Advancing knowledge in stepped-wedge cluster randomised controlled trials: design and analytical issues

Birmingham University Studentship

Advancing knowledge in stepped-wedge cluster randomised controlled trials: design and analytical issues

Lead Supervisor: Dr Karla Hemming

Institute: University of Birmingham, School of Health and Population Sciences, College of Medical & Dental Sciences

Funded by: Birmingham University School studentship

Applications are invited to work with Dr Karla Hemming, Dr Richard Riley and Professor Jon Deeks on an exciting PhD studentship within the biostatistics research team.

Cluster randomised trials (CRTs), in which clusters of individuals are randomised to intervention groups, are being used with increasing frequency to evaluate a range of interventions. Not only are the absolute number of CRTs increasing steadily, they are increasingly being used across different settings, including both health care organisations and non-healthcare organisations, and evaluating a range of interventions some of which target the entire cluster and others of which target the individual. The stepped-wedge design, is a modification to the conventional CRT, in which all clusters initiate as control clusters, with sequential roll-out of the intervention at various points in time, so that by the end of the study all clusters have crossed over to become intervention clusters.

Stepped-wedge designs have many features which appeal to commissioners and providers of healthcare, and have great potential to allow robust evaluation of services (or other interventions), which without recourse to this design, would most likely not be subject to any randomised evaluation. The number of stepped-wedge designs being implemented in practice has increased over recent years and it is highly likely that this trend will continue. Indeed, locally in the School of Health and Population Sciences at the University of Birmingham, at least three studies involving the stepped-wedge design are in progress. However, despite the appeal of these studies and the move towards increased uptake of these designs, coupled with the need to ensure effective (and pragmatic) collaboration between academia and with local providers (i.e. the NHS), there is little research on both the design and analytical issues which are specific to these types of trials.

Project Objectives

1. To undertake an empirical review of the quality of reporting and accuracy of implementation of stepped-wedge trials, based on studies already identified by systematic reviews of stepped wedge studies.
2. Identify issues surrounding concealment of allocation and biased recruitment in stepped wedge trials using a series of case studies.
3. Investigate analytical features of stepped-wedge designs, for example looking at the impact of intra-cluster correlations, imbalanced designs and optimal designs.

Person Specification
We’re looking for a highly motivated person with a passion for statistics research who either holds or realistically expects to obtain at least an Upper Second Class Honours Degree in a relevant subject.

How to apply
Enquiries or applications including a CV, names and addresses of two referees and a covering letter should be sent to, Karla Hemming University of Birmingham, Edgbaston, Birmingham B15 2TT or email [Email Address Removed]

The closing date is 11 July 2012