Bob Koure
Mar 19, 2023

--

Agreed that journalists tend to over-sensationalize study results.
That said, I don’t think hazard ratios are appropriate for this kind of study, and a responsible journalist would have written them as two different absolute risks (most people don't understand relative risk - something I've seen used to minimize the risk associated with drugs), plus how likely the results are to apply to everyone (confidence interval and p-value being similarly mysterious).
Easy explanation of all these at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7515812/
Finally, these seem to be effectively correlation studies, in spite of coming out of what might have been RCTs.

--

--

Bob Koure
Bob Koure

Written by Bob Koure

Retired software architect, statistical analyst, hotel mgr, bike racer, distance swimmer. Photographer. Amateur historian. Avid reader. Home cook. Never-FBer

Responses (1)