Rosenberg and coworkers caused an uproar in the immunotherapy community when they stated in a position paper that despite great advances in the field of tumour immunology, optimism about the clinical application of therapeutic vaccines lacked justification. They argued that such optimism was based more on surrogate end-points, which lacked robustness, than on proof of anti-tumour effects [Rosenberg SA, Yang JC, Restifo NP. Cancer immunotherapy: moving beyond current vaccines. Nat Med 2004; 10:909-15]. They pointed out that cancer vaccine trials in 440 patients, conducted at the NCI Surgery Branch, had an overall objective response rate of only 2.6%. This was comparable to the 4.0% response rate reported in 40 studies that involved 756 patients. They concluded that surrogate end-points with such low response rates lack robustness. The situation may be even more problematic. Results of a number of larger trials in the past few years have indicated that vaccination therapy can also have a detrimental effect and be associated with worse outcome. These findings need to be looked at seriously and should lead to a critical appraisal of how well we understand these outcomes, how vaccine trials should be designed, monitored and conducted and which opportunities should be considered to be implemented in vaccine trial designs to improve the rationale and chances for a positive outcome.