Parametric Survival Model

survival-analysisproportional-hazardsweibullexponential-distributionextreme-valuecensoringpoisson-regressiongeneralized-linear-model

Definition

A parametric survival model specifies the distribution of a time-to-event TT through a fully parametric hazard (or, equivalently, survival) function and estimates covariate effects by maximum likelihood, accounting for censoring. In the proportional-hazards form the hazard is h(t∣x)=λ(t)exp⁡(β′x)h(t\mid x)=\lambda(t)\exp(\beta'x) with a parametric baseline λ(t)\lambda(t) (exponential, Weibull, or extreme-value), giving survival S(t∣x)=exp⁡ ⁣(−Λ(t)eβ′x)S(t\mid x)=\exp\!\big(-\Lambda(t)e^{\beta'x}\big), Λ(t)=∫0tλ(u) du\Lambda(t)=\int_0^t\lambda(u)\,du. Aitkin-Clayton (1980) showed such models can be fit with ordinary Poisson generalized-linear-model software.

Key Ideas

Why It Matters

Open Questions

Related