mike_nyc said:

Relative and absolute effects need reading together.

Pushing back on mike_nyc here. I would add the less popular caveat: these trial populations under-represented several groups, older adults and the highest BMI categories among them. The results probably generalise, and "probably" should be stated as an assumption rather than dropped.

Adding the numbers, since they settle part of this. A quick sanity check on any figure quoted here: is it mean or median, is it intention-to-treat or completers, and what was the comparator. Three questions, and they resolve most disagreements in these threads.

FDA_TrackerJim said:

I would add the less popular caveat: these trial populations under-represented several groups, older adults and the highest BMI categories among them.

There is a second half to this that has not been said yet. The gap between trial results and real-world results is consistent and it is not fraud. Trial participants get titration by protocol, scheduled contact, free drug and dietetic support; removing that infrastructure costs a few percentage points every time it has been measured. When your own curve sits below the published mean, that is the likeliest explanation before anything about you or your material.

One thing that is still open after mike_mod’s answer:

How to read a result like this without either dismissing it or over-reading it, since the summaries all read like press releases?

Reporting back.

Follow-up: I read the paper rather than the summary and the qualifier I was missing was in the second paragraph of the results.