The safety and immunogenicity of vaccines are demonstrated before they are made available to doctors and the public. However, the protective activity against the disease for which a vaccine was designed can only be accurately measured in field trials. This involves its large-scale use in the target population. Epidemiology is the best approach to assess the results. This approach is based on the comparison of the incidence of the disease among vaccinated and non-vaccinated individuals. Three techniques can be used to estimate vaccine efficacy in a population: screening methods; cohort studies; and case control studies. These methods are most often used during epidemics, and this article discusses their application in this context. The validity of these methods depends on three criteria: that the administration of the vaccine to individuals in the population is random; that contacts between individuals in the population is random, and that vaccinated and non-vaccinated individuals have the same probability of encountering the infectious agent; that the population has equal susceptibility to the infection (other than the effect of the vaccine). If these conditions are not respected, the results of the analyses can be biased. The method involving calculation of the secondary attack rate in the families of index cases is the method where these conditions are most likely to be fulfilled.