Reposted by Maarten van SmedenAliëtte Jonkers @aliettejonkers.nl · 09/07/2026Lang gewerkt aan de reconstructie van dit verhaal. Zaterdag in NRC, vandaag al online: Heeft de geit het weer gedaan? Het bewijs is broodmager dat omwonenden longontstekingen krijgen door de geitenhouderij Cadeaulink 🎁 www.nrc.nl/nieuws/2026/...nrc.nlHeeft de geit het weer gedaan? Het bewijs is broodmager dat omwonenden longontstekingen krijgen door de geitenhouderijVolksgezondheid: Zijn geiten opnieuw een gezondheidsrisico, na de eerdere uitbraak van Q-koorts? Zorgen ze voor longontstekingen? Onderzoek dat dat zou aantonen, wordt „volledig mislukt” genoemd. 11911
Maarten van Smeden @maartenvsmeden.bsky.social · 02/07/2026Normality : a very rare shape for data www.linkedin.com/posts/maarte...linkedin.comStatistical terms: what they really mean Multicolinearity— these variables all look the same Heteroscedasticity— the variation varies Attenuation— being too modest Overfitting— too good to be… | Maar...Statistical terms: what they really mean Multicolinearity— these variables all look the same Heteroscedasticity— the variation varies Attenuation— being too modest Overfitting— too good to be true Co... 1123
Maarten van Smeden @maartenvsmeden.bsky.social · 15/05/2026Came across this gem again today. Still makes me giggle @richarddriley.bsky.social @gscollins.bsky.social Yes, it is published: pubmed.ncbi.nlm.nih.gov/32817260/ 4303
Reposted by Maarten van SmedenDarren Dahly @statsepi.bsky.social · 28/04/2026🎯So many important points packed into this editorial. ht @maartenvsmeden.bsky.social www.nature.com/articles/s41...nature.comShow us the evidence for the value of medical AI - Nature MedicineClaims that medical AI is improving care must be backed by appropriate evidence. 1207
Reposted by Maarten van SmedenEslam Maher @eslammaher.bsky.social · 25/03/2026Appraising ML prediction models @maartenvsmeden.bsky.social 1) Share model and code 2) Predictors and language 3) Class imbalance & calibration 4) Model validation sample size 5) Is the model clinically useful in practice? www.jclinepi.com/article/S089...jclinepi.comPreliminary appraisal of machine learning based prediction modelsA significant portion of healthcare research is devoted to the development of prediction models, yet the integration of these models into routine clinical care remains limited. Persistent barriers, su... 072
Reposted by Maarten van SmedenFrank Harrell @f2harrell.bsky.social · 17/03/2026Fantastic new paper casting doubt on explainability of explainable AI. To explain complex machine learning algorithms you need reproducibility of the explanation at a minimum academic.oup.com/ehjdh/articl... #machinelearning #Statistics #StatsSky @maartenvsmeden.bsky.socialacademic.oup.comSignal or noise? Evaluating commonly used attribution methods for explaining deep neural networks in electrocardiogram classificationAbstractAims. Attribution-based explainability methods are widely used in electrocardiogram (ECG) analysis to interpret predictions from ‘black-box’ deep n 13815
Reposted by Maarten van SmedenBen Van Calster @benvancalster.bsky.social · 10/03/2026Happy to see this in print! doi 10.1146/annurev-statistics-042324-123749 @maartenvsmeden.bsky.social @laurewynants.bsky.social @vanamsterdam.bsky.social and Ewout Steyerberg 0172
Maarten van Smeden @maartenvsmeden.bsky.social · 06/03/2026I have got the data and I figured out exactly how to do my cluster analysis all I need is a relevant question that my cluster analysis is going to answer 4362
Reposted by Maarten van SmedenDonald Szlosek @dszlosek.bsky.social · 03/03/2026What if you combine open datasets with AI? Apparently, a 3 fold increase in low quality research papers, mass-produced by paper mills. Interesting study in @jclinepi.bsky.social #academicsky #episky #medsky #Skystats Thanks to @maartenvsmeden.bsky.social for initially posting this on Linkedin! 195
Reposted by Maarten van SmedenBen Van Calster @benvancalster.bsky.social · 13/12/2025Our guidance regarding performance measures for medical AI models is finally out! - Stop bashing AUROC, although it does not settle things - Calibration and clinical utility are key - Show risk distributions - Classification statistics (e.g. F1) are improper www.thelancet.com/journals/lan...thelancet.comEvaluation of performance measures in predictive artificial intelligence models to support medical decisions: overview and guidanceNumerous measures have been proposed to illustrate the performance of predictive artificial intelligence (AI) models. Selecting appropriate performance measures is essential for predictive AI models i... 25225
Maarten van Smeden @maartenvsmeden.bsky.social · 05/12/2025NEW PAPER The use of explainable AI in healthcare evaluated using the well known Explain, Predict and Describe taxonomy by Galit Shmueli link.springer.com/article/10.1... 1264
Reposted by Maarten van SmedenCameron Patrick @cameronpat.bsky.social · 25/11/2025this is one of my favourite observations about sample size calculations. (afaik first articulated by Miettinen in 1985) 17721
Maarten van Smeden @maartenvsmeden.bsky.social · 25/11/2025For some research studies the optimal sample size should be estimated at 0 1636
Maarten van Smeden @maartenvsmeden.bsky.social · 18/11/2025“Data available upon reasonable request” is academic language for you can get my data OVER MY DEAD BODY 2383
Maarten van Smeden @maartenvsmeden.bsky.social · 14/11/2025Manuscript_Final_Version_actualFINALcopy_version9b_USETHISONE.docx 6460
Maarten van Smeden @maartenvsmeden.bsky.social · 08/10/2025Kind reminder: data driven variable selection (e.g. forward/stepwise/univariable screening) makes things *worse* for most analytical goals 23410
Maarten van Smeden @maartenvsmeden.bsky.social · 25/09/2025NEW FULLY FUNDED PHD POSITION Looking for a motivated PhD candidate to join our team. Together with Danya Muilwijk, Jeffrey Beekman and I, you will explore opportunities and limitations of AI in the context of organoids For more info and for applying 👉 www.careersatumcutrecht.com/vacancies/sc...careersatumcutrecht.comVacancy — PhD position on AI methodology for prediction of patient outcomes using organoid modelsAre you passionate about bringing personalized medicine to the next level and make real impact in healthcare? Join our team and develop novel AI methodology to improve predictions of relevant patient ... 198
Reposted by Maarten van SmedenDarren Dahly @statsepi.bsky.social · 22/09/2024Interpretable "AI" is just a distraction from safe and useful "AI" 1102
Maarten van Smeden @maartenvsmeden.bsky.social · 13/08/2025I wonder who those people are who come here dying to know what GenAI has done with some prompt you put in 151
Maarten van Smeden @maartenvsmeden.bsky.social · 12/08/2025If you think AI is cool, wait until you learn about regression analysis 512521
Maarten van Smeden @maartenvsmeden.bsky.social · 11/08/2025NEW PREPRINT Explainable AI refers to an extremely popular group of approaches that aim to open "black box" AI models. But what can we see when we open the black AI box? We use Galit Shmueli's framework (to describe, predict or explain) to evaluate arxiv.org/abs/2508.05753 66919
Maarten van Smeden @maartenvsmeden.bsky.social · 31/07/2025The healthcare literature is filled with "risk factors". This word combination makes research findings sound important by implying causality, while avoiding direct claims of having identified causal associations that are easily critiqued. 2241
Maarten van Smeden @maartenvsmeden.bsky.social · 23/07/2025When forced to make a choice, my choice will be logistic regression model over linear probability model 103% of the time 0352
Reposted by Maarten van SmedenTim Morris @timpmorris.bsky.social · 23/07/2025Post just up: Is multiple imputation making up information? tldr: no. Includes a cheeky simulation study to demonstrate the point. open.substack.com/pub/tpmorris... 33910
Reposted by Maarten van SmedenDarren Dahly @statsepi.bsky.social · 14/07/2025You can have all the omni-omics data in the world and the bestest algorithms, but eventually a predicted probability is produced & it should be evaluated using well-established methods, and correctly implemented in the context of medical decision making. statsepi.substack.com/i/140315566/... 44517
Maarten van Smeden @maartenvsmeden.bsky.social · 10/07/2025Depending which methods guru you ask every analytical task is “essentially” a missing data problem, a causal inference problem, a Bayesian problem, a regression problem or a machine learning problem 5586
Reposted by Maarten van SmedenRichard Riley (R²) @richarddriley.bsky.social · 27/06/2025* New preprint led by Joao Matos & @gscollins.bsky.social "Critical Appraisal of Fairness Metrics in Clinical Predictive AI" - Important, rapidly growing area - But confusion exists - 62 fairness metrics identified so far - Better standards & metrics needed for healthcare arxiv.org/abs/2506.17035 0105
Maarten van Smeden @maartenvsmeden.bsky.social · 27/06/2025Surprisingly common thing: comparisons of prediction models developed using, say, Logistic Regression, Random Forest and XGBoost with conclusion XGBoost is "good" because it yields slightly higher AUC than LR or RF using the same data Fact that "better" doesn't always mean "good" seems lost 2110
Reposted by Maarten van SmedenGeorg Heinze @georgheinze.bsky.social · 26/06/2025Published: the paper 'On the uses and abuses of Regression Models: a Call for Reform of Statistical Practice and Teaching' by John Carlin and Margarita Moreno-Betancur in the latest issue of Statistics in Medicine onlinelibrary.wiley.com/doi/10.1002/... (1/8)onlinelibrary.wiley.com 34818
Maarten van Smeden @maartenvsmeden.bsky.social · 17/06/2025What is common knowledge in your field, but shocks outsiders? Validated does not mean it works as intended. It means someone has evaluated it (and may have concluded it doesn’t work at all) 1236
Reposted by Maarten van SmedenRichard Riley (R²) @richarddriley.bsky.social · 02/06/2025**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/yrje52fnsciencedirect.comImportance of sample size on the quality and utility of AI-based prediction models for healthcareRigorous study design and analytical standards are required to generate reliable findings in healthcare from artificial intelligence (AI) research. On… 23415
Reposted by Maarten van SmedenEric Smith @ericsmithrocks.bsky.social · 31/05/2025People always ask me, “how do I know my manuscript is done?” There’s only one way, my friends. If your file name looks something like this: Manuscript - Final Draft 3.7 FINAL FINAL - FINAL (5).docx Then, and only then, is it time. 3063796
Maarten van Smeden @maartenvsmeden.bsky.social · 27/05/2025Re-proposing the Occam's taser: an automatic electric shock for anyone riding the AI hype train making their models unnecessarily complex 0132
Maarten van Smeden @maartenvsmeden.bsky.social · 27/05/2025Rule of thumb: If your model requires data to look like this (balanced after SMOTE), then maybe you want to use a different model 4193
Maarten van Smeden @maartenvsmeden.bsky.social · 19/05/2025We should really ban all use of AI from all our education because the use of AI will make our students dumb, and banning innovation has worked really well before 2253
Maarten van Smeden @maartenvsmeden.bsky.social · 19/05/2025As scientists it is difficult to stay in love with the questions once you fall in love with the answers 0529
Reposted by Maarten van SmedenTim Morris @timpmorris.bsky.social · 14/05/2025So glad they let us use this turn-of-phrase @maartenvsmeden.bsky.social doi.org/10.1016/j.jc... 191
Maarten van Smeden @maartenvsmeden.bsky.social · 12/05/2025So.... about using large language models (e.g. chatGPT) for writing motivation letters for a job I get it! And honestly, use all the technology you need to write the best letter you can But after reading dozens of letters with almost EXACTLY the same intro paragraph I do get a bit tired of it 6492
Reposted by Maarten van SmedenNicole Erler @nerler.bsky.social · 11/05/2025Only 5 days left to apply for our 3 PhD positions at the @umcutrecht.bsky.social 🎓 🏃🏃♀️ 031
Reposted by Maarten van SmedenTim Morris @timpmorris.bsky.social · 06/05/2025‘Wow, this treatment was no better than placebo – because the placebo was so effective!’ It’s such a shame the term ‘placebo effect’ is so widely known and used. There are fundamental challenges with identifying it and most of the claims don’t even try. It’s just change-from-baseline every time. 5408
Maarten van Smeden @maartenvsmeden.bsky.social · 01/05/2025Before citing a paper the least you can do is thoroughly read… 1st year student: the entire literature on the subject Grad student: the paper you cite Post-grad: title and abstract and scan the rest of the paper Professor: the title. PLEASE read the title 7559
Reposted by Maarten van SmedenGary Collins @gscollins.bsky.social · 30/04/2025recommendations on which prediction model performance measures to report and which not too by @benvancalster.bsky.social more discussion here… arxiv.org/abs/2412.10288 #memtab 0164
Reposted by Maarten van SmedenGary Collins @gscollins.bsky.social · 30/04/2025next up is @joieensor.bsky.social talking about his work on software implementation for sample size calculations for targeting precise risk predictions - code available to do this available here github.com/JoieEnsor/pm... #memtab 032