The 2024 Pediatric Sepsis Data Challenge has officially concluded, wrapping up a months-long global effort to improve early prediction of sepsis-related deaths in children.
Launched by the Pediatric Sepsis Data CoLaboratory (Sepsis CoLab) in partnership with the Institute for Global Health at BC Children’s Hospital and BC Women’s Hospitals + Health Centre, the World Federation of Pediatric Intensive and Critical Care Societies (WFPICCS), and Duke University, the challenge brought together students, clinicians, and data scientists from around the world.
Sixteen finalist teams from eleven countries came together on an open-source platform to collaborate, exchange ideas, and develop algorithms to address one of the most urgent challenges in pediatric healthcare: sepsis. Each team was tasked with building a model to predict in-hospital mortality in children with suspected sepsis. A defining feature of the challenge was its emphasis on collaboration – many participants met for the first time through the challenge platform, forming new, multidisciplinary teams that spanned countries, specialties, and time zones.
Using a synthetically generated dataset based on real patient data, teams developed predictive models using clinical, social, and demographic data collected at hospital admission. These models were then tested on unseen real-world data to see how accurately they could identify high-risk children early. By using synthetic data and open-source principles, the challenge lowered barriers to participation while protecting patient privacy.
After months of competition, the top three performing teams are:
- Team AIMS - a multinational collaboration from Cameroon, Uganda, Morocco, and Canada
- Team Epiphany - United States
- UCL Paediatrics - United Kingdom
Each team brought a unique approach to the challenge, combining technical innovation with clinical insight. Their work highlights how global collaboration can drive progress in tackling complex health problems, especially in low-resource settings.
Sepsis remains a leading cause of death in children under five, particularly in low- and middle-income countries. In 2017, children made up over half of the estimated 48.9 million global sepsis cases. Early identification is critical, and the tools developed through this challenge offer promising new strategies to support that goal.
With the final leaderboard and team profiles now released (more team profiles added daily!), the broader research and clinical communities are invited to explore the methods behind the models and consider how they can be translated into real-world tools to improve care and save lives.