In an era in which all aspects of medicine are being driven by evidence, one consequence has been an increased use of randomized controlled trials. This is not surprising as they representthe most rigorous method of assessment as, by their very nature, they eliminate much of the potential bias that makes other means of comparing 'treatments' less reliable. In order to produce clear evidence, however, any trial must be well constructed and carried out. Achieving this requires careful planning, and in this article we hope to cover the important issues that must be considered in relation to any randomized controlled trial. © 2003 Elsevier Science Ltd.