News and events - Tampere University of Technology

Data analysis helps prevent sepsis deaths among premature babies

Sepsis, or blood poisoning, is a common cause of death among premature babies. Researchers from TUT are participating in an international project that aims to create a method for predicting potential cases of sepsis. This would enable improved treatment and reduced death rates for premature babies.
Photo: credit photo Thinkstock
Photo: credit photo Thinkstock

FACTS:  Non-invasive monitoring of perinatal health through multiparametric digital representation of clinically relevant functions for improving clinical intervention in neonatal units (Digi-NewB)

  • A Horizon 2020 project that aims to promote the treatment of premature babies by utilizing big data
  • Funding: over 4 million euros
  • Duration: 2016–2020
  • Involved parties: Groupement de Coopération Sanitaire - Hôpitaux Universitaires du Grand Ouest (coordinator, France), Universite de Rennes I (France), Voxygen SAS (France), TUT (Finland), Inesc Porto - Instituto de Engenharia de Sistemas e Computadores do Porto (Portugal), National University of Ireland, Galway, Ireland, Syncrophi Systems Ltd (Ireland)
  • TUT’s role: developing a support system for premature infants’ treatment decisions. The system will be based on machine intelligence and machine vision.

Prematurely born infants are highly sensitive to infections, as their immunity is still underdeveloped. An infection can cause blood poisoning, which can, at worst, prove fatal to premature babies. To enable early care and keep the treatment as gentle as possible, diagnosing sepsis at an early stage is important.

Researchers from TUT’s Department of Signal processing have participated in the ‘Digi-NewB’ Horizon 2020 project launched in April. In the project, a method is being developed for the early detection of blood poisoning of premature babies. The development of the cardiorespiratory and sleep/neurobehavioural maturity of these babies is also being monitored. TUT’s particular expertise and task in this project is to develop a system that is based on machine intelligence and machine vision and supports treatment decisions concerning premature babies.

The pediatric units of seven hospitals from the Bretagne region in France are also involved in the project.

“The blood pressure, heart rate and oxygen saturation of preterm babies are monitored at the pediatric intensive care units of these hospitals with incubator equipment. A new innovation is that the incubators are now also equipped with cameras that enable monitoring the babies’ cry and movement,” project participants Alpo Värri and Hannu Nieminen from the Department of Signal Processing explain.

The measurement data is analysed using state-of-the-art methods. Previous data collected at the hospitals is also available for analysis. Based on this enormous data mass, it is possible to attain a certain statistical probability as to predicting which of the premature babies are susceptible to sepsis.

“Technology can never replace humans as caregivers to preterm infants, but the early indication of problems helps us monitor the babies more closely and start treatment early on if needed,” the researchers point out.

Data analysis also provides more detailed information on the development of premature babies’ neurological maturity. This facilitates the assessment of the need for intensive care and may help the babies be discharged to their parents’ care at an earlier stage than previously.

“This data collection may also have other positive impacts for the preterm infants’ parents, as the incubator camera could allow them to see their child through online connections at home.”

As a result of the four-year project, a prototype will be validated which will serve as a basis for creating an actual medical device for hospital use.

Further information:

Associate Professor Alpo Värri +358 (0) 40 849 0780,

Senior Research Fellow Hannu Nieminen +358 (0) 50 483 5824,

News submitted by: Tiina Leivo
Keywords: science and research