Letters

Quality of randomised controlled trials in head injury

Trials in head injury are more complex than review suggests

Authors' reply

If in doubt, declare competing interests

Trials in head injury are more complex than review suggests

EDITORThe review by Dickinson and colleagues¹ shows a remarkably narrow view of research in head injury and virtually ignores the need to match the design to the research question. Historically, many clinical trials have been underpowered, but the authors' premise that the main aim of head injury trials should be to detect changes of "a few per cent" in the rate of death or disability does not apply, for example, to phase I/II trials in the acute stage nor the later interventions used in many of their reviewed trials. The authors might find it useful to reread the article "Why do we need some large, simple randomized trials?" by Yusuf et al (note the word "some" in the title).²

Several factors influence the relevant effect size and hence the size of the trial. Some potentially powerful interventions in severe head injury are not widely practicable and are likely to be expensive, and therefore evidence of a substantial effect is required if budget holders are to be persuaded to support them. The focus on a 10% benefit has reflected a perception that funding could be obtained for a treatment that benefits 1 person in 10. However, even this may be optimistic. Despite the 13% benefit obtained from nimodipine treatment in subarachnoid haemorrhage,³ corresponding to a number needed to treat of eight, clinicians have had difficulties in gaining funding for the routine use of this drug. The effectiveness in individual patients is also relevant.

Dickinson and colleagues say that unfamiliarity among ethics committees and investigators with the idea of randomisation without consent obstructs recruitment. This is erroneous and displays a dangerously superficial attitude towards a complex area. What urgently needs to be clarified is the legal framework in which research in incompetent adults takes place. Recent legislation in the Scottish parliament contained no provision for an exception to the requirement to obtain informed consent. Equally it is not clear that any legal framework exists to allow research without consent in the rest of the United Kingdom.

The authors highlighted inadequate funding as one obstacle that has prevented large randomised controlled trials of widely practicable treatments for head injuries. The corresponding author is an applicant to the Medical Research Council for substantial funding for developing the CRASH study⁴ into a full scale trial, a study that is in part supported by the manufacturer of the agent under trial. In view of this, and his apparently strong position on this issue,⁵ it may be found surprising that no competing interests were declared.

Gordon D Murray, professor of medical statistics. 
Department of Community Health Sciences, Epidemiology and Statistics, University of Edinburgh, Edinburgh EH8 9AG Gordon.Murray{at}ed.ac.uk

Graham M Teasdale, professor of neurosurgery. 
University Department of Neurosurgery, Institute of Neurological Sciences, Southern General Hospital, Glasgow G51 4TF

Competing interests: As director of a charitable organisation, the European Brain Injury Consortium, Professor Murray has been active in providing statistical advice to several pharmaceutical companies on the design, conduct, and analysis of clinical trials in head injurynamely, Bayer, Cambridge Neuroscience, Novartis, Pharmos, SmithKline Beecham, and Synthelabo. In addition to extensive declared interests in head injury (BMJ 2000;230:1631-5), Professor Teasdale was a co-applicant to the Medical Research Council and a member of the steering committee for the pilot phase of the CRASH study but is not an applicant for funding for the full phase and has withdrawn from the steering committee.

Authors' reply

EDITORWe are pleased that Murray and Teasdale agree that clinical trials in head injury have been too small and that some large simple randomised controlled trials are needed. To date, there have been no such studies in head injury.

We are grateful to Professors Murray and Teasdale for identifying yet another obstacle to conducting large trials in head injury, the idea that to obtain funding a treatment must benefit at least 1 person in 10. There is no rational basis for the use of such a decision rule. Many factors impact on the decision to provide a treatment but considerations of efficiency require that priority is given to treatments that provide the greatest benefit per unit of cost. Even expensive treatments that benefit fewer than 1 person in 10 might be worth funding if the intervention offers an overall net welfare gain. In head injury, with high rates of long term disability, such a situation might easily occur.

When the effect size is large even small trials may be able to detect it. However, Murray and Teasdale fail to appreciate that both the size and the precision of the estimated treatment effect must be taken into account in therapeutic decision making. Large trials, with larger numbers of outcome events, provide more precise estimates of treatment effect, and the true treatment effect is likely to be close to what has been observed. Imprecise estimates of even large treatment effects from poor quality trials make clinical and funding decisions difficult.

We agree that the legal framework in which research in incompetent adults takes place needs to be clarified. Given that such senior investigators as Murray and Teasdale are unclear on this issue, we hope that we might be forgiven for suggesting that less experienced investigators also find this issue problematic.

In our paper we openly and publicly make the scientific argument for some large simple randomised trials in head injury. We openly and publicly acknowledge that the same scientific argument underpins the Medical Research Council's CRASH trial (corticosteroid randomisation after significant head injury), the first large simple randomised controlled trial in head injury. Open scientific argument in the pages of a medical journal does not constitute a conflict of interest and we are surprised that Murray and Teasdale think otherwise.

Finally, we would point out that the CRASH trial is sponsored by the Medical Research Council and not the manufacturers of the agent under trial. The manufacturers have donated the drug for the trial to the Medical Research Council, but the design, management, and finance of the trial are entirely independent of them.

Ian Roberts, senior lecturer in epidemiology. 
Ian.Roberts{at}ich.ucl.ac.uk

Frances Bunn, review group coordinator. 
Cochrane Injuries Group

Reinhard Wentz, information specialist. 
Cochrane Injuries Group

Phil Edwards, research fellow. 
Child Health Monitoring Unit, Institute of Child Health, University College London, London WC1N 1EH

Competing interests: None declared.

If in doubt, declare competing interests

EDITORFive years ago it was unusual for contributors to medical journals to declare competing interests even though they often had them. Now, increasingly, contributors do declare them, but there continues to be confusion over when to declare.

The BMJ started its campaign on competing interests by asking authors to declare any sort of competing interest, be it personal, political, religious, or whatever. Now we concentrate on financial competing interests because they are easier to define and there is stronger evidence that they matter.

Dr Roberts and others chose not to declare that they had applied to the Medical Research Council for a grant for a large trial of the treatment of head injury. The BMJ's guidance to contributors says: "A competing interest exists when professional judgment concerning a primary interest (such as patients' welfare or the validity of research) may be influenced by a secondary interest (such as financial gain or personal rivalry)." It seems entirely plausible that the view of Dr Roberts and others on the desirability of a large trial of treatment of head injury may be influenced by the Medical Research Council's being more likely to award them a grant if that view becomes widely accepted. In my judgment, they would thus have been wiser to declare their competing interest.

There is nothing wrong with having competing interests, and my advice to contributors is: "If in doubt, declare."

Richard Smith, editor. 
BMJ