Christopher Russell, MD, MS (he/his/him)
banner
cjrussellmd.bsky.social
Christopher Russell, MD, MS (he/his/him)
@cjrussellmd.bsky.social
Associate Professor, Pediatric Hospitalist, Stanford Medicine Children’s Health. Physician-scientist & clinical researcher studying children with medical complexity. Views my own. 🏳️‍🌈

More at https://med.stanford.edu/profiles/314127
🚨🚨Please amplify and consider joining the lawsuit or submitting an amicus brief: @academyhealth.bsky.social @atscommunity.bsky.social @ats-peds.bsky.social @ameracadpeds.bsky.social
September 7, 2025 at 2:23 AM
Or be clear in the abstract that it was underpowered (understanding that this was not due to any fault of the investigators). However, noting this as a randomized, multicenter, double-blinded study in the title but not acknowledging that it was underpowered in the abstract is concerning.
January 30, 2025 at 6:40 PM
Can’t reliably estimate the rare outcomes if you are so underpowered. IMO, underpowered RCTs, particularly those with null findings or non-inferiority design, should rarely be published. People see RCT and think high level of evidence, and assume the conclusions are sound.
January 30, 2025 at 6:40 PM
They also (incorrectly) state in the abstract that “Placebo appears to be non-inferior to amoxicillin in reducing fever duration” but then correctly say in the discussion that they “must reject the hypothesis of non-inferiority of placebo over amoxicillin.” Can’t move the goal posts…
January 30, 2025 at 6:27 PM
The primary outcome is odd—fever duration—and they are quite underpowered to detect differences in adverse events. The adverse events aren’t benign (~3% in placebo with RPA), which may have been prevented by upfront treatment. Not sure how to use this in my practice.
January 30, 2025 at 6:20 PM