Replies to post #222132 on Biotech Values

[DewDiligence]

I'm probing because I've seen cases where investigators have referred to analyses as prespecified when they weren't—and you've probably seen this too.

In this instance, the distinction we're discussing (raw hsCRP vs log-transform hsCRP) is a pretty small matter in the overall REDUCE-IT datatset.

[DewDiligence]

BTW: Technically, tertiary endpoints can’t be considered statsig unless all secondary endpoints have been met (in hierarchical order). Since REDUCE-IT’s lowest-ranking secondary endpoint—all-cause mortality—was not statsig (a close miss), there is no alpha for any of the tertiary endpoints.

The chart in #msg-144802758 shows the REDUCE-IT secondary endpoints in hierarchical order.

The above is a general nitpick; it doesn't change anything we said earlier in this thread.