SUPR
Digital Decision Support for Recommending Transport Destinations for Trauma Patients
Dnr:

NAISS 2024/23-432

Type:

NAISS Small Storage

Principal Investigator:

Stefan Candefjord

Affiliation:

Chalmers tekniska högskola

Start Date:

2024-07-02

End Date:

2025-06-01

Primary Classification:

20699: Other Medical Engineering

Allocation

Abstract

This project “Digital Decision Support for Recommending Transport Destinations for Trauma Patients” is part of the IoT Sweden project “ASAP-PoC (point of care)” (https://iotsverige.se/projekt/asap-poc-snabba-vardbeslut-i-kritiska-situationer?lang=sv) and the EU Interreg project “Kontiki – Artificiell intelligens (AI) som beslutsstöd för patienter och hälso- och sjukvård” (https://www.hb.se/forskning/forskningsportal/projekt/kontiki--artificiell-intelligens-ai-som-beslutsstod-for-patienter-och-halso--och-sjukvard/), a continuation of the previous EU Interreg project “Artifisiell intelligens (AI) som beslutningsstøtte – för en jämlik vård” ( http://beslutningsstotte-ai.com/ ) and is done within the Biomedical Engineering and Digital Health research groups at the Department of Electrical Engineering, by project leader Stefan Candefjord and PhD Student Anna Bakidou. The project aims to develop new decision support for the prehospital care based on artificial intelligence (AI), which are expected to create great opportunities for improvement of precision in decisions regarding priority, assessment of the right level of care and management for different patient groups. The model development is based on data from the Swedish Trauma Registry, SweTrau, and the Norwegian Trauma Registry, NTR. The methods are machine learning algorithms and statistical methods, such as Support Vector Machines (SVM), neural networks and logistic regression, implemented in Python, R and potentially Matlab. Several studies show that patients who have been subjected to severe violence have the greatest chance of surviving if they are transported to a university hospital, which has the highest competence and most resources to treat serious injuries. It is critical that these patients are transported directly to a university hospital, which unfortunately is often not done today. The challenge is that since not all injuries are visible, it is not obvious which patients are seriously injured, the patient can e.g. suffer from a life-threatening internal bleeding but appear relatively unharmed. The precision of the assessment is today relatively low and many patients with serious injuries do not receive optimal care. Precise decision support can help save lives and reduce disability. A decision support based on AI has the potential to, by creating a mathematical model based on measurable and observable parameters such as vital data and type of injury, provide a recommendation for transport destination and possibly also more decisions that are important for the patient to be given optimal care.