Department of Health Skip to content

Please note that this website has a UK government access keys system.

You are here:

MS49: Collaborative Stroke Audit and Research (COSTAR)

  • Last modified date:
    8 February 2007

Professor Barer


At the time this project was set up, research on stroke rehabilitation had had little impact on everyday clinical practice. The research effort was fragmented, there was no "Common Clinical Language" for stroke and many studies were of poor methodological quality. The COSTAR collaboration, involving many of the leaders in British stroke rehabilitation research, was set up to overcome some of these difficulties and to prepare for large-scale collaborative studies. Members were asked to rate the

Main aims of the group

In order of importance:

  1. To develop standard (qualitative and quantitative) methods of classifying, measuring and timing rehabilitation interventions ("opening the black box")
  2. To promote collaboration among centres and across disciplines in the scientific evaluation of stroke rehabilitation.
  3. To establish a set of recommended (and not recommended) functional and outcome measures for stroke, and to identify areas where new scales need to be developed.
  4. To plan large-scale, multi-centre clinical trials.
  5. To produce guidelines for the design and conduct of controlled trials in stroke rehabilitation, and a "standard basic protocol".
  6. To stimulate discussion about the organisation and evaluation of stroke services.
  7. To investigate ways of securing better funding and providing incentives for collaborative research.

Collaborative Projects

The present grant allowed the establishment of a COSTAR Co-ordinator to facilitate collaborative projects. A consensus process to identify "burning issues" in stroke rehabilitation and a "Stroke Trials Beauty Contest" were held, to select suitable projects for development into multi-centre studies. Three projects were selected:

  • Evaluation of a multidisciplinary Dysphagia Management Policy for improving the detection and management of swallowing problems in patients with acute stroke
  • Evaluation of Leisure Rehabilitation provided by Occupational Therapists in the community
  • Evaluation of the use of Wheelchairs as a temporary mobility aid in early in-patient rehabilitation

The first two of these projects developed into full-scale studies (CODA study and TOTAL trial) with independent funding, involving 12 different centres in all, and both have now been completed. The Wheelchair project group has completed a randomised controlled pilot trial in two centres and is seeking funding for a larger study. A fourth project (PEGASUS trial of early gastrostomy feeding in stroke patients with dysphagia) was developed shortly after the others, and a randomised controlled pilot trial has been completed in five centres.

Further grants totalling over £750,000 have been obtained through the COSTAR collaboration, and so far over 2,000 stroke patients have been recruited into COSTAR studies.

Common Clinical Language and Common Data Framework

COSTAR has worked together with the European Stroke Database (ESDB) project in developing a "Common Clinical Language" for stroke, by agreeing on standard terminology, classifications, clinical assessments and outcome measures, and a Common Data Framework which can be used in any clinical stroke study.

Common Methodological Standards for Stroke Rehabilitation Trials and Epi-analysis

As well as agreeing on a framework for describing rehabilitation interventions and a basic set of assessment measures, the COSTAR group have agreed on basic methodological standards for trials, including recommended approaches to obtaining consent, randomisation and blinding of outcome assessment.

Complex interventions such as those used in stroke rehabilitation must be evaluated on a collaborative basis, but conventional multi-centre trials leave little room for local "ownership" and can be prohibitively expensive. The common methodological framework developed by the COSTAR group allows pre-planned collaboration between separately funded autonomous trials in answering the same broad questions - "epi-analysis".

Use of the common framework allows research questions to be posed in precise terms, patient categories to be defined and outcomes to be measured in a standardised form, so that results from different studies can be combined in a far more efficient way than has hitherto been possible. Most importantly, epi-analysis involves prospective collaboration, so that hypotheses (including those involving patient subgroups) can be defined in advance, avoiding many of the potential biases of retrospective meta-analysis. Epi-analysis offers a practical alternative to conventional multi-centre trials, which allows small-scale local efforts to be joined into a wide collaboration, making a real contribution to answering important clinical questions.

Barriers to Collaborative Research

The experience of setting up collaborative studies through COSTAR has provided insight into many of the attitudinal, practical and political barriers to multi-centre and inter-disciplinary research, and these are discussed in detail. Some suggestions are made for providing incentives for collaboration rather than competition in clinical research.

Access keys