Adjusting for Treatment Switching in Randomised Controlled Trials in the Context of Economic Evaluation - Simulation Studies and a Novel Two-Stage Method

• Activity Number: 488
• Type: Topic Contributed
• Date/Time: Wednesday, August 6, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Nicholas Latimer*+
• Companies: University of Sheffield
• Keywords: Treatment switching ; Economic evaluation ; Survival analysis ; Treatment crossover ; Health technology assessment ; Prediction

Abstract

Economic evaluations must incorporate the overall survival impact of a new therapy and in the presence of treatment switching, an intention to treat analysis is likely to underestimate this. Several switching adjustment methods have been advocated in the literature, but it is unclear which are likely to produce least bias in realistic RCT-based scenarios. We simulated RCTs which permitted switching, varying several parameters over a range of scenarios. The performance of alternative methods was assessed based upon bias, coverage and mean squared error, related to the estimation of true restricted mean survival in the absence of switching. When the treatment effect was not time-dependent, Rank Preserving Structural Failure Time Models (RPSFTM) produced low bias. However in the presence of a time-dependent treatment effect these methods performed less well. The Inverse Probability of Censoring Weights (IPCW) method produced high levels of bias when switching proportions exceeded 85%. A novel two-stage accelerated failure time method produced low bias across most scenarios and provided the treatment switching mechanism is suitable, represents an appropriate adjustment method.