The decision to treat a patient is often based on the results of randomised controlled trials (RCTs). These important investigations inform physicians and patients with regard to the chance of favourable results of treatment and the risks of adverse reactions. However, research has shown that especially the smaller RCTs in which no or even an adverse effect of new interventions is found run a relatively high risk of remaining unpublished, which leads to publication bias and an overestimate of the efficacy of an intervention. This is undesirable. The best way to identify possible publication bias and reduce its negative effects is to register RCTs from their starting date onwards in a prospective and publicly accessible register. Other motives for such a register are the prevention of duplication of research work and its financial support, and informing patients. Currently, a prospective Dutch national trial register is being developed by the Dutch Cochrane Centre. This is not only of importance to researchers and patients, but also to grant-providing bodies and the editors of medical scientific journals.