A STATISTICAL TEST FOR THE ACTUARIAL SURVIVAL IN CANCER

The Actuarial survival is a macroscopic goal in cancer treatment, for which there has been no satisfac-toryevaluation procedure. A nonparametrio significance test is proposed for the differenco of cohort half-lives between sample and population. It is expressed as the ref-lected statistic the test-based statistio and used for evaluating the actuarial survival. Kaplan-Meier or Berkson-Gage estimate is utilized via the exponential interpolation equation for adjusting the effect of censorship so that it is suitable for various kinds of data: censoredor uncensored, grouped or ungrouped. It reduces to the classical median test when there is no censoring as all generalizations for adjusting the effect of censorship have the reducibility. A worked out example illustrates the methodology.