An Analysis of the Survival of Gall Bladder Patients in a Tertiary Cancer Center in India using Accelerated Failure Time Models

Abstract

Objective: Accelerated Failure Time (AFT) models are an useful alternative of Cox- PH model to determine the significant predictors affecting the survival of the patients. This article aims to determine the significant prognostic factors of hospitalized Gall Bladder Cancer patients in Rajiv Gandhi Cancer Institute and Research Center, New Delhi, India by applying AFT Models. To the best of our knowledge, this is the first study to be carried out in India identifying the factors of Gall bladder patients using AFTM. Materials and Methods: The data are taken from original proformae of 652 hospital admitted Gall Bladder patients from a tertiary care hospital from Delhi from the period January 2012 to December 2016. These models take the logarithm of survival time, S(t) as dependent variable and prognostic factors as independent variables. Thereby, effect of these prognostic factors is multiplicative and therefore these models can be easily interpreted. AFTM demonstrates the predictor’s effect in terms of time ratio (TR). Analysis was implemented on R software version 3.5.1. Results and Conclusions: In the Gall Bladder data considered in this article, shape of hazard function, H(t) and the exploratory data analysis falls in line with the Lognormal AFT model. AFT models give an estimate of Time Ratio which helps doctors, clinicians, epidemiologists etc. to determine the effect of treatment in terms of an increasing/decreasing survival time.