Algorithms and IT – new weapon against HAI
Healthcare-associated infections are the source of great suffering and major costs every year – 750,000 extra hospital days and SEK 6.5 billion. And 1,500 people die. In Sweden alone. This is why a nationwide initiative is underway to find more innovative approaches to prevention and treatment.
How can algorithms reduce healthcare acquired infections?
Currently it is very difficult to know when someone is affected. Therefore patient records are analyzed using AI. We aim at better understanding how and when infections may be prevented – and new AI-tools are developed to support healthcare staff.
"VRI Proaktiv" (HAI Proactive) is a nationwide innovation and research project, in which a variety of participants have joined forces to develop new, more efficient approaches to prevent healthcare-associated infections (HAI). Another objective is to improve treatment for those who nevertheless become infected.
We aim at better understanding how and when the infections may be prevented
"We want to be able to identify patients who are at risk of contracting healthcare-associated infections, but also to find those patients who do have an infection at an early stage and follow up their treatment," says Dr. Pontus Nauclér, an infectious disease specialist who is one of the driving forces in the project.
New IT support guides staff
The growing quantity of information from different sources provides greater opportunities for proactive decisions, but also increases staff needs for support in sorting through the maze of data.
For this reason the IT support project was developed to flag high-risk patients in the computerized medical records systems and indicate what procedures reduce the risk of healthcare-associated infections. The decision-support program includes both known risk factors and identification of new risk factors and effective measures.
It will also result in a better overview for care providers.
AI analysis of patient charts
"It is currently difficult to measure exactly how many patients contract healthcare-associated infections – even though the care process is carefully documented in patient charts," says Hercules Dalianis, Professor of Computer and Systems Sciences at Stockholm University.
For this reason, the project is now analyzing a large number of charts using artificial intelligence (AI), to teach the computer system which patients are affected.
"Automatic reporting to the hospital administration and national bodies can show the number of infections at specific clinical departments, a specific hospital, or throughout the country. Whether the number of infections increase or decrease can be linked to specific causes."
Prediction models developed
VRI Proaktiv has calculated that the Swedish healthcare system could free up a large number of hospital beds and billions of Swedish kronor annually by developing and improving the right areas. Today 75 percent of improvement projects focus on emergency medical services, which only account for 25 percent of total care. By developing predictive methods and preventive measures, an estimated one third of all healthcare-associated infections could be prevented.
Phone: +46 (0)765696718
HAI Proactive (VRI Proaktiv in Swedish) is a nationwide initiative that is coordinated by Karolinska University Hospital and aims to develop new, more effective methods to prevent healthcare-associated infections and to improve treatment of HAI.
The project includes development of IT support to identify high-risk patients and to improve patient care for those affected.
Project partners: Swedish Association of Local Authorities and Regions (SALAR), Karolinska University Hospital, Region Östergötland, Västerbotten County Council, Karolinska Institutet, Stockholm University, Tieto, SAS Institute, Treat Systems and Vinnova.
Hospital services involved: Inflammation and Infection Theme, as well as Quality and Patient Safety.
Coordinator at the Innovation Center: Per S. Englund.