byGroupLOS: Length of stay by groups

In n8thangreen/HESmanip: Hospital length of stay data preprocessing and analysis

Description

Batch calculate all the cumulative hazards, transition probabilites and length of stay for each of the organism groups in turn

Usage

byGroupLOS(survData.list, type, standerr = FALSE)

Arguments

survData.list
type	Model type: cause-specific "", subdistribution with discharge alive as event of interest "alive" or subdistribution with in-hospital death as the event of interest "dead".
standerr	Should the standard error be calculated. This may take some minutes.

Value

Returns list of lists. Below initial sample stratification:

mvna.data	Output from mvna contained in mvna
clos.data	Excess length of stay output
se	Bootstrap standard error estimate
etm.data	Output from etm contained in etm

See Also

byGroupLOS.boxplot

Examples

require(mvna)
require(etm)
data(survDataByGender)

foo <- byGroupLOS(survDataByGender,"")
foo$Male$clos.data



## The function is currently defined as
function (survData.list, type, standerr = FALSE) 
{
    startTime <- Sys.time()
    namesGroup <- names(survData.list)
    progressbar <- txtProgressBar(min = 0, max = length(namesGroup), 
        style = 3)
    counter = 0
    tra <- matrix(FALSE, 3, 3, dimnames = list(as.character(0:2), 
        as.character(0:2)))
    tra[1, 2:3] <- TRUE
    tra[2, 3] <- TRUE
    results <- list()
    for (i in namesGroup) {
        counter = counter + 1
        setTxtProgressBar(progressbar, counter)
        if (i == "") {
            stop("no group name")
        }
        if (nrow(survData.list[[i]]) == 0) {
            results[[i]] <- list(NA, NA, NA)
        }
        else {
            results[[i]] <- listLOS(survData.list[[i]], tra, 
                type, standerr)
        }
    }
    close(progressbar)
    cat(Sys.time() - startTime)
    results
  }

n8thangreen/HESmanip documentation built on March 21, 2020, 12:20 a.m.