A methods-focused walkthrough of the head-to-head trial, because effect-size headlines are doing a lot of unearned work in this comparison.
Reading SURMOUNT-5 critically: endpoints, dropouts, and what the 95% CIs hide
Replies
No replies yet. Be the first to add one.
The treatment-regimen-estimand vs. efficacy-estimand distinction is the single most misreported aspect. Both numbers are real; they answer different questions.
Also worth flagging the open-label design — hard to avoid with distinct devices, but it matters for patient-reported secondaries.