Create a Cox Proportional Hazards Model with Splines and Covariates

Constructs a Cox model using a survival formula and allows for the inclusion of both regular covariates and spline-transformed continuous variables.

Usage

cox_create_model(
  cox_model_data,
  normal_vars = c("edu"),
  spline_vars = c("age_bs", "bmi"),
  surv_formula = "Surv(tstart, tstop, diagnose) ~ exposure",
  id_var = "ID"
)

Arguments

cox_model_data

A data frame containing the variables required for the model, including survival time variables and covariates.

normal_vars

A character vector of covariate names to be included as standard (non-spline) variables in the model (e.g., `"edu"`).

spline_vars

A character vector of continuous variables for which natural cubic splines will be used (via `splines::bs()`).

surv_formula

A character string specifying the survival outcome in `Surv()` notation. Currently not parsed dynamically; the function uses `Surv(tstart, tstop, diagnose)` internally.

id_var

A string giving the column name for individual-level IDs to be used in the `coxph()` call.

Value

An object of class `coxph` representing the fitted Cox proportional hazards model.

Details

- The function validates variable presence and ensures spline variables are not entirely missing (`NA`). - Splines are applied using `splines::bs()` to the specified `spline_vars`. - The survival object is internally defined as `Surv(tstart, tstop, diagnose)`.

Examples

if (FALSE) { # \dontrun{
model <- create_cox_model(
  cox_model_data = cox_model_data,
  normal_vars = c("edu"),
  spline_vars = c("age_bs", "bmi")
)
summary(model)
} # }