NAISS
SUPR
NAISS Projects
SUPR
Implementing AI for sequential decision-making in Intensive Care: Challenges illustrated by intracranial pressure prediction
Dnr:

NAISS 2026/4-290

Type:

NAISS Small

Principal Investigator:

Peter Galos

Affiliation:

Uppsala universitet

Start Date:

2026-03-10

End Date:

2027-04-01

Primary Classification:

30201: Anesthesiology and Intensive Care

Webpage:

Allocation

Abstract

This project investigates challenges in implementing AI-based sequential decision-making in intensive care, using intracranial pressure prediction as a case study. A machine-learning model generates minute-by-minute risk scores within simulated patient trajectories and is compared with clinical interventions to assess alert burden, causal relevance, and practical usability. The study aims to support the translation of predictive models toward clinically deployable decision-support systems.