Connecting Threshold Regression and Accelerated Failure Time Models

摘要: The accelerated failure time model is one of the most commonly used alternative methods to Cox proportional hazards when assumption violated. Threshold regression a relatively new for analyzing time-to-event data with non-proportional hazards. It based on first-hitting-time models, where can be modeled as at which stochastic process interest first hits boundary or threshold state. This paper compares and models demonstrates situations becomes special case model. Three illustrative examples from clinical studies are provided.