcomes setting. Chris Roberts (University of Manchester) The Analysis of Non-Pharmacological Treatment Trials with Partial Nesting and Multiple Therapists per Patient Tuesday 15 February 12:30 13:30 (Room A011) In trials of physical and talking therapies nesting of patients within care-provider can have design and analysis implications analogous to cluster randomized trials. There can nevertheless be additional complexity due to heterogeneity in the pattern of clustering including partial nesting. In some trials, a patient may receives a therapy from multiple care-providers, so that the data structure is no longer strictly hierarchical. This talk will consider methods of analysis of trials with Spring term seminars 2011 partial nesting and multiple-membership due to therapist or therapy group. Methods and issues will be illustrated by data from trials of therapies for musculoskeletal and psychological disorders. Sample size and power for trials with multiple membership due to therapist will also be considered. Nick Parsons (WMS) Tuesday 22 March 12:30 13:30 (Room A011) Can Weather Forecasts be used to Predict Admission Rates to UK Trauma Units?