For evaluating measles vaccine efficacy (VE) in the field, some investigators have suggested that an overall attack rate level of 5% or more in a randomly mixing population is sufficient to provide equal exposure to the viral agent in both vaccinated and unvaccinated groups.
It is not clear, however, if this assumption is valid given the effect of herd immunity.
We created a computer simulation model based on the stochastic extension of the Reed-Frost model and tested for variation in bias in VE estimation due to herd immunity, based on runs of 200 trials.
At higher levels of attack rate, the degree of herd immunity decreases, as does the percentage of trials with bias in VE estimation.
The two main factors that affect the level of attack rate are the probability of adequate contact and the number of susceptibles.
At a given level of attack rate, the number of susceptibles is positively associated with the percentage of biased trials in VE estimation.
Since vaccination reduces the number of susceptibles, we also observe that when controlling for attack rate, higher vaccination coverage results in lower bias in VE estimation.
The results show that the assumption of no bias when the attack rate is 5% or more becomes increasingly true when a large percentage of a randomly mixing population is immune.
Mots-clés Pascal : Rougeole, Virose, Infection, Vaccination, Prévention, Efficacité, Homme, Modèle mathématique, Simulation ordinateur, Méthodologie
Mots-clés Pascal anglais : Measles, Viral disease, Infection, Vaccination, Prevention, Efficiency, Human, Mathematical model, Computer simulation, Methodology
Notice produite par :
Inist-CNRS - Institut de l'Information Scientifique et Technique
Cote : 95-0474713
Code Inist : 002B05C02C. Création : 01/03/1996.