nomo2.crr: Estimate Cumulative Incidence Rate
In jixccf/QHScrnomo: Construct nomograms for competing risk models

Description Usage Arguments Details Value Note Author(s) See Also Examples

Calculate Estimated Cumulative Incidence Rate Calculate predicted cumulative incidence rate based on a competing risks regression model.

1	nomo2.crr(x, f.crr, time)

`x`	a vector of sum of linear predictors for each subject.
`f.crr`	a saved model fitted by function `crr.fit`
`time`	expected evaluation time

This function is usually used to transform regular failure probabilities to competing risks adjusted probabilities, when a nomogram of competing risks regression model is constructed started from a regular survival model. It is not often called externally.

a vector with each element being the predicted cumulative incidence rate at the expected time.

internal function

Michael W. Kattan, Ph.D. and Changhong Yu.
Department of Quantitative Health Sciences, Cleveland Clinic

pred2.crr crr.fit crr

data(prostate.dat)
library(Hmisc,TRUE)
library(rms,TRUE)
dd <- datadist(prostate.dat)
options( datadist = "dd")
f.cph <- cph(formula = Surv(TIME_EVENT, EVENT_DOD == 1 ) ~ rcs(AGE,3) +
    CLIN_STG + rcs(PSA, 3),
    data = prostate.dat, x = TRUE, y = TRUE, surv = TRUE)
# Now construct the crr object based on the above cph object.
# 0s are censored observations.
# 1s are deaths from prostate cancer and
# 2s are deaths from other causes (competing risks).
f.crr <- crr.fit(fit = f.cph, cencode = 0, failcode = 1)

# Estimate cumulative incidence rate by 6 year
pred.lp <- predict.cmprsk(f.crr,lps = TRUE) # calculate X beta values
QHScrnomo:::nomo2.crr(pred.lp, prostate.crr, time = 60)