MIMIC-IV / eICU
ICU EHR datasets used for clinical outcome, adverse-event, and PK/PD benchmarks.
Kind
data-platform
Composite
89.4
Benchmarks tracked
14
Direct-linked
0
As of
2026-05-12
Host organisation
MIT Lab for Computational Physiology
Primary contacts
Leo Anthony Celi, Alistair Johnson, Roger Mark
Founded
2016 / 2020 (v4)
License model
PhysioNet credentialed
Official site
GitHub
Count methodology
PhysioNet + BigBio MIMIC-IV benchmarks 2026-05: 14 derived benchmarks (mortality, LOS, readmission, sepsis, AKI, drug dosing, phenotyping).
Rubric
rigor
5
coverage
4
maintenance
5
adoption
5
quality
5
accessibility
3
industry_relevance
4
Breakdown
outcome_prediction
6
drug_dosing
3
adverse_event
3
phenotyping
2
Notes
Canonical for clinical ML. US-centric.