Red Door Analytics When the hazard ratio misleads

Statistical tools

When a single hazard ratio misleads

A Cox model reports one hazard ratio. But if the treatment effect is delayed, wanes, or the curves cross, that single number hides the story. Pick a non-proportional scenario and see what a Cox model would report — against what actually happens to survival, RMST and the hazard ratio over time.

Effect over time

Effect parameters
Control survival
RMST horizon (τ)

What you'd report vs what's true

Red Door Analytics

Designing and analysing trials beyond proportional hazards

RMST, milestone survival, weighted log-rank, and flexible-parametric models with time-dependent effects — choosing and pre-specifying the right estimand is what merlin and our trial-design work are for.

Talk to us

See also: Survival DGM explorer· Multi-state explorer· Sample-size & power calculator· how we pre-specify an analysis →

The reported hazard ratio is a Cox model fitted to a simulated trial of the chosen mechanism; HR(t), RMST and survival are computed from the model. For teaching and estimand illustration.