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

Informations projet

Digi-NewB

N° de convention de subvention: 689260

DOI

10.3030/689260

Projet clôturé

Date de signature de la CE 11 Decembre 2015

Date de début 1 Mars 2016

Date de fin 31 Mai 2020

Financé au titre de

• SOCIETAL CHALLENGES - Health, demographic change and well-being

Coût total

€ 4 457 680,25

Contribution de l’UE

€ 4 457 680,25

Coordonné par

GCS HOPITAUX UNIVERSITAIRES GRAND OUEST
France

Periodic Reporting for period 3 - Digi-NewB (Non-invasive monitoring of perinatal health through multiparametric digital representation of clinically relevant functions for improving clinical intervention in neonatal units (Digi-NewB))

Période du rapport: 2019-03-01 au 2020-02-29

Résumé du contexte et des objectifs généraux du projet

In the EU around 8% of newborns are preterm births and around 6.5 % have a birth weight of less than 2500g (Europeristat data) meaning that more than 300 000 infants are hospitalised each year in a European neonatal care unit. Prematurity is a developmental period critical to later neurobehavioural function. During this period, the risk of sepsis is high with a high risk of mortality and many environmental exposures have been shown to impact long-term neurobehavioural outcomes.
Started in 2016, Digi-NewB is a an EU-funded project that aimed to improve care for the newborns in neonatal units by reducing mortality, morbidity and health costs with a better monitoring of sepsis and newborn’s development. Using the clinical and signal data (cardiac and respiratory traces, video quantification of movement and sound) from a large cohort of hospitalised newborns recordings, Digi-NewB proposed a new decision support system (DSS). This DSS aims to assist the clinician in his/her decision-making through non-invasive monitoring of sepsis risk and of cardio-respiratory and neurobehavioral maturations.
Sub-objectives included i) the implementation of a large clinical study that consisted in monitoring more than 700 newborns in 6 different hospitals; this includes the design and setting up of a dedicated data acquisition system acquiring physiological signals and video and sound data ii) the identification of relevant parameters and related composite indices to be displayed on a dedicated interface in neonatology units, and iii) the improvement and validation of the Decision support system with regards to diagnosis relevance and to user experience.
To achieve these goals, a consortium has been constituted, under the coordination of a paediatric clinical network (GCS HUGO-France), with two SME (Voxygen-France and Syncrophi-Ireland) and four university groups with multidisciplinary expertise: in signal processing (INESC-Portugal and LTSI-UR1 France), in multivariate analyses (TUT-Finland) and in user centred design approach (NUIG-Ireland). The clinical study was implemented in Rennes, Angers, Nantes, Brest, Tours, Poitiers. The proposed new class of monitoring aim to ensure novel preventive and therapeutic strategies to counteract late diagnosis of sepsis and inappropriate evaluation of maturity.

Travail effectué depuis le début du projet jusqu’à la fin de la période considérée dans le rapport et principaux résultats atteints jusqu’à présent

The Digi-NewB project achieved a novel and non-invasive sepsis warning score for neonates, supporting neonatologists in taking decisions based on early alerts from several hours to several days before clinical observation. This decision-support system is running real-time and is based on continuous monitoring of physiological vital signs. It also includes a running video monitoring system, and an option enabling the visualisation of the newborn maturation during hospital stay.

The Digi-NewB project also achieved the following results:
- Database including 781 newborns in the Clinical study, which equals to 7 years of data recordings in total (+2200 recordings) that is stored into a centralised database. Data includes vital signs monitoring, video/sound and clinical records data
- Data analysis results leading to the development of novel features based upon graph theory for Heart rate variability analysis, novel algorithms for the detection/characterisation of cardiorespiratory events, movement, cries, and other features relevant for the early detection of sepsis and quantification of maturation
- Achievement of an integrated maturation graph enabling real time access to quantified indicators regarding the newborn maturation process: sleep, movement, cries, cardio-respiratory events.
- Development of a Sepsis AI-based algorithm based on deep learning approaches and software integration into the full decision support system
- Development and testing in hospital setting and on 4 newborns of the real-time sepsis early detection system and of video and sound acquisition
The project also led to 10 international papers in peer-reviewed journals, 5 softwares were licensed to software protection authorities and 2 patents. Two new patents are currently in patentability studies.

The system is being currently tested in Rennes University Hospital and will aim to clinically validated in a large randomized controlled trial. Contacts with industrial players and further research pathways are already established.

Progrès au-delà de l’état des connaissances et impact potentiel prévu (y compris l’impact socio-économique et les conséquences sociétales plus larges du projet jusqu’à présent)

The design of a real-time newborn sepsis monitoring device based on a mathematically driven Decision support system using composite indices extracted from video, sound and physiological signal data is unique in the world. The DSS aims to detect sepsis risk earlier than current monitoring devices by proposing predictions, and provides valuable insights into cardio-respiratory and sleep maturation of newborns. It follows a non-invasive/contactless approach, an approach that has never been undertaken in paediatrics.
To achieve this objective, a number of scientific advances were achieved during the project duration:
- Discovery of a new marker for sepsis and maturation for premature newborns
- Early detection of sepsis in premature newborns through a unique AI-based method
- Integrative multi-scale multi-level approach with sepsis and maturation indices including clinical evolution, cardiac and respiratory activities, video and sound analysis, and sleep monitoring
- Improved knowledge on clinical risks monitoring and related treatment decisions, with predictive value
- Integration of user-friendly and robust DSS including software and hardware components
- Implementation of a user-centered approach to achieve a high-quality DSS implementation in neonatology services
At the end of the project, the fully integrated Digi-NewB system was tested on long recording periods for clinical use and further industrial cooperation will set the basis to be tested in a European randomized clinical trial and prepare for commercial developments.
Digi-NewB brought further scientific progress on accurate evolution of perinatal health and to better knowledge of sepsis dynamics, and improve the management of such clinical complex situatios. With the RCT, it is expected to measure expected reduced health costs by better health of premature newborns and more limited impacts related to antibiotics use.
On the longer term and to be proved by the RCT, Digi-NewB’s ambition is to reduce mortality and morbidity in neonatology services by an improved detection and treatment of sepsis, and maturation monitoring. It will aim to be used by a large number of European neonatology services, and evolve with the medical knowledge and user needs.

Digi-NewB logo