ABSTRACT
This chapter focuses on topics related to absolute risk, including prognosis following disease diagnosis, handling missing data on cause of death, and time-varying covariates or health state. It also discusses the applications of absolute risk for individual counseling and for public health prevention strategies. The chapter indicates how the ideas in previous chapters are related to other applications and analytical approaches and provides some key references. Absolute risk estimates often rely on survival analysis methods that were first developed for estimating pure risks following disease diagnosis. Perhaps the first example of covariate modeling in survival analysis concerned the pure risk of death following diagnosis of acute myelogenous leukemia as a function of baseline white cell count. Multistate models require that the time-varying covariate be discrete. Although it is possible to categorize continuous markers and put them into the multistate framework, this approach may result in loss of information.
