A Hyperbolastic Model for Survival Data

• Activity Number: 75
• Type: Contributed
• Date/Time: Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
• Sponsor: Biometrics Section
• Author(s): Zoran Bursac*+ and Mohammad Tabatabai and David K. Williams and Karan P. Singh
• Companies: University of Arkansas for Medical Sciences and Cameron University and University of Arkansas for Medical Sciences and University of North Texas Health Science Center
• Keywords: hyperbolastic survival model ; Cox model ; Weibull model ; log-likelihood ; survival probability prediction ; log-logistic model

Abstract

Modeling the outcomes that factor in the element of time has been extensively studied and applied in a wide range of medical and biological studies. In continuation of our work on a family of hyperbolastic growth models (Tabatabai et. al. 2005, Bursac et. al 2006) a new hyperbolastic survival model is introduced. The model is utilized for analysis of a published survival data set and the results are compared with those obtained from commonly used survival models (e.g. Cox, Weibull, log-logistic and log-normal). With the same number of parameters as in the classical models, the new model performs better in terms of values of log-likelihood and survival probability prediction. The new hyperbolastic survival model produces flexible hazards, accommodates a mix of covariates and may be a useful predictive tool in various research fields that employ time-to-event data.