在做一个生存分析,发现一个比较有意思的现象,在比较两组生存率的时候LOGRANK的结果和WILCOXON结果不一致。
如下文:
. sts test case, logrank
failure _d: outcome
analysis time _t: hospitalduration
Log-rank test for equality of survivor functions
| Events Events
case | observed expected
------+-------------------------
1 | 74 58.83
2 | 70 85.17
------+-------------------------
Total | 144 144.00
chi2(1) = 7.34
Pr>chi2 = 0.0067
. sts test case, wilcoxon
failure _d: outcome
analysis time _t: hospitalduration
Wilcoxon (Breslow) test for equality of survivor functions
| Events Events Sum of
case | observed expected ranks
------+--------------------------------------
1 | 74 58.83 862
2 | 70 85.17 -862
------+--------------------------------------
Total | 144 144.00 0
chi2(1) = 2.98
Pr>chi2 = 0.0843
.
大家平时是否遇到?
该如何解读呢?请大家不吝赐教。