ostermayer on Nostr: here is an example of how a pharmaceutical company can create a trial to make a drug ...

here is an example of how a pharmaceutical company can create a trial to make a drug ( in this case xofluza) look good while being essentially useless. many urgent cares are giving flu patients this med to "treat their flu".

first they create a primary subjective outcome (time to symptom improvement)

that endpoint showed "29 hr time" benefit for symptom improvement but when looking at return to pre-illness health it showed no significant difference compared to placebo (126.4h vs 149.8h, p=0.46)

they will choose healthy patients and exclude those who may suffer adverse effects - and they exclude hospitalized patients who may not even show a benefit

they will also focus on surrogate outcomes such as viral load irrespective of symptoms correlated to viral load

then they will bury a red flag deep in the paper where 9.7% of baloxavir-treated patients with H3N2 developed resistance mutations during treatment.

then they will perform a modified intention-to-treat (mITT) analysis and exclude 2.6% of patients from non-GCP-compliant site to potentially minimize data on patients with side effects that prevent their trial participation

all of this creates a publication that can headline as a great drug but really fails to improve patient health compared to placebo.

https://files.ostermayer.co/ison2020.pdf