R/CM_ECMR_C_IMF.R

In surveyPrev: Mapping the Prevalence of Binary Indicators using Survey Data in Small Areas

##' CM_ECMR_C_IMF imr 5 years prior to survey.
##'
##' BR
#' @param Rdata  data.frame from surveyPrev::getDHSdata
#' @param mort.cut  age group cutoffs for hazard calculation. Constant hazards are assumed to be within the specified cutoffs.
#' @param mort.year The number of years preceding survey. Set to be 5.
#'
#' @return A partially processed data.frame that will be used in surveyPrev::getDHSindicator. The whole function can be used as a parameter in surveyPrev::getDHSindicator
#'
#' @examples
#' \dontrun{
#' dhsData <- getDHSdata(country = "Zambia", indicator = "CM_ECMR_C_IMF", year = 2018)
#' }
#'
#' @export
#'
CM_ECMR_C_IMF<- function(Rdata, mort.cut = c(1, 2, 6, 12, 24, 36, 48, 60), mort.year = 5){


  Rdata$b5 <- labelled::to_factor(Rdata$b5)
  Rdata <- subset(Rdata, v008-12*(mort.year + 5)-b3 < 0)
  Rdata$strata <- NA
  # Get birth coded as person-months
  # year.cut is specified for a wide range to avoid SUMMER's rule of dropping partial year observations for now
  raw.dat.tmp <- SUMMER::getBirths(data = Rdata,  strata = c("strata"),
                                   year.cut = c(1900, 2100), month.cut = mort.cut)
  # now we need to remove person months not in the [-X, 0] range (between -X-5 and -X)
  # id.new is the row number in the original input
  Rdata$id.new <- 1:dim(Rdata)[1]
  raw.dat.tmp1 <- left_join(raw.dat.tmp, Rdata[, c("id.new", "v008")])
  raw.dat.tmp1 <- subset(raw.dat.tmp1, v008 - mort.year*12 < obsStart)
  raw.dat.tmp <- raw.dat.tmp1[, -which(colnames(raw.dat.tmp1) == "v008")]
  raw.dat.tmp$value <- raw.dat.tmp$died

  raw.dat.tmp <- raw.dat.tmp %>%
    subset(age %in% c("0", "1-1", "2-5", "6-11"))%>%
    droplevels()


  return(raw.dat.tmp)


}