Sensitivity analysis for missing outcomes in time-to-event data with covariate adjustment

J Biopharm Stat. 2016;26(2):269-79. doi: 10.1080/10543406.2014.1000549. Epub 2015 Jan 30.


Covariate-adjusted sensitivity analyses is proposed for missing time-to-event outcomes. The method invokes multiple imputation (MI) for the missing failure times under a variety of specifications regarding the post-withdrawal tendency for having the event of interest. With a clinical trial example, we compared methods of covariance analyses for time-to-event data, i.e., the multivariable Cox proportional hazards (PH) model and nonparametric analysis of covariance, and then illustrated how to incorporate these methods into the proposed sensitivity analysis for covariate adjustment. The MI methods considered are Kaplan-Meier multiple imputation and covariate-adjusted and unadjusted PH multiple imputation. The assumptions, statistical issues, and features for these methods are discussed.

Keywords: Covariate adjustment; multiple imputation; sensitivity analysis; time-to-event data.

Publication types

  • Research Support, N.I.H., Extramural

MeSH terms

  • Computer Simulation
  • Data Interpretation, Statistical
  • Humans
  • Kaplan-Meier Estimate
  • Outcome Assessment, Health Care / statistics & numerical data*
  • Proportional Hazards Models*
  • Randomized Controlled Trials as Topic / statistics & numerical data*