Skip to contents

Time-to-event data example 1 for non-proportional hazards working group

Source: R/Ex1delayedEffect.R
Ex1delayedEffect.Rd

Survival objects reverse-engineered datasets from published Kaplan-Meier curves. Individual trials are de-identified since the data are only approximations of the actual data. Data are intended to evaluate methods and designs for trials where non-proportional hazards may be anticipated for outcome data.

Usage

data(Ex1delayedEffect)

Format

Data frame with 4 variables:

  • id: Sequential numbering of unique identifiers.

  • month: Time-to-event.

  • event: 1 for event, 0 for censored.

  • trt: 1 for experimental, 0 for control.

References

Lin, Ray S., Ji Lin, Satrajit Roychoudhury, Keaven M. Anderson, Tianle Hu, Bo Huang, Larry F Leon, Jason J.Z. Liao, Rong Liu, Xiaodong Luo, Pralay Mukhopadhyay, Rui Qin, Kay Tatsuoka, Xuejing Wang, Yang Wang, Jian Zhu, Tai-Tsang Chen, Renee Iacona & Cross-Pharma Non-proportional Hazards Working Group. 2020. Alternative analysis methods for time to event endpoints under nonproportional hazards: A comparative analysis. Statistics in Biopharmaceutical Research 12(2): 187--198.

See also

Ex2delayedEffect, Ex3curewithph, Ex4belly, Ex5widening, Ex6crossing

Examples

library(survival)

data(Ex1delayedEffect)
km1 <- with(Ex1delayedEffect, survfit(Surv(month, evntd) ~ trt))
km1
#> Call: survfit(formula = Surv(month, evntd) ~ trt)
#> 
#>         n events median 0.95LCL 0.95UCL
#> trt=0 121     86   5.04    4.18    6.21
#> trt=1 240    132   7.66    6.54    9.48
plot(km1)

with(subset(Ex1delayedEffect, trt == 1), survfit(Surv(month, evntd) ~ trt))
#> Call: survfit(formula = Surv(month, evntd) ~ trt)
#> 
#>        n events median 0.95LCL 0.95UCL
#> [1,] 240    132   7.66    6.54    9.48
with(subset(Ex1delayedEffect, trt == 0), survfit(Surv(month, evntd) ~ trt))
#> Call: survfit(formula = Surv(month, evntd) ~ trt)
#> 
#>        n events median 0.95LCL 0.95UCL
#> [1,] 121     86   5.04    4.18    6.21