Latent Events in Clinical Trials

Résumé du livre

[Abstract] A stochastic model of drug versus placebo parallel arms clinical trial with binary outcomes is given in this paper. The model takes into account for therapeutic and adverse effects. These effects are manifest but their dependence on latent effects is made explicit. The model obtained can be used to discriminate between direct and indirect toxic effects of drug administration. Several hierarchically ambedded models are considered so that the choice of the best one can be made by likelihood ratio tests. The model can also be designed to accomodate a dose effect relationship. Statistical model : the SORE model = Sure Outcome of Random Events