Important issues in planning and conducting multi-centre randomised trials in cancer and publishing their results

The principles and practical issues involved in planning and conducting multi-centre randomised trials in cancer and in publishing their results are reviewed for the benefit of those conducting trials or planning to do so. There is increasing public awareness of the importance of randomised trials and this needs to be nurtured. Questions addressed by trials must be clinically worthwhile, and trials must be planned and conducted to provide reliable results that will convince the clinical community and influence future clinical practice to patients' benefit. Trial protocols must justify the trial and give clear and unambiguous instructions to collaborating centres. Only data necessary for answering the trial questions should be collected, and the reliability of data should be checked. All trials should be published in peer-reviewed journals, and ways should be considered of communicating their results to patients, their families and the public. Wide consultation and independent review are needed to achieve these aims.