Risk Management
Innocens BV needed a faster, more reliable way to detect signs of late-onset sepsis in vulnerable newborns because NICU teams were relying on manual interpretation of many data sources and could miss critical changes for hours.
Reported impact
The Innocens platform gave clinicians a path to detect severe sepsis earlier, preserve patient-data locality, and build a reusable AI foundation for broader neonatal decision support.
Innocens BV worked with IBM to turn a decade of neonatal intensive care data into an explainable clinical decision-support platform. Using IBM Watson Studio on IBM Cloud Pak for Data, the team trained machine learning models on monitoring signals, reports, diagnoses, and patient-record data from Antwerp University Hospital. The implementation was designed around three architectural pillars identified by the team: a predictive model, an interpretable user interface, and a robust edge-oriented architecture. Explainable AI features in Cloud Pak for Data were used to help caregivers understand why the models were flagging risk, while federated learning let the models evolve without moving raw patient data outside the hospital. Parameters could move through the cloud, but raw NICU data stayed on premises behind hospital firewalls, creating a privacy-preserving operating model for continuous sepsis monitoring.
Monitoring signals, reports, diagnoses, and electronic patient-record data remain inside the hospital environment, Watson Studio models analyze the patterns through Cloud Pak for Data, and federated-learning parameters move in the cloud while clinicians receive explainable risk insight locally.
"Can help detect sepsis hours faster than medical staff"
"Innocens BV used IBM Watson Studio to train its solution's machine learning models"
"The UZA NICU is a highly digitized environment with multiple data sources"
"The raw data will remain on premises. Federated machine learning does this without moving the data."
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
IBM Sample Apr 1, 2026 MyLÚA Health MyLÚA Health needed an enterprise-ready way to deliver culturally attuned maternal support between appointments while protecting sensitive health information and identifying risk earlier in underserved populations. 48% of participants completed proactive wellness activities IBM Cloud · IBM Cloud Code Engine · IBM watsonx Orchestrate IBM Sample Dec 10, 2025 Global Pharmaceutical Manufacturer A global pharmaceutical manufacturer needed faster, more standardized supply-chain risk analysis because manual classification, fragmented documentation, and limited visibility into alternate suppliers and materials… >1500 risk events assessed, triaged and escalated to executives as needed AWS cloud services · IBM watsonx · IBM watsonx Orchestrate 
In-house Sample Jul 28, 2026 Novartis Novartis R&D needs to turn large volumes of scientific, clinical, operational, and real-world data into faster, higher-quality decisions across drug discovery and development, reducing cycle times, surfacing safety… less than 10 percent of drug candidates historically reach the market AI models · AI-enhanced analytics · Generative AI 