A Case Study Evaluating Time Varying Covariate and Competing Risk for Efficacy Evaluation in Survival Analyses

• Activity Number: 423 - SPEED: Biopharmaceutical Statistics, Medical Devices, and Mental Health
• Type: Contributed
• Date/Time: Tuesday, August 1, 2017 : 2:00 PM to 2:45 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Yang Xu*
• Companies: BioStat Solutions, Inc.
• Keywords: Survival Analysis ; Time Varying Covariate ; Competing Risk ; Cox Proportional Hazard (Cox PH) Model ; Cumulative Incidence Functions (CIF)

Abstract

In clinical studies, survival endpoints are generally modeled using Cox PH with baseline information as static covariates. With longitudinal data, using time varying covariates in Cox PH model to assess the prodromal influence is more appropriate. Using a case study, it demonstrates applying covariates measured over time displays a significant impact on survival endpoints compared to only considering baseline information. Survival analyses are further complicated by subjects who die without experiencing the study primary outcome. Multiple reports show that cause-specific hazard approach overestimates the risk of disease by failing to account for the competing risk. The CIF provides both a nonparametric and semiparametric solution. Instead of the log rank test, Gray's test is applied in nonparametric scenarios. The hazard ratio is used when modeling the cumulative incidence compared to the standard Cox PH model. Case study shows CIF could detect statistical difference between survival curves of treatment and placebo while log rank is not able to do so. Both methods can be implemented conveniently in SAS 9.4, and the benefit of these methods will be demonstrated in a case study.