There’s lots of guidance on developing, evaluating, and reporting prediction models, e.g. TRIPOD+AI or the BMJ series by @gscollins.bsky.social, @richarddriley.bsky.social etc. Many key steps, like checking calibration & comparing with existing tools, are missing here. Peer review has failed again.

Show Post

What is needed to show this tool is actually useful? 1) Show that adding the photo to what doctors already know improves predictions 2) Externally evaluate how accurate these predictions are, in the settings it’s meant for 3) Run a trial to show it improves decision making

Show Post

Prof. Daniela Witten from University of Washington @uwbiostat.bsky.social, will deliver the prestigious Distinguished Lecture Series in Statistical Sciences in person in Toronto and virtually via Zoom. Oct 1 & 2, 2026 | Register ➡️ https://bit.ly/4h11452 #Statistics #BioStatistics

Show Post

About once a month, for the last few years, I've been asked for a sample size in a prediction study and have been able to either point to pmsampsize or just run a quick sim myself. This package is such a blessed contribution to #rstats and #statsky @joieensor.bsky.social @richarddriley.bsky.social

Show Post

Show Post