Threshold Regression for Time-to-Event Analysis: The Stthreg Package

Abstract

In this article, we introduce the stthreg package of Stata commands to fit the threshold regression model, which is based on the first hitting time of a boundary by the sample path of a Wiener diffusion process and is well suited to applications involving time-to-event and survival data. The threshold regression model serves as an important alternative to the Cox proportional hazards model. The four commands that comprise this package for the threshold regression model are the model-fitting command stthreg, the postestimation command trhr for hazard-ratio calculation, the postestimation command trpredict for prediction, and the model diagnostics command sttrkm. These commands can also be used to implement an extended threshold regression model that accommodates applications where a cure rate exists.