Medical Statistician at KU Leuven. My brain is like a snail but it gets there in the end (or not).https://bsky.app/profile/benvancalster.bsky.socialhttps://bsky.app/profile/benvancalster.bsky.social/post/3mgoxwhlyis2hHappy to see this in print! doi 10.1146/annurev-statistics-042324-123749 @maartenvsmeden.bsky.social @laurewynants.bsky.social @vanamsterdam.bsky.social and Ewout Steyerberg10 Mar 2026 08:40 +0000at://did:plc:xeo3ghgqq5pfe4eixb2c5hlq/app.bsky.feed.post/3mgoxwhlyis2hhttps://bsky.app/profile/benvancalster.bsky.social/post/3m7urgs2zbs2cOur 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 https://www.thelancet.com/journals/landig/article/PIIS2589-7500(25)00098-6/fulltext13 Dec 2025 14:03 +0000at://did:plc:xeo3ghgqq5pfe4eixb2c5hlq/app.bsky.feed.post/3m7urgs2zbs2chttps://bsky.app/profile/benvancalster.bsky.social/post/3lxjn7opu5s2fIn our latest work, we show that risk estimates for patients are HUGELY uncertain due to model, data, and population uncertainty. Even for well performing models (c statistic, calibration, utility) based on large N. @laure_wynants @ESteyerberg @lasaibarrenada.bsky.social arxiv.org/abs/2506.17141 https://arxiv.org/abs/2506.1714129 Aug 2025 08:27 +0000at://did:plc:xeo3ghgqq5pfe4eixb2c5hlq/app.bsky.feed.post/3lxjn7opu5s2fhttps://bsky.app/profile/benvancalster.bsky.social/post/3lxcbuhfub226Universities love open science Small print: unless money is involved26 Aug 2025 10:15 +0000at://did:plc:xeo3ghgqq5pfe4eixb2c5hlq/app.bsky.feed.post/3lxcbuhfub226https://bsky.app/profile/benvancalster.bsky.social/post/3lovu73zaos2rI recently had a paper in a journal where the fee for open access was lower than the fee for closed access. Is that common? @grahamkendall.bsky.social11 May 2025 16:17 +0000at://did:plc:xeo3ghgqq5pfe4eixb2c5hlq/app.bsky.feed.post/3lovu73zaos2rhttps://bsky.app/profile/benvancalster.bsky.social/post/3ll4lr5fdyc2pPROBAST+AI is out! https://www.bmj.com/content/388/bmj-2024-08250524 Mar 2025 10:54 +0000at://did:plc:xeo3ghgqq5pfe4eixb2c5hlq/app.bsky.feed.post/3ll4lr5fdyc2phttps://bsky.app/profile/benvancalster.bsky.social/post/3lkigclwlnc27Quousque tandem? @kuleuvenuniversity.bsky.social @fwovlaanderen.bsky.social A lot of publications in MDPI journals also in Flanders, despite it being listed on predatoryjournals.org. [contains quote post or other embedded content]16 Mar 2025 10:23 +0000at://did:plc:xeo3ghgqq5pfe4eixb2c5hlq/app.bsky.feed.post/3lkigclwlnc27https://bsky.app/profile/benvancalster.bsky.social/post/3lk66tw6wfs2mWe tried to look at ways to obtain flexible calibration plots in clustered (e.g. multicenter) validation studies. Work with @lasaibarrenada.bsky.social @laurewynants.bsky.social @bavodccampo.bsky.social arxiv.org/abs/2503.08389 https://arxiv.org/abs/2503.0838912 Mar 2025 08:43 +0000at://did:plc:xeo3ghgqq5pfe4eixb2c5hlq/app.bsky.feed.post/3lk66tw6wfs2mhttps://bsky.app/profile/benvancalster.bsky.social/post/3ljz43em2kc23It is a good day today, I got the reviewer comment "please consult a statistician" again.10 Mar 2025 08:10 +0000at://did:plc:xeo3ghgqq5pfe4eixb2c5hlq/app.bsky.feed.post/3ljz43em2kc23https://bsky.app/profile/benvancalster.bsky.social/post/3ldfxcum5yk22New work in preprint! "Performance evaluation of predictive AI models to support medical decisions: Overview and guidance". Under the wings of the STRATOS initiative. But @maartenvsmeden.bsky.social said it better already 😜 arxiv.org/abs/2412.10288 http://arxiv.org/abs/2412.1028816 Dec 2024 08:53 +0000at://did:plc:xeo3ghgqq5pfe4eixb2c5hlq/app.bsky.feed.post/3ldfxcum5yk22https://bsky.app/profile/benvancalster.bsky.social/post/3lc3rkzeuns2xIs this something everybody knows already but me? The joy of EHR! https://www.youtube.com/watch?v=xB_tSFJsjsw29 Nov 2024 14:19 +0000at://did:plc:xeo3ghgqq5pfe4eixb2c5hlq/app.bsky.feed.post/3lc3rkzeuns2x