ICUSafeNotes: Safe Actionable Querying of ICU Clinical Notes

TP17

Home > Our Research > Algorithms & AI/ML Models > ICUSafeNotes: Safe Actionable Querying of ICU Clinical Notes
Bairbre McNicholas

Bairbre McNicholas

Lead Translational Researcher

Thematic area:

Algorithms and AI/ML Models

Product:

A large language model platform tailored to ICU environments, capable of summarising unstructured clinical notes, integrating medical device data, and generating accurate, time aligned clinical insights to support critical care decision making.

Market:

Focuses on European ICUs (8,500 units, 100k beds), with an estimated €210M annual market at €2,000/bed/year. Customers include hospitals, ICU data platform vendors, EHR companies, and clinical trial technology providers.

Fit:

Current ICU documentation is fragmented and overwhelming. ICUSafeNotes differentiates itself by combining prompt engineered LLMs with multimodal signal integration and clinical validation. It enhances safety, reduces cognitive load, and improves communication—offering strong potential for licensing or a digital health spinout.

Other Projects