Evaluating differences in survival between two groups of patients.

摘要: MEDICAL investigations often result in a comparison of the survival experience two groups patients. The usual method making this involves computing curves for both and evaluating any apparent differences on basis test significance. From observed such statistics as 5-yr rate or median time may be used comparative measure. Differences found real they due to distribution with respect factors that are associated survival. For example, Cutler, Asire Taylor [I] analyzing data series disseminated breast cancer patients who experienced natural menopause subsequent diagnosis their primary had better following dissemination than those prior (see Fig. 1). They pointed out, however, general longer interval between disease occurrence dissemination. Since previously reported was related [2]), concluded advantage could explained disease-free interval. This example is discussed detail section 4 illustrate methodology presented sections 2 3. Mantel [3] has proposed significance while taking into account thought Associated statistic called relative odds mortality. association can considered differ significantly from unity when chi-square large. While primarily having similar distributions might survival, he indicated analysis readily extended situation where not case. We wish consider Mantel’s suggested extension procedure which do have over important