On the utility of statistical modelling in reproductive studies

The usefulness of mathematical modelling is discussed, with reference to a beta-binomial model which is used to describe both the implantation and birth statistics after assisted procreation. The relatively heavy computing effort required is emphasized, and contrasted with the rather simpler calculations associated with traditional statistical methods. The advantages of deriving a concise description of the data with a few important, and easily interpretable, parameters are also discussed. Numerical data provided by an IVF clinic are used to illustrate the mathematical procedures and to provide an assessment of the success of the modelling exercise.