Discussion about this post

User's avatar
Frank Harrell's avatar

Very nice post John. One minor quibble: "Major bleed was 69% more in ..." implies something different than a hazard ratio of 1.69 to most people. I'd recommend reporting "The hazard rate of major bleed was at least 1.23 x larger in DOAC (with 0.95 confidence; point estimate 1.69), with a cumulative incidence at ... months raised from ... in the control arm to ... in DOAC (0.95 CI for difference in cumulative risk ...)." Supplementing that with a Bayesian probability that the risk difference is clinically important would be nice.

How did minor bleeds fare? What would be the result of an analysis that used bleeding more as a continuous variable (e.g., longitudinal analysis)? Or an analysis that distinguished GI bleeding from the worse cerebral bleeding (e.g., ordinal longitudinal analysis)?

Expand full comment
Amanita's avatar

Frail -AF. That is quite the descriptive name for this study.

Expand full comment
30 more comments...

No posts