R/contribution.coxph.R

In rineq: Concentration Index and Decomposition for Health Inequalities

#' @importFrom stats model.matrix weights coefficients
#' @export
contribution.coxph <-
function(object, ranker, correction = TRUE, type = "CI", intercept = "exclude") {
    
    # extract the outcome of the coxph object
    outcome <- object$linear.predictor
    
  
    # extract the weights of the coxph object -> in case of NULL (none provided) generate unit weights
    wt <- weights(object)
    
    if(is.null(wt))
      wt <- rep(1, length(outcome))
    
    # extract the model matrix & coefficients of the coxph object
    # `intercept` has no influence here since coxph models do not have intercepts
    mm <- model.matrix(object)[,names(object$coefficients), drop = F]
    betas <- coefficients(object)

    # call the backend decomposition function
    results <- decomposition(outcome, betas, mm, ranker, wt, correction, citype = type)
    return(results)
}