ABSTRACT
This chapter presents an equation that expresses absolute risk in terms of cause-specific hazards. If the cause-specific hazards depend on covariates through a proportional hazards relationship, it is possible to estimate covariate- and cause-specific hazards by combining relative and attributable risk estimates from case-control or cohort data or subsamples of cohorts, as in the case-cohort design, with estimates of age-specific incidence from a registry. This approach is applicable even if the covariates in equation are time-dependent. The ability to combine information on relative and attributable risk with registry data to estimate absolute risk is one of the major advantages of the cause-specific formulation of absolute risk. Disease registries usually cover a well defined population and collect information on the characteristics of persons diagnosed with a particular disease. These characteristics typically include age at diagnosis, sex, race/ethnicity, and features of the disease, such as cancer stage.
