Skip to main content
conference lunch move company map contacts lindholmen lindholmen 2 travel info


AI will find more patients with sepsis while still in the ambulance

The PreSISe project, being run by Prehospital ICT Arena at Lindholmen Science Park, is one of seven projects that has been granted funding in Vinnova’s “Artificial Intelligence for Better Health” call for proposals. Vinnova has awarded the project SEK 3.7 million in funding.
Photo: Unsplash

With a total project budget of approximately SEK 5.5 million spread over two years, the project will enable an increase in the precision of prehospital assessment of sepsis patients. Vinnova is contributing SEK 3.7 million, and the rest will be funded by the parties to the project. Seven parties are taking part in the project in addition to PICTA: Chalmers, the University of Borås, Aweria AB, InterSystems Sweden AB, Karolinska Institute, MedITeQ AB and Skaraborg Hospital.

Johanna Bergman, Project Manager for PreSISe (Prehospital Support for Identifying Sepsis Risk), says that this exact partner constellation and the long-term national effort are just some of the project’s strengths, and continues:

“We have carried out thorough preparatory work that led to the application, and while waiting for a decision we continued pursuing activities that will support the project moving forward, such as a recently conducted workshop around utilizing AI in health care. PreSISe is an important project in many ways. Sepsis is a relatively unknown condition that claims many lives every year — there is always potential here to make a major difference. The project will be a guide for how AI projects can be run over the long term to benefit to health care as regards access to data, regulatory issues, visualization and above all collaboration among different players. It is fantastically gratifying to be able to set the project going in earnest now!”

Sepsis is a serious condition in which the body’s immune defense attacks the body’s own tissues in order to defend itself against invading bacteria. Up to 40,000 people in Sweden are afflicted each year, and mortality is around 20%. Early identification and rapidly initiated treatment has proven to be crucial to the outcome. Identifying possible sepsis while still in the ambulance has the potential to shorten the time until antibiotic treatment once at the hospital. Checklists and mnemonics based on vital parameters are currently used in order to identify the risk of sepsis in prehospital treatment, which has been criticized for being too unspecific. In PreSISe, artificial intelligence (AI) is used to produce decision-making support that can predict the risk that a patient is suffering from sepsis. This support will work alongside existing information systems to automatically notify ambulance staff if there is an increased risk of sepsis. In this initial project phase, data from previous sepsis studies will be used to produce a prototype that will be tested in a simulated environment.

Professor Bengt Arne Sjöqvist of Chalmers, Program Manager of PICTA at Lindholmen Science Park, explains why PreSISe is such an important project:

“PreSISe is an excellent example of the type of project we would like to conduct in PICTA, and is also difficult for any other party to design and manage. The reason for this is that PICTA is a neutral arena for collaboration with a focus on innovation and producing useful applications, and is currently a well-known and established player in the field of prehospital care with a broad network in medical care, academia and the private sector. In short, PICTA has the knowledge and the support needed to carry out this project. In addition, the project will whet the appetite for the national AI venture that will be taking place here at Lindholmen.”

The project will begin immediately and run through June 2020. Information on the project will be presented on a regular basis on the PICTA website.

Read more about PreSISe here

Read more about the AI venture

Read about the projects that received funding in the call for proposals here