In October 2006, the World Health Organisation (WHO) convened a meeting of experts to discuss appropriate methods for evaluating the efficacy of malaria vaccines in pivotal phase III trials. The participants included regulatory, industry and donor representatives and clinical trialists, epidemiologists and statisticians from both developed and developing countries. The consultation also considered the regulatory requirements for registration of a malaria vaccine and public health issues that clinical development plans need to address before deployment of a malaria vaccine in developing countries. This report summarizes the discussions and conclusions reached during the course of the meeting. While the global public health burden of malaria is unquestionable there has been considerable variation in the ways in which a case of clinical disease due to malaria has been defined. In designing trials of malaria vaccines it is important that, to the extent possible, definitions of both clinical malaria and severe malaria are agreed that have high specificity and good sensitivity. There was general agreement on how these definitions should be determined, which should facilitate the clinical evaluation of vaccine candidates in paediatric populations in malaria endemic countries. There was agreement that trials of products that might be expected to have lower efficacy than most other vaccines in routine use for other diseases was justified as even partially effective malaria vaccines may be an important tool for reducing the large burden of disease due to malaria globally. Such products would be most easily deployed if they were designed to be administered with other EPI vaccines, which would be appropriate as the greatest burden of malaria is in infancy and early childhood. The conduct of pivotal trials poses special challenges both because the expected efficacy of immediately foreseeable vaccines is likely to be less than 50% and while malaria is a very common disease, distinguishing it from other conditions is far from straightforward. Therefore, in order to facilitate the interpretation of the results from trials, in particular for regulatory decision-making, it is essential that, insofar as is possible, methods that are used to define the clinical endpoints in such trials are standardised and validated. Cogent cases can be made for using either uncomplicated malaria disease or severe disease as the primary endpoint in pivotal trials, as both impose an enormous public health burden. The decision on which of these is most appropriate will be influenced by both scientific and non-scientific factors. Public health authorities might be more likely to accelerate introduction of a vaccine if an effect on severe disease had been demonstrated in a pivotal trial. Such decisions would also be influenced by knowledge of the efficacy of the vaccine in different malaria endemic settings and by knowledge of the duration of protection conferred post-vaccination. While phase IV studies may be necessary to generate some of this information, it is important to design pivotal trials to provide this information to the extent possible.