pssm.survivalcurv: time to progression and time to death function for a "pssm"...

In pssm: Piecewise Exponential Model for Time to Progression and Time from Progression to Death

Description

Generates a function of time for specified covariate values from a "pssm" object created by pssm that fits a joint proportional hazards and survival model using a piecewise exponential underlying hazard function

Usage

pssm.survivalcurv(x, cov1, cov2, timeToProgression = FALSE, covariance = TRUE)

Arguments

x	A pssm object created by pssm
cov1	a a*b matrix of values of the b covariates affecting the time to progession
cov2	a matrix of values of the covariates affecting survival after progression with the same number of rows as cov1
timeToProgression	if FALSE estimates the survival curve, if TRUE estimates two probabilities, the probability of being disease free before t and the probability of progressing before t but surviving after t
covariance	if TRUE the covariance matrix is returned as an attribute of the function value

Details

pssm.survivalcurv returns a function the argument of which is the vector of times for which survival probabilities are desired.

Value

A function is returned, the input to the function is a vector of times, and an optional parameter indicating the prior precision on the estimate of the -log hazard ratio of the effect of survival after progression on the last covariate in the survival model (presumed to be treatment) and the output is a data frame with columns described below:

Note that to conduct the bayesian analysis the Covariance needs to be set to T.

rep	indicates what is estimated (see below), values are "s1" or "s2"
time	Time, t
covariates	Columns indicating covariates for survival and progression
estimate	Estimate, If timeToProgession is TRUE and the estimation was done with both survival and time to progession the "s1" value is the probability that a patient will progress before time t but survive longer than t. In that case the value at "s2" is the probability a patient will be disease free before t. Otherwise rep will only equal "s1" and it will be the probability that survival or progression occurs latter than t as the case may be.

Author(s)

David A. Schoenfeld

See Also

pssm-class, pssm-package, pssm.generate.data, pssm.object, pssm, pssm.simulate,

plot-methods, pssm.power

Examples

#generate data for plot
u<-pssm.generate.data(theta1=.5,theta2=0,phaz.progression=rep(log(-log(.3)/4),5),	
    phaz.survival=rep(log(-log(.15)/4),15),accrual=2,followup=2.9,m=5,
    n=50,times=c(1,2,3),delta=0.5)
#estimate parameters
ps<-pssm(surv(tprog0,tprog1)~rx,surv(tdeath,cdeath)~rx,dat=u,intervals=3)
#calculate survival curves
vs<-pssm.survivalcurv(ps,cov1=matrix(c(0,1),2,1),cov2=matrix(c(0,1),2,1),covariance=TRUE)
t=c(0,2,4,4.99)
curves=vs(t)
#plot survival curves
plot(t,curves$estimate[curves$rx==0],type='l',lty=2,ylim=c(0,1),
     main='Survival Curve',xlab='Time',ylab='Probability of Survival')
points(t,curves$estimate[curves$rx==1],type='l',lty=1,xlim=c(0,5))

pssm documentation built on May 2, 2019, 11:12 a.m.