akm-rmst's Introduction

IPW-adjusted Kaplan-Meier restricted mean survival times

This project contains an R function for inverse probability weighted (IPW) adjusted restricted mean survival times in observational studies. We also provide a working example using the dataset 'lung' from the {survival} package.

We acknowledge the source code of F. Le Borgne and Y. Foucher, authors of the 'adjusted.KM' function from {IPWsurvival} package, and Hajime Uno, Lu Tian, Angel Cronin, Chakib Battioui, and Miki Horiguchi, authors of the 'rmst1 function' from {survRM2} package.

akm_RMST.R

This function will produce IPW-adjusted RMSTs, differences and ratios in RMSTs for pairwise 
comparisons, and adjusted Kaplan-Meier survival curves. 
The function requires the following arguments:

Time: time to event
Status: 0 if censored, 1 if event
Group: factor variable for the exposure of interest
Weights: to be obtained previously, ie through logistic models
Tau: a user-specified truncation point. 
     If not specified, the default will be the minimum of the each groups' last event time 
Xaxismin: an optional minimum value for the x axis of the Kaplan-Meier plot.
     If not specified, the default is 0.
Xaxismax: an optional maximum value for the x axis of the Kaplan-Meier plot.
     If not specified, the default is the maximum survival time.

lung.R

We show how to apply our function to the lung dataset, available in the {survival} package. 
We examine the mean survival time (days) between high and low Karnofsky performance score 
rated by a physician, adjusted for sex, age, calories consumed, and ECOG performance score. 
We obtain inverse probability weights using a logistic model, and apply these weights in our 
function to obtain IPW-adjusted Kaplan-Meier curves and RMSTs.

Karnofsky score is dichotomized according to clinical criteria:
http://www.npcrc.org/files/news/karnofsky_performance_scale.pdf

Recommend Projects