Recommendation for confidence interval and sample size calculation for the Pearl Index

Abstract

A new guideline on the clinical investigation of steroid contraceptives in women, which has been released by the European Agency for the Evaluation of Medicinal Products (EMEA), calls for the calculation of a confidence interval for the Pearl Index, a widely used measure to describe the effectiveness of a contraceptive method. However, the interpretation of the Pearl Index as a statistical parameter, for which a confidence interval can be calculated, needs further clarification. The guideline does not provide the necessary definitions. In this paper, two statistical models, the Bernoulli model and the Poisson model, are compared; both can be used for the calculation of the Pearl Index and its upper confidence limit. The Poisson model proved to be more suitable, because it can accommodate incomplete treatment cycles. Unambiguous definitions and statistical formulae for the calculation of overall Pearl Index and the Method Failure Pearl Index are given. Finally, the sample sizes required to fulfill the EMEA's guideline are given.