SJD Barcelona Children’s Hospital has published a new article highlighting the development of an AI-based algorithm for the early prediction of sepsis in critically ill children, as part of the PHEMS project.
The article showcases the work being carried out by professional from SJD Barcelona’s Paediatric Intensive Care Unit, Department of Innovation and Data Management Department. The team has developed an artificial intelligence-based algorithm capable of predicting the onset of sepsis up to six hours before clinical symptoms appear.
The development is one of the clinical use cases within PHEMS, demonstrating how paediatric health data can be used to developed innovative solutions to support earlier diagnosis and clinical decision-making.
To develop the predictive tool, researchers retrospectively analysed 9,975 paediatric intensive care admissions involving 6,984 patients treated between 2019 and 2025. The analysis identified around 60 clinical variables that show different patterns in patients who subsequently develop sepsis, providing potential early indicators for prediction.
The algorithm is currently undergoing prospective evaluation in patients admitted to the Paediatric Intensive Care Unit, with the aim of integrating it into existing clinical screens, apps and alert systems.
From a local use case to European collaboration
The sepsis prediction use case illustrates the type of real-world clinical challenge that PHEMS is addressing: how to unlock the value of paediatric health data while maintaining privacy and security.
The work at SJD Barcelona Children’s Hospital is an important exampe of how these approaches can contribute to more precise, predictive and personalised paediatric medicine.
Read the full article here.