伴随Berkson-Gage法的一个悖论

A paradox with the Berkson-Gage procedure

目的 :本文考虑 Berkson-Gage法中校正处于危险数的合理性 .方法 :按生存率估计值对终检样本进行分解复得生存率估计值 .后者理应等于前者 .结果 :就 Berkson-Gage法而论 ,后者大于等于前者 ,从而形成悖论 .取消该校正 ,后者便等于前者 ,该悖论即解 .结论 :因此 。

Objecdtive:This paper concerns the rationality of the adjusted number at risk in the Berkson\|Gage procedure. Methods:Decomposing a censored sample based on the estimate of survival rate may yield the estimate again. The latter should be equal to the former. Results:As far as the Berkson\|Gage procedure is concerned, however, the latter is greater than or equal to the former. This is a paradox. The paradox is immediately solved after removing the adjustment and the equality is recovered. Conclusion:Thus the option of the adjustment is open to question.