Conference Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 277 - SPEED: Biometrics and Environmental Statistics Part 1
Type: Contributed
Date/Time: Tuesday, August 9, 2022 : 10:30 AM to 12:20 PM
Sponsor: Biometrics Section
Abstract #322499
Title: R Package AdjKM.CIF and Shiny Application for Creating the Covariate-Adjusted Kaplan-Meier and Cumulative Incidence Functions
Author(s): Biwei Cao* and Jonphil Kim
Companies: H. Lee Moffitt Cancer Center & Research Institute and H. Lee Moffitt Cancer Center & Research Institute,University of South Florida
Keywords: AdjKM.CIF; Bootstrap method; Competing Risks; Confidence Interval; Cox proportional hazards regression model; Fine-Gray subdistribution hazard regression model

In retrospective studies and prospective observational studies, the groups being compared may be imbalanced with regards to the prognostic factors that are significantly associated with the time-to-event outcomes, and thus the unadjusted Kaplan-Meier (KM) function and the cumulative incidence function (CIF) may be incompatible with the results from the multivariable model. To our knowledge, no R package is available for creating the stratified model-based, covariate-adjusted KM or CIF functions. To address these concerns, the covariate-adjusted KM functions and CIFs can be estimated by the multivariable Cox and Fine-Gray regression model, depending on the presence of the competing risks. The Direct Adjustment method is used to create the covariate-adjusted KM functions and CIFs. The Gail and Byar method and the Storer method are offered for the application of the stratified Cox model and the stratified Fine-Gray regression model. In addition, the bootstrap percentile method is used for computing the pointwise confidence intervals. The R package AdjKM.CIF and shiny application for all practitioners have been developed to create the covariate-adjusted KM functions and CIFs.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2022 program