Modeling of Time to Event for Healthcare Data Using Survival Analysis

Survival analysis is an aggregate of statistical methods for analyzing data where the outcome variable is time until a specific event occurs. This approach contains the type of problems where the covariate deliberates, the necessity to take into consideration "censored data". The time represents any interval of time from the start of an individual's follow-up until an event occurs. Cox provides clear examples of accepted techniques for the study of survival data (1972).In survival analysis the term 'failure' is used to describe the incidence of the event of interest

(although the event can really be a 'success' such as restitution from therapy). The term 'survival time' denotes time taken for failure to occur. In survival analysis, the time covariate known as "survival time" denotes that a person has endured over a number of follow-up periods. (Kalbfleisch and Prentice, 1980). In many different domains, survival analysis is also used to analyse data relating the length of time between two events. It is also called as event history analysis, lifetime data analysis,

consistency analysis or time to event analysis (Hosmer and Lemeshow,1999). A covariate that is directly relevant in many studies is the amount of time that passes before an event happens. The study of survival information has been targeted on predicting the likelihood of

response, survival or mean period of time, comparison of the survival distributions of human patients or of animals under study. The identification of risk or prognostic factors associated with response, survival and therefore the development of an illness has become equally vital (Lee, 1992). Due to the sensitivity of some of the approaches employed to underlying assumptions, it is very crucial to examine any assumptions as part of the analysis. (Cox and Oakes, 1984).