Risk research and risk analysis have to be modeled as a fairly complex system including multivariate regression modeling for risk factors in etiology, Markov models in pathogenesis, and a construct of mechanistic and hermeneutic variables for clinical outcome analysis. The McPeek index is proposed as an example. Several prophylaxes for risk reduction in the perioperative period produce risk reduction as well as risk augmentation in different types of outcome. These unexpected findings were observed not only in clinical trials, but also in animal experiments and in isolated tissues. This demonstrates a basic problem of handling complexity in the real clinical setting.