Estimation of the population effectiveness of vaccination

This paper presents a simple method for estimation of population vaccination effectiveness, which is the fraction of disease cases prevented by a vaccination programme. The method is based on the susceptible-infectious-recovered (SIR) model for the spread of an epidemic in a heterogeneous population under non-homogeneous mixing. The required data are the attack rates among vaccinated and unvaccinated individuals in each stratum of the population. Some information on the mixing pattern may be useful, but in most cases is not absolutely necessary. One can extend the estimation method to predict the attack rates expected in the same population for different vaccination strategies. I apply the new methods to data from a measles outbreak and investigate the bias, standard error and robustness of the estimation method in a stochastic simulation study.