RT @jwjacobs12: @AABB @DrGSBooth1 @DevaSharma1981

In this issue of @NEJMEvidence @jwjacobs12 and I interpret the recent individual-patient meta-analysis for RBC transfusion in myocardial infarction by Carson et al and make a case for nuanced interpretation Article:

@katy_milkman @Wharton sigh

@AaronGoodman33 I've been doing the math thing too. Really fun. Not motivated by physics in my case

Should we eliminate leukocyte reduction for whole blood? It's an intriguing concept with some (mainly logistical) benefits In @AABB journal, my brilliant co-authors argue it's a bit premature, and removal of LR deserves deeper, empirical scrutiny

Our correspondence with @Nature advocating that scholarly journals support direct link-sharing to @Bluesky amidst the exodus of many academic communities from X (formerly Twitter) @jwjacobs12 @path_brian @DrGSBooth1

@kepano Sharing my .json for anyone who just wants to import it!

@getjonwithit Knowing very little of the physics you mentioned but working with models in other areas: sometimes counterintuitive predictions of models turn out to be insightful ± true when tested. How do you think through what to test and discard? Feasibility, usefulness, other criterion? TY.

@AnilMakam Different context

How does a blood donation become a transfusable blood product? A new resource I helped develop for @CanadasLifeline provides an extensive walkthrough on the process of blood processing and delivery--for medical and curious non-medical folks #transfusion

Applications and advantages of multi-agent AI over single-agents for healthcare use

@tony_breu I would combine 3 with 8 and 9. I never want to see another red cell transfusion trial with a "placebo" arm which ignores iron deficiency, the most prevalent cause of anemia anywhere you look

@graham74GC These GWAS-y studies can be quite tricky to interpret. I haven’t found a way to separate unexpected disease associations from quirks in billing codes or biases introduced at data collection. Still useful for generating hypotheses

@iamlindachang @PNASNews Great paper. I came across this idea as the "McNamara Fallacy", good to see a rigorous demonstration of it, and intriguing theory of how it arises

@AaronGoodman33 For me it’s less about disagreements and more about the increasingly dizzying irrelevance of algorithm and bot posts. I used to be able to block this out, now it’s the majority of what I am seeing. Not planning on leaving, the useful-information density has dropped though.

@ruchika_goel1 @margitangyal @TeamCaptainJohn @CPSkorupski

@AdamRodmanMD @RichardLehman1 Fascinating! Adam, would you say based on these findings that the current best use of these untrained LLMs is simply to feed them straight facts rather than facts + clinical impressions?