The aICU research initiative: Developing next-generation decision support tools for intensive care
aICU is a clinically embedded research initiative at Karolinska Institutet and Södersjukhuset, focused on developing and evaluating AI-based decision support for intensive care. We aim to create robust, clinically relevant tools and to study how they can be tested, integrated and used safely in real-world healthcare settings.
About the aICU programme
The programme brings together ICU clinicians, AI researchers, large-scale health data, secure research infrastructure and human-centred evaluation. The aim is to develop decision-support tools that are clinically useful, technically robust and prospectively evaluated before any use in patient care.
The work is led from Karolinska Institutet and Södersjukhuset, in close partnership with the ML@CL research group at the University of Cambridge, together with additional national and international collaborators.
Research focus
Our current work focuses on AI-based decision support for complex ICU decisions, starting with mechanical ventilation and extubation. We study how models can help predict duration of mechanical ventilation and identify patients at risk of extubation failure, while accounting for evolving respiratory function over time. Broader areas include modelling organ-specific patient status and clinical trajectories, cardiac arrest-related trajectories and prognostication, longitudinal modelling of treatment response, and methods for prospective ICU model evaluation.