Hi!
Let's say I have a Cox Model, for example:
cox <- coxph(Surv(Last_FU, as.numeric(Status)) ~Histology + Metastisis + Recurrence + Adjuvant.Treatment + Transcriptomic_Classification, data = clinical_data)
From this i can obtain a Forest plot with the p-values and hazard ratios, after confirming that all variables fullfill the model's assumptions.
Then I passed this cox to an anova over CoxPH as suggested in this tutorial https://thomaselove.github.io/2020-432-book/cox-regression-models-for-survival-data-example-2.html , using this package https://rdrr.io/cran/survival/man/anova.coxph.html .
anova(cox)
The results obtained show Transcriptomic_Classification variable as the only variable with P<0.01.
My intuition tells me to conclude from this that Transcriptomic Classification emerged as the most significant variable at predicting patient's survival, however I have not seen any scientific paper with anova(cox).
Therefore I would like to ask the community, what is this anova(cox) doing in detail? And if it could be used to draw such conclusions?
Best Regards and thank you.
