Conditions of validation and use of the screening method for vaccine efficacy evaluation.

ABSTRACT

If PV is the proportion of the target population vaccinated and PCV the proportion of cases in the same population who have been previously vaccinated, vaccine efficacy can be computed as (PV-PCV)/[PV(1-PCV)]. In the screening method, this formula is currently applied to sample values. The objective of this paper is to provide the corresponding statistical tools. Taylor series expansion has been used to derive the bias, variance and confidence interval of the estimate. The bias may be very high when vaccine efficacies are low. It is below 1% for vaccine efficacies higher than 80%. Monte Carlo simulations were used to assess the departure from normality of the estimate. The number of subjects required by the method for given precisions are presented.