A statistical evaluation of implantation after replacing one or more human embryos

All the available data relating to pregnancy and multiple implantations from Bourn Hall Clinic are used to assess the accuracy of two statistical models of implantation. The simple binomial model was found to be incapable of describing both the pregnancy rates and the incidence of multiple implantations, whereas a two-parameter model, incorporating both patient and embryo variability, provided a good fit to the data. The estimates obtained for the two parameters in the second model suggested that 36% of the embryos could implant, and that 43% of patients were capable of sustaining implantation. The Maximum Likelihood method of constructing 95% confidence limits on the estimates of the parameter values is demonstrated.