R/fludiff.R

In kmcconeghy/flumodelr: An R Package for Estimating Attributable Influenza Morbidity and Mortality

#' @title fludiff: Compute excess mortality
#' 
#' @description This function takes fitted data from a serflm() or similar 
#' procedure and computes difference between observed and predicted mortality.    
#' 
#' @param data A dataframe class object, must contain observed, fitted variables
#' 
#' @param obsvar named observed outcome variable  
#' 
#' @param fitvar named fitted trend variable. Can specify either fit line, or 
#' upper threshold limit.  
#' 
#' @param serfrule Logical, default=F. If T the function will apply the 
#' Serfling rule which only counts consecutive observations of excess 
#' mortality. 
#' 
#' @return an object of class dataframe  
#' 
#' @export
#' 
#' @examples
#' require(flumodelr)
#' df <- flumodelr::fludta
#' serf_fit <- fluserf(fludta, outc =fludeaths, time =yrweek_dt)  
#' excess_fludeaths <- fludiff(serf_fit, obsvar=fludeaths, fitvar=y0)
#' 
#' @import rlang 

fludiff <- function(data=NULL, obsvar=NULL, fitvar=NULL, serfrule=F) {
  
  #tidy evaluation  
  obs_eq <- enquo(obsvar)
  fit_eq <- enquo(fitvar)
  
  if (serfrule==F) {
    data <- mutate(data, y_diff = if_else(((!!obs_eq) - (!!fit_eq))>0,
                                            (!!obs_eq) - (!!fit_eq), 0)
    )
  } else if (serfrule==T) {
    
  }
  
  #return dataframe
  return(data)
}