If you are a digital health company struggling to keep diabetes patients engaged with your app — this project developed an integrated self-management system combining predictive glucose models, automated e-coaching based on behavioral science, and personalized action plans. It was validated through randomised trials across 3 countries with both T1 and T2 diabetes patients, giving you a clinically tested blueprint to upgrade your platform from passive tracking to active, predictive coaching.
AI-Powered Diabetes Self-Management System with Predictive Coaching and Real-Time Monitoring
Imagine having a smart health coach on your phone that knows your diabetes patterns better than you do. It watches your blood sugar data in real time, predicts what will happen in the next few hours or months, and nudges you to take action — like adjusting meals or exercise — before problems hit. The system was tested with real patients in 3 countries (Netherlands, Germany, Spain) through proper clinical trials, combining short-term glucose predictions with long-term complication risk scoring to keep people on track with their care plans.
What needed solving
Diabetes patients struggle to stay on top of their condition between doctor visits — blood sugar swings, missed medication, and lifestyle drift lead to expensive complications and hospital admissions. Current apps mostly track data passively without predicting what comes next or coaching patients to change behavior. Healthcare providers and insurers pay the price when prevention fails and patients end up in emergency care.
What was built
The project built a complete integrated self-management system including: a decision support engine combining 3 predictive models (short-term glucose, medium-term progression, long-term complication risk), an automated e-coaching recommender engine with personalized action plans based on behavioral change science, a cloud-based data integration service connecting personal devices and health records, and both web and mobile user interfaces — all validated through randomised clinical trials in 3 countries.
Who needs this
Who can put this to work
If you are a health insurer dealing with rising costs from diabetes complications and hospital admissions — this project built a cloud-based decision support system that processes real-time personal device data and electronic health records to predict complications before they happen. The system was tested with 8 consortium partners across 6 countries, showing how predictive prevention can shift spending from expensive crisis care to cheaper early intervention.
If you are a medical device company looking to add intelligence to your hardware — this project created a data integration service that collects readings from personal devices and feeds them into predictive models for short-term glucose levels, medium-term disease progression, and long-term complication risk. The system includes both web-based and mobile GUI components, giving you ready-designed interfaces to turn raw sensor data into actionable patient guidance.
Quick answers
What would it cost to license or integrate this technology?
The project was coordinated by TNO, a major Dutch applied research organization, with 3 SMEs in the consortium. Licensing terms would need to be negotiated directly with the consortium. Based on available project data, no specific pricing model is published, but the mix of research and industry partners suggests technology transfer pathways exist.
Can this scale to handle large patient populations?
The system uses a cloud-based Data Integration Service designed to collect and process data from personal devices and EHR/PHR systems in real time. It was validated across 3 country trials (NL, DE, ES), demonstrating cross-border scalability. The architecture with 55 deliverables covering data processing, recommender engines, and GUI components suggests enterprise-grade design.
What is the IP situation and who owns the technology?
IP is shared among 8 consortium partners across 6 countries (AT, DE, ES, FR, NL, TR). TNO as coordinator likely holds key patents on the predictive models. Specific IP arrangements would follow Horizon 2020 grant agreement rules, where each partner typically owns the IP they generated.
Does this comply with medical device regulations?
The system completed randomised clinical trials in 3 countries, which is a strong foundation for regulatory approval. However, as a Horizon 2020 research project, full CE marking or FDA clearance would still require additional regulatory steps. The clinical trial evidence significantly de-risks the regulatory pathway.
How long would it take to integrate this into an existing platform?
The project produced multiple prototype releases and an integrated system with both web-based and mobile GUI components. The modular architecture — separate predictive engine, recommender engine, action plan engine, and data processing layer — allows partial integration. Based on the 55 deliverables produced, substantial technical documentation exists to support integration.
What clinical evidence supports this system?
The project completed a full randomised trial campaign (D5.4.3) with field trials in the Netherlands, Germany, and Spain. Three successive prototype releases were tested with real T1 and T2 diabetes patients. This level of clinical validation is unusually strong for an EU research project.
What patient data does the system need to work?
The system processes data from personal monitoring devices (glucose meters, wearables) and electronic health records (EHR/PHR) in real time through its cloud-based Data Integration Service. It combines this with predictive models for plasma glucose, disease progression, and complication risk scoring. Data privacy compliance would follow GDPR standards given the EU consortium.
Who built it
The POWER2DM consortium brings together 8 partners from 6 countries (AT, DE, ES, FR, NL, TR), led by TNO — one of Europe's largest applied research organizations. With 3 industry partners and 3 SMEs making up 38% of the consortium, this is a balanced mix of research muscle and commercial intent. The presence of SMEs signals real market interest, while the multi-country trial setup (Netherlands, Germany, Spain) demonstrates the system works across different healthcare systems and patient populations. For a business looking to license or build on this technology, TNO as coordinator provides a credible, well-resourced entry point for technology transfer discussions.
- NEDERLANDSE ORGANISATIE VOOR TOEGEPAST NATUURWETENSCHAPPELIJK ONDERZOEK TNOCoordinator · NL
- ACADEMISCH ZIEKENHUIS LEIDENparticipant · NL
- SRDC YAZILIM ARASTIRMA VE GELISTIRME VE DANISMANLIK TICARET ANONIM SIRKETIparticipant · TR
- SERVICIO ANDALUZ DE SALUDparticipant · ES
- SALZBURG RESEARCH FORSCHUNGSGESELLSCHAFT M.B.H.participant · AT
TNO (Nederlandse Organisatie voor Toegepast Natuurwetenschappelijk Onderzoek), Netherlands — contact through their technology transfer office or via SciTransfer for a facilitated introduction.
Talk to the team behind this work.
Want to explore how POWER2DM's clinically validated diabetes prediction technology could fit your product or patient population? SciTransfer can arrange a direct introduction to the consortium and help you evaluate licensing or integration options.