NonCumHaz: NonCumHaz

In DySeq: Functions for Dyadic Sequence Analyses

Description

Computes the non-cumulated hazard from a vector containing the cumulated hazard. Can be applied directly to survfit-object (no need to extract the hazard first!). If the vector contains only hazard information for some but not all time intervalls, e.g. intervals with a hazard of zero are left out, a second vector is needed to match the hazard to the corresonding time intervals.

Usage

NonCumHaz(x, t = NA, plot = FALSE)

Arguments

x	a vector containing cumulated hazard or a survfit-object from the 'survival'-package
t	optional: vector containting time reference for x (is required for plot)
plot	logical value indicating if non-cumulated plot should be generated

Examples

# Example 1: Short artificial data
# example cumulated hazard with time referenz
cumhaz<-c(0.2 ,0.21 ,0.31 ,0.44 ,1.1  ,1.1  ,1.12 ,1.2)
time<-c(4     ,5    ,6    ,7    ,10   ,14   ,15   ,16)

NonCumHaz(cumhaz, time, plot=TRUE)



# Example 2: Every hazard entry represents one point of time

# if every hazard entry repesents one point of time
NonCumHaz(cumhaz, 1:8, plot=TRUE)



# Example 3: real data and real researchquestion
## Not run: #install.packages("survival") 
library(survival)

# How long till the last stress signal
my.last<-LastOccur(CouplesCope[,50:97],1)

# If last stress signal was in time intervall 48,
# stress did not end till the observation duration
event<-rep(1,length(my.last))
event[my.last==48]<-0

# Coxregression
my.surv<-Surv(my.last,event) # creates a object for survival time analysis
my.fit<-survfit(coxph(my.surv~1)) # fits coxregression without covariates on the data

plot(my.fit) # survival curve
plot(my.fit, fun="cumhaz") # cumulated survival curve

# Different uses for NonCumHaz
NonCumHaz(my.fit, plot=TRUE)
NonCumHaz(my.fit$cumhaz, my.fit$time, plot=TRUE) # if over packages than 'survival'are used

## End(Not run)

DySeq documentation built on May 1, 2019, 10:14 p.m.