Randomised clinical trials provide the most valid means of establishing the efficacy of clinical therapeutics. Ethical standards dictate that patients and clinicians should not consent to randomisation unless there is uncertainty about whether any of the treatment options is superior to the others ("equipoise"). However, true equipoise is rarely present; most randomised trials, therefore, present challenging ethical dilemmas. Minimising the tension between science and ethics is an obligation of investigators and clinicians. This article briefly reviews several techniques for addressing this (...) issue and suggests that unbalanced randomisation, a technique rarely employed in current clinical trial practice, may be useful for enhancing the ethical design of human experimentation. (shrink)