Richard Hooper

A useful reminder of the original meaning of a "stepped wedge".

The #UKCTUNetwork is committed to fostering a supportive, collaborative & transparent environment across CTUs—driven by continuous improvement, knowledge-sharing, and mutual respect for experience and insight.

Learn more: ukcrc-ctu.org.uk/network-stra... #ClinicalTrials #UKCTUNetworkStrategy

I'm not sure that clickbait always gets me.

Good news is we all survived another trip to the vet for an annual check-up. Bad news is that this guy needs more dental work.

One of our excellent trial managers Lana reflecting on what makes a successful trial

Our lead of data management Arouna sharing the clinical trial life cycle

Beautiful day for the @qmul-wiph.bsky.social showcase! It has been fabulous to hear all the work that is being carried out across our 6 Centres. But all of it underpinned by excellence in methods, as highlighted by some of our staff in @wiphcem.bsky.social

It's quite nice out

Just overheard a young boy saying "the trouble is that all the really good athletes in our class are also taking part in the maths challenge". There's hope for the next generation.

I was actually trying to think last week "What's that conference that all the methodologists rave about?" Oh, yeah, this one.

Even just making a little hole in the fence, or leaving out some old logs, could help.

I'm seeing a lot less content on Bluesky about stepped wedge trials than on the blue bird website. Maybe no bad thing, but think of @steppedwedgehog.bsky.social who is poised to re-post any interesting content.

Stepped wedge devotees step up, and cluster enthusiasts get ready to cluster: it's conference time again.

Great to contribute to a session today on how our Institute can support applicants for Fellowships; not so great it was scheduled from 12-2 with no refreshments. The messaging "we support you, but lunch is for wimps" was a bit off... I did feed back to the organisers with a plea for more sustenance!

A picture of the title and abstract of the research paper: All sizzle, no steak: AI tools are not able to act as credible knowledge brokers by summarising evidence in mathematics education Lucy Rycroft-Smith, Darren Macey University of Cambridge Current discourse around AI (artificial intelligence) tools suggest they can offer users quick access to research evidence, with the goal of improving our practice and policies in mathematics education. However, recent evaluations suggest that AI often falls short of human capabilities in accurately summarising complex information, raising questions about trust, credibility, and reliability. In this paper, we investigate whether AI tools are able to successfully perform the educational knowledge brokering function of finding, selecting, and summarising research with implications for practitioners. In particular, we ask if AI tools are able to find research papers that actually exist (accuracy), are relevant to the question (relevancy) and are able to interpret findings into useful, actionable implications for practitioners (applicability). We find that generally they are not, and even when they are, they do so in a way that is so inconsistent as to be a serious threat to credibility. Keywords: artificial intelligence; AI; knowledge brokering; large language models; mathematics education

I am delighted to announce that my new paper with @darrenmacey.bsky.social is now published, entitled 'All sizzle, no steak: AI tools are not able to act as credible knowledge brokers by summarising evidence in mathematics education'

bsrlm.org.uk/wp-content/u...

📢UK undergraduates thinking about careers in medical statistics, clinical trial management, or other areas of applied health research: here's a great opportunity to join a multidisciplinary team on a paid internship (London-based) this summer. Pass it on! qmul-jobs.tal.net/vx/mobile-0/...

Door not dear, duh

Such a beautiful time of year in suburban South London: I step outside my dear and hear a blackcap singing -- the nightingale of the north -- and even (with the wind in the right direction) the contact calls of a pair of nesting peregrine falcons. The small stuff helps me deal with the big stuff.

A.I. Is Getting More Powerful, but Its Hallucinations Are Getting Worse

“On one test, the hallucination rates of newer A.I. systems were as high as 79 percent.”

Well, that’s… useless.

www.nytimes.com/2025/05/05/t...

Importance of sample size on the quality and utility of AI-based prediction models for healthcare Rigorous study design and analytical standards are required to generate reliable findings in healthcare from artificial intelligence (AI) research. On…

**New Lancet DH paper**

"Importance of sample size on the quality & utility of AI-based prediction models for healthcare"

- for broad audience
- explains why inadequate SS harms #AI model training, evaluation & performance
- pushback to claims SS irrelevant to AI research

👇
tinyurl.com/yrje52fn

Study calls for more diverse representation in statistical methodology research Embedding equality, diversity, and inclusion (EDI) into statistical methodology research may improve outcomes and better serve the public, a new Le

A new Leicester study highlights the need for greater diversity in #StatisticalMethodology research, showing that embedding equality, diversity and inclusion may improve outcomes and better serve the public.

@ethnichealthres.bsky.social ‪@proflauragray.bsky.social
arc-em.nihr.ac.uk/news/study-c...

Description Please note that job descriptions are not exhaustive, and you may be asked to take on additional duties that align with the key responsibilities ment...

🔈Job alert! We are looking for an ambitious senior trial statistician to join an exciting new partnership between Imperial College London & Imperial College Healthcare NHS Trust, the Fleming Initiative, which aims to tackle antimicrobial resistance (AMR). www.imperial.ac.uk/jobs/search-...

I'm going to leave the snails alone to enjoy the clematis too

Change from baseline and analysis of covariance revisited The case for preferring analysis of covariance (ANCOVA) to the simple analysis of change scores (SACS) has often been made. Nevertheless, claims continue to be made that analysis of covariance is bia...

Senn 2006? onlinelibrary.wiley.com/doi/epdf/10....

If your organisation is accredited with Athena Swan, or would like to be (hopefully that's most UK universities) then a simple thing you can do to show commitment is not to schedule meetings outside core hours of 10 to 4. I'm surprised at how often I'm still asked to join meetings at other times.

Perhaps worth noting that for medical research with patients as research participants, we don't insist on, or at least encourage, things like protocols and statistical analysis plans for "philosophy of science reasons".

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/...

Asking us to get rid of "personal" items from offices is all very well, but where am I going to put my giant Wordle?

Core GRADE 1: overview of the Core GRADE approach This first article in a seven part series presents an overview of the essential elements of the GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach that has proved extre...

Researchers grappling to make sense of the body of GRADE guidance (35 papers and growing) may be next-level-bewildered or relieved (not sure which) by a new series of Core GRADE in BMJ, which re-starts at 1. I think useful, but also hoping they can stop at the promised 7. www.bmj.com/content/389/...

George is bewildered as we repeatedly sing the words "Why's. My. Cat. So. Stinkeee-ee-ee-ee" to the tune of the UK Eurovision entry. Who knew we would sing so many songs to him?

Just helped this beautiful rose chafer beetle out of my house and let in a bluebottle. One in, one out, I guess.