Aalborg University in Denmark has published a press release about the development of an AI-powered system to identify critical medical cases requiring admission to emergency departments and to avoid admitting other medical cases to reduce pressure on hospitals.
Researchers will use data from ambulances to avoid unnecessary hospital admissions.
Ambulance crews need better tools to assess patients and distinguish between critical cases and those that do not require hospitalization. Therefore, researchers from Aalborg University, in collaboration with the Northern Denmark Region and Treat Systems ApS, are developing a system that continuously assesses a patient's condition inside an ambulance.
The new system aims to give ambulance crews a better basis for decision-making, so that more patients can receive appropriate treatment in the right place, and unnecessary hospital admissions can be avoided.
Currently, many patients are being transported to the hospital by ambulance, even though they could receive treatment elsewhere within the healthcare system. This puts a strain on emergency departments and hospitals, which are already facing shortages of medical staff and beds.
In collaboration with the North Jutland region and Treat Systems ApS, researchers from Aalborg University have developed a prototype AI-based system that continuously assesses the patient's condition and is already present in the ambulance.
The system aims to help ambulance crews identify patients who need emergency hospital treatment more quickly, and is designed to be a decision-making tool to support crew assessments, according to project manager and clinical assistant professor at the Aalborg University Prehospital and Emergency Research Centre, Morten Breinholt-Sovsow.
Today, medical staff assess patients using a scale calculated based on several factors, including pulse, blood pressure, and other vital signs. However, this current scale is not accurate in predicting who is at risk of deteriorating. As a result, some patients are hospitalized when a quick assessment by their own physician would suffice. Our system appears capable of reducing the number of patients misdiagnosed as critically ill by almost half.
The rating may become outdated quickly.
In addition to requiring manual intervention from medical staff, modern patient assessments are often conducted only once or twice. Therefore, they provide only a brief overview of the patient's condition, increasing the risk of missing any deterioration in time.
“The key difference is that the new system automatically and continuously assesses the condition every time new measurements are received,” says Morten Brennholt Sofsow, adding that the researchers focused on ensuring that staff could see the basis on which the specific assessment was made: “For example, if a patient’s respiratory rate is high, the system can show that this is exactly what is driving the assessment toward a higher risk level.”
The model can support patient safety
This is a prototype currently being trained using ambulance data from North Jutland. The next step is to work towards obtaining CE marking and testing the system in ambulances, possibly through a randomized clinical trial.
According to the researchers leading the project, this approach is very promising, as emergency services outside of hospitals are among the areas where shared systems and vast amounts of data are available nationwide. This would facilitate scaling up the solution if it proves effective.
In the pre-hospital service, which includes the ambulance service in North Jutland, they also see great potential in the project, especially with the increasing number of elderly patients and the rise in complex diseases, which puts additional pressure on emergency preparedness.
Martin Rosgaard-Knudsen, Medical Director of the Pre-Hospital Service, explains: “Time is a critical factor in our work, so if artificial intelligence can help our staff improve their assessments of patients, it will ultimately contribute to enhancing patient safety and will also help us ensure that they receive the right treatment in the right place. We are always open to new technologies and artificial intelligence as tools that complement our skilled professional staff.”.
Facts about ambulance services in Denmark (according to the press release and Ritzau):
In December 2024, Prehospital Business issued a statement entitled
“Analysis of the pre-hospital area in North Jutland: Reports indicate that the North Jutland Emergency Management Authority (AMK) emergency center receives approximately 47,000 calls annually on the 112 number related to acute illnesses or potentially life-threatening accidents.
The analysis shows that the number of incoming calls to the 112 number that are not related to acute or serious cases exceeds this number.
According to the analysis, North Jutland saw an increase of 27% in the number of incoming calls via the 112 number during the period 2017-2023. The region also saw an increase in the number of pre-hospital medical trips, especially trips made via the 112 number, with an increase of 22% during the same period.
