icsurvROC: Interval Censored survival ROC

In icsurvROC: Interval Censored Survival ROC

Description

Nonparametric and semiparametric estimations of the time-dependent ROC curve for the interval-censored failure time data.

Usage

  icsurvROC(Time, Status, Marker, pred.time, method, wt=NULL, span=NULL)

Arguments

Time	Monitoring time
Status	Event indicator (1: event occurs before or at monitoring time; 0: event occurs after monitoring time)
Marker	Predictior or marekr value
pred.time	Prediction time of the ROC curve
method	"np" for nonparametric model using local NPMLE; "sp" for semiparametric proportional hazard model
wt	Weight, such as inverse probablity weighting
span	Smoothing bandwidth parameter for method="np"

Details

It estimates a time-dependent ROC curve for the current status data based on the local NPMLE if method="np" or proportional hazards model if method="sp". For method="np", optimal bandwidth selection based on maximum likelihood cross varlidation is used if span is not specified or span=NUL. For method="sp", span is not needed to be specified.

Author(s)

Yunro Chung [cre]

References

Yunro Chung, Tianxi Cai, Yingye Zheng, Estimating Diagnostic Accuracy Measures for Current Status Survival Data with Application to Prostate Cancer Active Surveillance Study (in progress)

Examples

Time=   c(1,2,5,3,9,8,9,4,6,4)
Status= c(1,1,1,0,1,1,1,0,0,0)
Marker= c(8,2,6,3,1,4,5,1,3,7)

#np at year 3
nobs=length(Time)
span=sd(Marker)*nobs^(-1/7)
RES1=icsurvROC(Time, Status, Marker, pred.time=3, method="np", span=span)
print(RES1)

#sp at year 3
RES2=icsurvROC(Time, Status, Marker, pred.time=3, method="sp")
print(RES2)

icsurvROC documentation built on Dec. 3, 2018, 5:04 p.m.