Analyze time-to-event outcomes

After cardiac surgery, when did the event happen and what changed along the way? These recipes begin with the observed follow-up and then move from familiar survival displays to an additive hazard model when the shape of risk needs more than a Cox model.

Choose the next recipe

Your question	Recipe	Why this one
How does freedom from an event change over follow-up?	Survival plots	Draws the Kaplan-Meier view with the information readers expect.
When is risk highest or lowest?	Hazard and nonparametric curves	Shows instantaneous risk rather than cumulative survival.
How large is the clinical difference between two survival curves?	Number needed to treat and survival difference	Turns a survival contrast into a count a clinician can use.
Which phases belong in an additive hazard model?	Fit additive hazard models	Starts with the empirical curve before selecting early, constant, or late phases.
What does the fitted model predict for this patient profile?	Predict from a hazard model	Produces survival, cumulative-hazard, instantaneous-hazard, and phase-specific curves.
Does the hazard model agree with the observed data?	Check a hazard model	Compares fit, calibration, selection uncertainty, and competing outcomes.
How do I move a PROC HAZARD analysis into maintained R code?	Move a legacy HAZARD analysis into R	Carries the model structure into an R workflow that we can keep.