ABSTRACT
This chapter aims to outline the mathematical and epidemiological basis of the notion of risk, and to discuss in detail a number of general aspects of the construction and application of systems of risk assessment. It seeks to highlight the typical methodological problems inherent in risk assessment and outlines risk scoring procedures in the field of obstetrics by the use of examples. The tests of statistical significance carried out as part of the application of discriminant analysis, however, necessitate the assumption of multivariate normality; as a result of the categorical or ordinal nature of most risk factors in obstetric and pediatric applications, this assumption is generally not realistic. Separate analyses for subgroups such as the multigravidae may be necessary since the inclusion of obstetric history factors in a system for all pregnancies can lead to inappropriate assumptions concerning missing data.
