The big reveal starts. #escmidglobal

Following up, Gunnar also shows the problems with the screening tests. It's a challenge in high R area #escmidglobal #escmid2025

Totally agree, not sure why it's that relevant.... (Or we could just repeat the test?)

Gunnar from EUCAST talking about the complexities of susceptibility testing using oxacillin disc screening. #escmid2025

It zero shots stuff in R, complex plots with in built linear models, saves to a pdf, everything works. Incredible for simulations.

NEW PREPRINT!

Enjoyed working with @tom-parks.bsky.social and @petedodd24.bsky.social on immortal time bias

In its simplest form, ITB results from misallocation of person time, i.e. pre treatment time in intervention group being counted as time on treatment

www.medrxiv.org/content/10.1...

[1/n]

Ah thanks I should have done this!

We are looking for patient, clinician and researcher insight into priorities for community acquired pneumonia!

This is a JLA Priority Setting Partnership and will drive the UK research agenda for years to come - please share!!

⛄🎄⛄🎄!!

forms.office.com/Pages/Respon...

Just a short piece on FQ prescribing. Basically, it's gone downhill, in relation to comparator antibiotics and absolutely. Still waiting for the MHRA to release the report with new information...

NEW LETTER! Me and @gushamilton.bsky.social argue that, if you want to demonstrate all oral treatment for pneumonia is adequate, then systematically excluding from sample people who are switched to IV after failing oral therapy is problematic

www.clinicalmicrobiologyandinfection.com/article/S119...

No... But they offer a great service generally!

Just really cool.

What an absolutely lovely paper. Important topic. I wonder what it would cost today and how much the total cost was. Such a lovely bit of work. Will keep this for talks and presentations.

As in, we think shorter is better. So why would there be increased harm with a day less antibiotics? I just can't really square it with e.g. balance. And the CRP arm mortality was about the same as PCT arm. I think it's just small numbers.

I thought about this but I think it's just a bit too imprecise and I think the mechanism for increased mortality is not really there. So I *think* we are just seeing noise.

regardless, a very, very impressive trial and well run. Congrats to all authors! @darknatter.bsky.social

I don't think you can easily claim that PCT reduces abx use but that the mortality is the same. It's just too imprecise to call mortality. We can't infer much about the effect on mortality. I am not sure the NI margin is what patients/clinicians would accept.

Biomarker-Guided Antibiotic Duration for Patients With Suspected Sepsis This clinical trial assessed whether treatment protocols for monitoring C-reactive protein or procalcitonin safely reduced the amount of time critically ill adults with suspected sepsis received antib...

Interesting trial. If we took the point estimates, then we would accept 0.9 days less abx per person, for a 1.6% absolute mortality increase. Not sure I would take that. Obviously, imprecision means both of these are less precise; but I think not a clear 'win' for PCT
jamanetwork.com/journals/jam...

I saw one just this week! If only you'd published it last week...

Don't
I've got too much to do

I definitely think this is underexplored. lots of appropriate datasets in which to study this. Very interesting!

Colin, wouldn't one analysis compare different ages as we know drinking has reduced in younger people. So if we see larger effects in younger folk makes less likely to be alcohol related.

This is lovely work Paul. Lovely. We are just thinking about in vitro systems and phages and how to start testing this systematically. Such interesting data. Thanks for sharing.

This must be in party diagnosis of fatty liver and increased use of e.g. fibroscans that pick up these. But not a hematologist or an epidemiologist. But I can't quite tie it together.

And what a paper to choose ;)

Not gonna disagree with a Nobel prize winner on the economics here, but people underestimate Trump and the power being crazy has. If people think he will do something crazy, they change their behaviour in funny ways(and sometimes in his favour). I don't think it's a rational play, that's the point.

Yes I meant to do tom too then fat fingered and not used to blue sky!

Although I suppose cough is biology 🤣

What's the best book for understanding LLMs and their ilk? I understand basic statistics and computing (i.e. I can use bash, python, and HPC clusters). So something that works from there upwards. Like explains transformers.

Oh and I would love to see more about the smoking association. MVMR would seem super interesting here. Lots to think about. I always wonder about increased diagnosis in smokers due to cough as well as actual biology. Can't wait to see new preprint!