Measles in England and Wales--III: Assessing published predictions of the impact of vaccination on incidence.

Published predictions of the impact which different levels of vaccination should have on measles incidence are discussed in the context of observed data on vaccine uptake and measles incidence in England and Wales. Discrepancies are noted between prediction and observation with regard to the effect of vaccination on epidemic periodicity, epidemic size, age distribution of cases, and the disappearance of measles from communities. These errors are attributable to failures to recognize the implications of seasonal trends in incidence, age dependence of risk, and the non-homogeneity of human populations. Predictive models are useful hypotheses, but should be assessed critically against observation.