Open AccessBrief introduction to parametric time to event model
Author(s) -
HyeongSeok Lim
Publication year - 2021
Publication title -
translational and clinical pharmacology/translational and clinical pharmacology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.166
H-Index - 6
eISSN - 2383-5427
pISSN - 2289-0882
DOI - 10.12793/tcp.2021.29.e7
Subject(s) - covariate , proportional hazards model , accelerated failure time model , weibull distribution , parametric statistics , statistics , parametric model , hazard , survival analysis , logistic regression , econometrics , event (particle physics) , time point , computer science , mathematics , physics , quantum mechanics , philosophy , chemistry , organic chemistry , aesthetics
This tutorial explains the basic concept of parametric time to event (TTE) models, focusing on commonly used exponential, Weibull, and log-logistic model. TTE data is commonly used as endpoint for treatment effect of a drug or prognosis of diseases. Although non-parametric Kaplan-Meier analysis has been widely used for TTE data analysis, parametric modeling analysis has its own advantages such as ease of simulation, and evaluation of continuous covariate. Accelerated failure time model is introduced as a covariate model for TTE data together with proportional hazard model. Compared to proportional hazard model, accelerated failure time model provides more intuitive results on covariate effect since it states that covariates change TTE whereas in proportional hazard model covariates affect hazard.