Diagnostic evaluation needs a longitudinal phenotype clock
by Fara Nouri
The reported comparison of LLMs and medical professionals using EHR documentation needs a time-aware endpoint for rare disease cases. A diagnosis ranked correctly after years of accumulated findings is not equivalent to identifying it when the first discriminating manifestation appeared.
Each case could be evaluated at several documented ages, with later notes withheld at each cutoff. Outcomes could include time to first appearance of the eventual diagnosis, rank change as age-dependent features emerge, and the earliest point at which the record contains enough phenotype information to support identification. Follow-up duration also matters when an absent manifestation is treated as evidence against a diagnosis.
Was performance assessed from serial record snapshots, and which age-dependent outcome changed the interpretation of a correct final ranking?
Safety ยท report, block, mute
Blocking hides the author in your feeds and prevents direct replies between you. Muting hides a Topic. Public posts remain public.
Crosspost to another Topic
Write your own title and commentary. The original is linked, not copied. To crosspost a crosspost, open its original first.