失访和迄今存活病人的生存率

癌症临床工作的关键指标是生存率，尤其是那些失访和迄今存活病人即终检（也称删失）病例的生存率。后者尚无估计方法。本文提出生存率终检模型，利用该模型从整个样本生存率分离出终检病例生存率，并按二项分布理论构造其方差和置信限。其统计学意义是它揭示出生存率的终检成份和非终检成份，临床意义是它实现了对失访和迄今存活病人生存率的估计。与以往的笼统生存率估计相比，这显然提高了一步。附有工作实例描述其临床应用

Survival Rate in Lost and UP-to-Date Alive patients

The key measure in cancer clinical work is survival rate, especially the survival rate in lost and up-to-date alive patients, ie. the censored cases. The latter, however, there has been no way to estimate. This paper proposes the censoring model of survival rate, separates the estimate for survival rate in censored cases from that for whole sample using the model, and derives its variance and confidence interval for it based on binomial theory. The statistical significance of the model is that it discloses the censoring element and the uncensoring element of survival rate and the clinical significance is that it realizes the estimation for the survival rate in lost and up-to-date alive patients. This is a remarkable progress by contrast with the previous estimation for survival rate in whole sample. A worked example illustrates its use in clinical practice