Using The Generalized Gamma Model To Estimate The Survival Function

Abstract

In many medical studies, the outcome of interest is the time remaining for an event to occur. These include deaths, disease progression, or hospitalization. To aid in decision making, the hazard function is estimated from parametric models which aim to quantify the benefits to patients. Time-to-event (TTE) data with complete follow-up are rarely available. As survival function data, whether complete data or monitoring data of various types, is often in the form of a time series represented by a time period of survival for each patient or for each machine, we notice when analyzing survival data that it suffers from problems of instability or high fluctuation as a result of the abnormal phenomena of this. The idea in the research process was to process survival function data through generalized linear models by using one of the methods of generalized linear models (GLMs), represented by the generalized gamma model, to analyze time-to-event (TTE) data, through a link function linked to people at risk. At the time t that the patient is exposed to, using two methods,  maximum likelihood and weighted iterative maximum potential, to predict survival data.