Reader Comments

Post a new comment on this article

Survival analysis results should be reviewed

Posted by StevenMcKinney on 24 May 2013 at 02:05 GMT

Figure 2 survival curve panel lower right plot shows two survival curves that lie on top of one another, consist of data from only a few dozen cases, with relatively few events, yet shows a p-value of 0.0118. Reported results of the survival analysis do not appear to be reasonable. The p-value of this lower right set of curves is smaller than the p-value of the curves above or to the right, yet survival curves above or to the right show greater separation. Interpretation of reported findings related to these survival curves should be undertaken with caution.

No competing interests declared.