Survival Data

Summary

Develops survival analysis as its own world built on censored time-to-event outcomes, starting with right censoring through loss to follow-up, competing causes, and administrative end-of-study mechanisms. Builds the Kaplan-Meier product-limit estimator as a staircase function dropping at each event time, the log-rank test for comparing curves, median survival, and the assumption of independent censoring. Develops Cox proportional hazards regression with its semiparametric baseline hazard and partial likelihood, then covers diagnostics including Schoenfeld residuals, time-dependent covariates, and competing risks through cause-specific hazards and the Fine-Gray subdistribution approach.