Tied Survival Times In Survival Analysis

• Main Authors: Durdu Karasoy, Sena Keskin Kaplan
• Format: Article
• Language: English
• Published: Istanbul University 2017-06-01
• Series: Alphanumeric Journal
• Subjects: Cox Regression; Survival Analysis; Tied Survival Times
• Online Access: http://alphanumericjournal.com/media/Issue/volume-5-issue-1-2017/yasam-cozumlemesinde-es-zamanli-yasam-sureleri.pdf

Survival analysis is generally defined as a set of methods for analyzing data where the outcome variable is the time until the occurrence of an event of interest. One of the methods commonly used in the survival analysis is Cox regresion model which is used to determine the factors that impact on survival times. Cox regression model has assumptions. One of them is proportional hazards assumption and the another one is there is no tied data between event times. However, in real applications, tied event times are commonly observed and Cox's partial likelihood function needs to be modified to handle ties. It is well known methods that the Exact method, Breslow method, Efron method and Discrete method for handling tied event times. Firstly, the methods are analysed during the study, Breslow, Efron and Exact methods, which is applied on a stomach canser data set (there is tied data between event times) It was decided that Cox regression with Exact Method is the best model. Than this methods is applied Acute Myocardial Infarction data set which has no tied data between event times and it is found the same resuts at all methods.