Fig. 1
From: Examining evidence of time-dependent treatment effects: an illustration using regression methods
Graphical summary of the regression-based modelling approaches when non-proportional treatment effects are present. Estimated hazards (y-axes) and treatment effects from the Cox PH, the Weibull, the FPM PH and FPM TD models when non-proportionality of the true hazards (dashed lines) were present for a hypothetical situation. The arrows indicate the magnitude and direction of treatment effect as measured from the modelled baseline hazard (solid light blue line) to the modelled treatment line (solid purple line)