News for hackers. Stuff that matters.
Re:New treatment destroys head and neck cancer tumours in trial by addaon on Monday October 11, 2021 @11:30AM (#28829465)
The challenge is that they found strong results only for those patients with high level of PD-L1, which was a subset of the patient population. When you start looking at subsets, you either need a larger population to begin with, or you need to select for that subset in your test population.
Re:New treatment destroys head and neck cancer tumours in trial by catlikesshrimp on Monday October 11, 2021 @02:52PM (#28831581)
^This

You can't draw conclussions from subsets of the study. This is a rule that even professionals often forget.