by
A.Antoniadis, J. Berruyer and A. Filhol
1996

MacSurvival (English) / MacSurvie (French) are two versions of the same a Mac OS 9 software designed for the analysis of censored data, i.e. for the study of lifetime, failure time or survival time.

These applications were never ported to a modern operating system. However you can still run them through a tool like SheepShaver (for Linux or Mac OS X)  which makes it possible to install a Mac OS 9.0.4 runtime environment on a modern computer.

MacSurvival was developed for the type of modern statistical data analysis most useful in practice and especially for clinical and nonclinical investigators who want to apply appropriate statistical tools for censored data. The program is designed for easy use and full interactivity. It eliminates many of the set-up requirements and unnecessary constraints common in similar applications.

Survival analysis originated in the study and analysis of length of times to a response. The response is an event that occurs at a specific point in time, and is often called a death or a failure, hence the terms lifetime, failure time or survival time. Fields of study other than medicine have given other names to the identical methodology discussed here. Because of the focus of most of the examples in this manual we call it Survival analysis.

The ability to use data from cases for which the response has not yet occurred distinguishes survival analysis techniques from other statistical methodologies. Such data are called incomplete or censored. They may arise from loss to follow-up (that is the patient may have moved away, or refuse to participate further in the study), death from causes other than the one under study, or no response before the end of the study.

The methods that are implemented in MacSurvival assume that the chance of censoring is not related to the individual's survival times. This assumption is reasonable when entries and losses to follow-up occur randomly among patients. If certain patients are lost as a consequence of treatment, their chances of being censored are not independent of their "death" times. Modeling of survival times is based on two distinct approaches: parametric (mainly used in the industrial and manufacturing world) and nonparametric (widely used in clinical trials). The examples in this document cover both approaches.