Choosing statistical tests for survival analysis

Abstract

Survival analysis is a very specific type of statistical analyses. Survival analysis is aimed to analyze not the event itself but the time lapsed to the event. This time of interest is also referred to as the failure time or survival time. The time used in survival analysis might be measured in different intervals: days, months, weeks, years, etc. The lengthy studies as a matter of course are preferred for being analyzed since they provide stronger evidence and more reliable results. However, it is practically unfeasible for some of the events to be observed over a long period of time. For example, in a study of the pancreatic cancer, one of the most lethal and rapidly growing type of cancer; researchers might get a very low median for survival time, which may indicate that half of the participants died within just a three month period. The studies, perhaps, would not be stopped at the moment of reaching three or six month period and may continue up until five years, but just on the miniscule, if any, number, of participants. The events in the survival analysis are usually deleterious in nature. The death is the prototypical event for the analysis, termed usually as a failure. Other events, such as an occurrence of a disease, relapse, smoking and drinking resumption, complication of the disease, might be of the research interest as well. The survival analysis methods can be used in other than medicine fields as well: in economics, political science, sociology, engineering.