Dozens of AI disease-prediction models were trained on dubious data The models are designed to predict someone’s risk of diabetes or stroke. A few might already have been used on patients.

Some of the models seem to have been used in clinical settings although it’s not clear whether this has led to flawed diagnoses

go.nature.com/3Q7scE0

Dozens of AI disease-prediction models were trained on dubious data The models are designed to predict someone’s risk of diabetes or stroke. A few might already have been used on patients.

Outstanding work by @alexdgibson.bsky.social to uncover the use of highly dubious data sets from @kaggle.com being used in hundreds of research papers and potentially even informing clinical practice. If you're re-using data, take the time to confirm that it's real. www.nature.com/articles/d41....

Scientific datasets are riddled with copy-paste errors Initial results from scanning through Excel files belonging to 600 published scientific papers.

Markus Englund built a tool for identifying repetitive sequences in scientific datasets! Read up here: www.sciencedetective.org/scientific-d...

Evidence of Unreliable Data and Poor Data Provenance in Clinical Prediction Model Research and Clinical Practice Clinical prediction models are often created using large routinely collected datasets. It is essential that prediction models are developed with appropriate data and methods and transparently reported...

Evidence of unreliable data and poor data provenance in published clinical prediction models and clinical practice 📣

My first PhD study is available on MedRxiv: www.medrxiv.org/content/10.6...

Evidence of Unreliable Data and Poor Data Provenance in Clinical Prediction Model Research and Clinical Practice Clinical prediction models are often created using large routinely collected datasets. It is essential that prediction models are developed with appropriate data and methods and transparently reported...

Evidence of unreliable data and poor data provenance in published clinical prediction models and clinical practice 📣

My first PhD study is available on MedRxiv: www.medrxiv.org/content/10.6...

Learning R for Good Research Practices: Part 1 | Alexander Gibson In this multi-part blog series I outline some steps to help start learning R. Part 1 is a background on R, packages and resources and tips on how to get started learning.

New Blog: Learning R for Good Research Practices‼️

Read Part 1: alexdgibson.com/blog/r_1/

This is the first of a multipart series where I go into my experiences learning R, highlighting tips and resources that were essential in my learning.

This would be so interesting! Will there be a recording?

We need more people reading more books! Such an important skill and has done so much for me.

There’s nothing better than a good book, a change in perspective, beliefs or the generation of new ideas.

Last week, #AusHSI PhD student @alexdgibson.bsky.social presented at #AIMOS2025, highlighting a new global issue of unreliable data within #clinicalprediction model research and clinical practice, which has potential to influence patient outcomes and evidence-based decisions. @aimosinc.bsky.social

Genuinely enjoying how much layered wrongness is here.

A kind of beautiful tiered wedding cake of misconceptions.

Just BATHE in the warm rolling folds of daftness.

Speaker and slide

First up: Alexander Gibson @alexdgibson.bsky.social: Poor Data Provenance in Published Clinical Prediction Model Research.
I was seriously concerned about some of the Kaggle Datasets - those on stroke and diabetes raised concerns about data being fake.
#AIMOS2025

A model of good research practice in clinical prediction - AusHSI When it comes to health and medical research, doing the right thing is critical, especially when it impacts patient outcomes. Alexander Gibson's research focuses on identifying statistical and researc...

📣 New Blog 📣

Learn more about my #PhD Research and work @aushsi.bsky.social

www.aushsi.org.au/a-model-of-g...

Ok, time for a short thread about this paper.

My sense over the past six months or so is that chain-of-thought prompting as used in e.g. ChatGPT o.3 improves substantially upon previous systems such as ChatGPT 4.o, at least for certain tasks.

But how revolutionary is it?

🚨Job Alert! 🚨 #AusHSI is seeking a new Research Project Officer to work collaboratively with a team of leading #healthservices researchers and partners, providing support across a range of major research projects to deliver exceptional outcomes.

👉 Find out more and apply today: bit.ly/4k1ba4h

Reading over my brother’s undergraduate assignment.

Criteria has a section: “Is there an appropriate amount of detail in this section so that you could replicate this study?”

I don’t remember learning about replication in my undergrad! Nice to see it getting some air time.

If you're currently working—or have worked in the past 5 years—in consulting or collaborative research as a biostatistician, we’d love to hear from you.

📅 Closes 11th July

This survey has been approved by the QUT Human Research Ethics Committee (approval #9691).

And if I do need to learn something new I just go to the documentation of the function/package. I trust the devs more than an LLM at this point in time.

I feel the time taken to specifically explain what I want the output code to be from an LLM to be, could be spent just writing the code. Maybe I’m not great at prompting?

I’ve basically never use them. Every time I have tried to use LLMs for coding it hasn’t output what I’m after. Nearly always have to make major changes that anecdotally take longer than if I were to write the code from scratch.

International #ResearchIntegrity conference taking place in beautiful Sydney Australia on 16-18 November 2025 🐨 🌏
Great speakers including @elisabethbik.bsky.social @jdwilko.bsky.social @jasonchin.bsky.social
For more info contact
@simongandevia.bsky.social 🧪

Lazy cross post, please help. I have a bad feeling about this.

Cross stitch embroidery of beetles and the words: But I am very poorly today and very stupid and hate everybody and everything. C Darwin to C Lyell 1861

I thought my first post on Bluesky should be something positive and motivational

I’ve not been on Bluesky for a while but there seems to be so many interesting and engaging people! Lots of good #research

Fineee, I’ll add another to my long tbr list

Thanks for the add and kind words :) appreciate it Andrew

Should My PhD Exist? | Alexander Gibson We don’t need more science, we need better science.

Should my PhD Exist?

My latest blog ✍️: alexdgibson.com/blog/good_sc...

“My hope is one day science will be so rigorous, that all focus can be on progressing science forward, not identifying common problems.”

#researchintegrity #metaresesrch #phd

#AusHSI Prof Will Parsonage is one of more than 20 experts from across the world involved in a new @thelancet.bsky.social Commission calling on the medical profession to treat coronary #heartdisease as a lifelong condition, which has the potential to save 8.7 million lives every year: bit.ly/3XEMd5G

PROBAST+AI: an updated quality, risk of bias, and applicability assessment tool for prediction models using regression or artificial intelligence methods The Prediction model Risk Of Bias ASsessment Tool (PROBAST) is used to assess the quality, risk of bias, and applicability of prediction models or algorithms and of prediction model/algorithm studies....

*NEW PAPER*

PROBAST+AI: an updated quality, risk of bias & applicability assessment tool for prediction models using regression or AI methods

PROBAST+AI consists of two distinct parts:
- model development (quality assessment tool)
- model evaluation (risk of bias tool)

www.bmj.com/content/388/...

A great experience!