This book is a practical guide to the quantitative estimation and stratification of medical risk. It is not about legal risk, or about the business risk that is associated with capitated insurance plans. Although some of the concepts described in the pages to follow could provide useful information for evaluating aspects of the risk pertinent to other professional domains, we will not discuss such applications in this book. Medical risk, as we define it here, is the probability that a specific kind of event or outcome will occur following or in connection with a medical intervention. The intervention and the outcome can be anything, as long as they can be measured with reasonable accuracy, so that a relationship between them can be evaluated.

Risk stratification is a branch of clinical research, but it differs from classic efficacy-oriented medical research in an important way. Rather than evaluating the efficacy of a treatment, it determines whether events that occur in a local population are accounted for by the risk factors in that population. Said another way, risk stratification determines whether events, such as deaths following surgery, infections in the hospital, etc., that occur in a particular population can be explained by risk factors that are known to produce such events. If more events occur in a population than would be expected based on risk factors, some other source of the excess event rate is inferred to be present in the population. In a formal risk-stratification study, a local population is standardized by comparison to the well-known risk factors of an established model.