In mdbrown/survNRI: Estimate NRI for survival data

survNRI

This package contains a single function survNRI which calculates the NRI for survival data using up to five different estimators. The different methods available are:

• KM = Kaplan- Meier estimator
• IPW = Inverse probability weighted estimator
• SmoothIPW = Smooth inverse probability weighted estimator
• SEM = Semi-parametric estimator
• Combined= Combined estimator as described (along with all other estimates) in the reference paper.

Tutorial

library(survNRI)
#some simulated data for example. 
data(SimData)

#take a look 
head(SimData)

Calculate the NRI using all estimators at future time 2.

#bootstrap only 10 times to reduce computation for this example. 
survNRI( time  = "stime", event = "status", model1 = "y1", model2 = c("y1", "y2"), data = SimData, 
         predict.time = 2, method = "all", bootMethod = "normal",  bootstraps = 10, alpha = .05)

Now only estimate using the smooth IPW, SEM and combined methods.

tmp <- survNRI( time  = "stime", event = "status", model1 = "y1", model2 = c("y1", "y2"), data = SimData, 
         predict.time = 3, method = c("SmoothIPW", "SEM", "Combined"), bootMethod = "percentile", bootstraps = 10, alpha = .01)

#look at the results
tmp

#access estimates and ci's
tmp$estimates
tmp$CI

for more information see ?survNRI.

References

Lifetime Data Analysis 2012 Dec 20. [Epub ahead of print] Evaluating incremental values from new predictors with net reclassification improvement in survival analysis. Zheng Y, Parast L, Cai T, Brown M. PMID: 23254468

mdbrown/survNRI documentation built on May 22, 2019, 3:24 p.m.