Paul Dijkstra

How to read a paper involving artificial intelligence (AI) by Paul Dijkstra et al

Reading a research paper that involves using AI in clinical settings?

This article has questions to help you critically appraise AI papers:
bmjmedicine.bmj.com/content/4/1/...

Gonna write a paper about a paradox. Systematic reviews squash primary studies into tables; meta-analyses sum the findings of similar studies. 'Homogeneity' tends to be valued. But the REALLY pivotal studies are often unique, distinctive, imaginative, break-the-rules kind. Thoughts? #academicsky

"This paper is destined to become a citation classic."

Thank you, Reviewer 2. Now, to your 40 suggestions for improvement...

#academicsky

Plenty studies showing that your postcode is a better predictor of your health outcomes than your genome.

Vaccine advisory committees and transparent practices Why changes to ACIP are concerning.

Who’s on RFK’s new vaccine advisory group? What are their conflicts of interest? Read and weep. open.substack.com/pub/sheenacr...

Today I asked AI to summarise an academic paper. The actual results were the opposite to what was hypothesised (and to what others had shown). The AI summary assumed the study had found what everyone else had found. Hot tip: don’t use a stochastic parrot to summarise an outlier paper. #academicsky 🧪

OUR NEW PAPER: the 'final findings' paper from the Remote by Default 2 study:

"After the disruptive innovation: How remote and digital services were embedded, blended and abandoned in UK general practice"

(will do thread on this soon)
www.journalslibrary.nihr.ac.uk/hsdr/publish...

#academicsky 🧪

How to read a paper involving artificial intelligence (AI) This paper guides readers through the critical appraisal of a paper that includes the use of artificial intelligence (AI) in clinical settings for healthcare delivery. A brief introduction to the diff...

Putting together some teaching materials on AI in healthcare - what a useful paper this is bmjmedicine.bmj.com/content/4/1/...

Germany's Plan for an Open and Independent PubMed Safety Net - Absolutely Maybe A few months ago, I wrote about reasons to be concerned about the reliability of PubMed under the new regime at the…

This is so exciting! The German National Library of Medicine @zbmed.bsky.social just had its first open meeting on the plan for an open & independent PubMed safety net. Here's my write-up @plos.org on the meeting & how institutions & others can help absolutelymaybe.plos.org/2025/05/14/g...
#MedLibs

Investigating the impact of trial retractions on the healthcare evidence ecosystem (VITALITY Study I): retrospective cohort study Objective To investigate the impact of retracted trials on the production and use of healthcare evidence in the evidence ecosystem. Design Retrospective cohort study based on forward citation searchi...

Retraction of RCTs (because they were at best flawed & at worst fraudulent) leads to substantial changes in evidence syntheses (systematic reviews and meta-analyses). How many more crap RCTs are lurking? Time to stop hero-worshipping this design. GARBAGE IN GARBAGE OUT.

www.bmj.com/content/389/...

Linda McMahon, Secretary of Education, sent Harvard a letter.

They graded it.

Bwahahaha.

Edith Pechey Lecture 2025 - How digitalisation increases health inequities Edith Pechey Lecture 2025 - How digitalisation increases health inequities with Professor Trish Greenhalgh, University of Oxford

I’m giving the Edith Pechey lecture in Leeds tomorrow (6th May) at 5.30 pm UK time on how digitalisation increases health inequities. Join in person (from 4.30 for drinks) or online from 5.30). medicinehealth.leeds.ac.uk/medicine/eve...

medicinehealth.leeds.ac.uk/faculty-/eve...

Warning signs realised: Trump’s war on science and the fight to save American academia In early March, I warned of authoritarian like attacks on science and universities. Six weeks later and the Trump administration has gone even further than feared.

In early March, I wrote about the proto-authoritarian attacks on science and universities by the Trump administration and listed some warning signs to look out for in the coming weeks.

Now six weeks later and those warning signs are being realised.

christinapagel.substack.com/p/warning-si...

Thanks Alan - however, when you download the PDF they are correctly labelled 1 to 10

Super useful reference including key AI definitions and an approach on how to read a paper including AI in its research. Reminder that “AI… is not a panacea, and it raises both practical and ethical challenges”.

Sometimes good decent people die within hours of meeting nasty evil people. It’s coincidence.

Periodic reminder....😀

#Transparency is one of the six core principles underpinning the WHO guidance on #ethics and #governance of #ArtificialIntelligence for health

--> who.int/publications...

Important to adhere to the TRIPOD+AI reporting standards🙏
--> bmj.com/content/385/...

#trustworthyAI

An excellent addition to the "How to read a paper" fold - focus is clearly on health but the introduction has wide relevance as do many of the questions 📚 #medlibs

How to read a paper involving artificial intelligence (AI) This paper guides readers through the critical appraisal of a paper that includes the use of artificial intelligence (AI) in clinical settings for healthcare delivery. A brief introduction to the diff...

Our new paper!! Led by @drpauldijkstra.bsky.social with @jessrmorley.bsky.social and awesome medical student Yosra Mekki. #digitalhealth #medsky #episky

How to Read a Paper Involving AI.

bmjmedicine.bmj.com/content/4/1/...

How to read a paper The basics of evidence-based healthcare Seventh Edition Trisha Greenhalgh and Paul Dijkstra Wiley Blackwell

How to Read a Paper: The Basics of Evidence-Based Healthcare

The new edition of this text is now available in the e-library. It guides those who wish to ‘assess the scientific validity and practical relevance’ of healthcare literature. #RPSLibrary

ow.ly/kYsU50VuzV3

Book in packing box

It's out! 5th edition of #qual #methods for health research now published, and in a library near you soon!

If, as a doctor, your best scientific take on cancer incidence /recurrence is "I seem to be seeing a few more cases in my own clinic since mRNA vaccines were approved", you need to bone up on cancer epidemiology or shut up.

We are learning the wrong lessons from lockdown And unless we learn the right ones, we risk making the same mistakes in a new pandemic

Lockdowns are terrible. So are deadly pandemics. @chrischirp.bsky.social tells it straight.

christinapagel.substack.com/p/we-are-lea...

In this current harsh world, let's warm a few hearts remembering the events of 7-8 March 1991, when the 90,000 people of Brindisi woke up to find 25,000 Albanians roaming the streets of their town. For perspective, that's roughly equivalent to 25k people suddenly landing in Hastings [Thread] >> 1

Our new paper on the evidence base for Physician Associates and Anaesthetic Associates in the UK is out today! Coauthor @martinmckee.bsky.social has done a thread here. #healthpolicy #academicsky 🧪

Physician associates and anaesthetic associates in UK: rapid systematic review of recent UK based research Objective To summarise research on the efficacy and safety of UK physician associates and anaesthetic associates in the context of an ongoing policy review. Design Rapid systematic review. Search st...

Paper here www.bmj.com/content/388/...

Systematic reviews of non-RCT evidence: building dry stone walls To bake good cookies, start with good cookie dough. To use a different metaphor, to build a brick wall, take a large collection of bricks—all the same size and in perfect shape—and line them up neatly...

In the excitement over our new review of PAs, I forgot to tell you about the OTHER paper published in BMJ today, my commentary on the challenges of doing systematic reviews of evidence when there's no RCTs and most of the primary studies were unique and non-standardised.🧪
www.bmj.com/content/388/...

This is why @kityates.bsky.social post is so relevant today

kityates.substack.com/p/science-is...

and why my dystopian piece is perfectly plausible.

christinapagel.substack.com/p/a-dystopia...