Wouter van Amsterdam Profile picture
MD, PhD-candidate at University Medical Centre Utrecht. BSc. of physics, loves stats, ML, causal inference and music
21 Jul
Better outcome predictions lead to better treatment decisions, right?

Turns out it’s not that simple. In fact, accurate outcome predictions can in some cases lead to WORSE clinical decisions.

Blog with numerical example here: vanamsterdam.github.io/posts/goodpred…
Thread here 👇:
For a certain type of cancer there are two treatments: A and B. From randomized trials it is known that A leads to better overall survival. There is no known variation in treatment effect among patient subgroups.
Treatment A takes longer than B and has more side effects, so it’s only recommended for patients with a >10% chance of surviving 1 year. In current practice, the 1-year survival is estimated using covariates X.
Read 14 tweets