Hello all,
I wanted to create a spline curve to have the independent variable on x axis. Just wondering for y-axis should I calculate the HR(x;x_ref) on the y-axis?
like HR(x;x_reference)= exp(xbeta − xbeta_reference)? or no need to consider any reference just calculate the HR on the y-axis?
or is there any resource to learn about the cubic spline in survival?
Thank you,
Thank you, Mark, for your answer.
I was wondering when creating the spline curve, should I plot the hazard ratio (HR) directly on the y-axis, or should I calculate HR relative to a reference value (e.g., HR(x; x_ref) = exp(xbeta − xbeta_ref))?
Which one is more appropriate?
Log(Hazard Ratio) — Most common
Interpretation: The natural logarithm of the hazard ratio (HR) for the covariate modeled by the spline.
Why used: Because hazard ratios are multiplicative, plotting them on a log scale provides a symmetric view around the null value (HR = 1 → log(HR) = 0).
Y-axis label: log(HR) or Log Hazard Ratio
- Badges
- Science topic
- Similar topics
- Computer Science
- Computer Communications (Networks)
- Internet