Niall Harrison

Fig 1A from Kobak et al 2025, showing words that have increased in use in biomedical abstracts in the last few years and thus may be indicators of LLM use

Zoomed in part of the previous figure, highlighting "steatotic", which does not seem like it would naturally come up much in LLM output, which may instead be due to the new disease nomenclature from 2023.

You know your disease rebrand has been successful when the key term sees as much of an uptick in use as tell-tale words from LLM content (see alt text): www.science.org/doi/10.1126/...

Science research gets more engagement on Bluesky than X, study finds Bluesky posts referencing scholarly articles ‘find substantially higher levels of interaction’ than on Elon Musk’s platform

NEW: Bluesky is becoming place to discuss scientific research - it gets more proper engagement & attention, new study finds.

For all the talk of decline - borne out by several metrics - the science community seems to have adopted Bluesky more than others.

Story: www.theguardian.com/technology/2...

Text reads: “Deaths from viral hepatitis-related causes are increasing. 2019: 1.1 million deaths 2022: 1.3 million death

Today is World #Hepatitis Day.

Did you know❓ Hepatitis is the second leading infectious cause of death globally, with hepatitis B and C claiming 3 500 lives every day.

Learn how to protect yourself and your loved ones 👉 bit.ly/hepatitis2025

🆕CPG in press❕

EASL Clinical Practice Guidelines on the management of hepatitis B virus infection

Find it here👉https://bit.ly/4deZHf1

#LiverSky
@easlnews.bsky.social

NEW PAPER: Updated CONSORT-2025 for reporting randomised trials is now available in the @bmj.com @jama.com , @thelancet.bsky.social, @plos.org and @naturemedicine.bsky.social

—> www.bmj.com/content/389/...

#openscience #transparency #medsky #statssky #episky

A pie chart with three segments. The largest segment, over half the chart, is labelled "authors who have not ventured an opinion". The next largest, about a third of the chart, is labelled "authors who have stated they opposed lie charts". The smallest segment, about one-eighth of the chart, is labelled "authors who are well aware that other actors dislike pie charts, but committed pie chart crimes anyway." The composer of this alt text notes that if pie charts habitually came with alt text they would be *marginally* less opposed to them.

Members of the ACCORD group are debating data visualisation options for their upcoming @ismpp.bsky.social poster.

Clinical Trials Information System designated as WHO primary registry | European Medicines Agency (EMA) Designation facilitates data sharing, promotes transparency and trust in clinical research

The Clinical Trials Information System #CTIS has been designated as primary registry by the @who.int !
This designation represents a significant step forward in promoting transparency and trust in clinical research in the EU.
#clinicaltrials #clinicalresearch www.ema.europa.eu/en/news/clin...

Anyone who attended this year's #CROI2025 will tell you it was a very strange experience given the current political climate. Still, plenty of interesting studies -- here's a Really Rapid Review of selected highlights. #IDSky #HIVSky blogs.jwatch.org/hiv-id-obser...

Two new cases of HIV remission after stem cell transplant for cancer reported at #CROI2025. These have become so routine that they're posters instead of oral presentations. We know an HIV cure is possible -- now we need to use clues from these rare cases to develop widely applicable approaches.

Yes, of course. I meant my comment very narrowly, with reference to the outage over the weekend, which does now seem to have been resolved. But agree that in the medium-to-long term the politics could make that sort of technical failure more common.

Current PubMed issues appear to be (fingers crossed) more technical than political, but good to remind yourself of the alternatives.

This afternoon! Or this morning for those in the US. Or this evening for those in AP, I guess. Either way: not many hours from now.

Read our editorial on the need for communication and collaboration when developing reporting guidelines @equatornetwork.bsky.social

Always powerful to see the graph showing the impact of PEPFAR blogs.jwatch.org/hiv-id-obser...

What's Happening Inside the NIH and NSF

A long post about what’s happening to the science funding agencies in the US and why. As mentioned, this one just kept getting longer even as I kept stripping curse words from it.

www.science.org/content/blog...

Fascinating comment in passing there about ASCO having been spammed with AI submissions -- does anyone know anything more about that? #ISMPPeurope

(Someone has just made this point! A balloon debate on publication tools might be interesting next year... #ISMPPeurope)

I do think AI was a little short-changed, on the grounds that it's a tool not a format, so it's not quite a fair comparison #ISMPPeurope

An occasional issue with the day 2 repeated-plenary format is that if the plenary leans on presentation or panel discussion, round 2 can be a bit stale. This format turns the structure - the fact that the audience is what changes - into a strength. #ISMPPeurope

The balloon debate on publication formats made for an interesting session, and clever choice of format #ISMPPeurope

Enjoying the "maintaining quality in a fast-paced world" session. I think I might actually quite enjoy a 48-hour "all the authors in a room writing the draft" challenge... #ISMPPeurope

Important update: found a stand offering Venchi. I approve even more. #ISMPPeurope

Key takeaway from the exhibition: lindor are the giveaway of choice. I approve. #ISMPPeurope

Picture of a food label that says "garlic lemon boneless chicken tights with roosted sweet potatoes"; presumably meant to be chicken thighs and roasted sweet potatoes

The lunch labels may have needed a proofread! #ISMPPeurope

In mice is on Bluesky as well! #ISMPPeurope

A photo of a plenary meeting room full of people, listening to Rob Matheis deliver the Chair's Welcome at the ISMPP EU 2025 meeting

The @ismpp.bsky.social EU meeting is underway! #ismppeu2025

Clinical Trial Readouts Are A-Comin'

Some clinical trial results to look forward to this year:

Artificial intelligence is infiltrating health care. We shouldn’t let it make all the decisions. AI paternalism could put patient autonomy at risk—if we let it.

Interesting piece in @mitpress.bsky.social Tech Review about #AI paternalism in healthcare. i.e. "the doctor knows best" being replaced by AI knows best, both dismissing patient experience: www.technologyreview.com/2023/04/21/1...

Ah yes, that early January "suddenly it's only three weeks to @ismpp.bsky.social why didn't I do more prep in December" feeling

Now you have to make a medcomms starter pack!