Adding global P value to logistic regression and cox proportional hazards model - Which tests are used? is it the Wald test?

Question

I have been asked to add a global p-value to my categorical variables in my multivariable logistic regression model and in my multivariable cox proportional hazard model.

I have used the function add_global_p() when using the function tbl_regression (which summarizes my results into a table) but I cannot figure out which tests have been used to calculate the global p-value. Is it the wald test? Do I need to specify the test or does it automatically select the test according to the model you have used?

Answer 1

If you use coxph() function from survival package to fit your cox proportional hazard model, you can use summary() call to get the global p-value. There are three p-values calculated using three different tests (likelihood ratio test, Wald test and score logrank test) offered on the bottom of the output.

library(survival)

model_fit <- coxph(Surv(time, status) ~ covariate_1 + ... + covariate_n, data = your_data)

summary(model_fit)

Answer 2

According to Add the global p-values, the function uses car::Anova().