Patient Monitoring Roundtable No°8 | 2025: High Quality Data – New Opportunities in Acute Care and Neurology

Save the Date: The first roundtable of the new year 2026 will take place on January 28 at BeST (Berlin Simulation and Training Center) on the topic of “Beep! Alarm or just noise? The potential of UX sound design in healthcare.”
Interested? Then book your ticket here for on-site or online participation!

Under the theme “High Quality Data: New Opportunities in Acute Care and Neurology”, we met on November 27 at the BIH Digital Labs for the eighth and final Patient Monitoring Roundtable (PMRT) of the year. Our goal: to strengthen the exchange between clinicians, researchers, and industry on how monitoring data can serve as a reliable foundation for AI-driven innovation in acute care and neurology.

In his keynote, Prof. Meisel presented how high-quality continuous monitoring data and AI-supported models can help detect neurological diseases earlier, understand them better, and support more proactive clinical decision-making.
He demonstrated how continuous monitoring and “physiomes” derived from clinical as well as home-based data provide valuable insights to improve understanding of disease trajectories and enable proactive rather than reactive care.

Computational Neurology in Practice
Prof. Meisel showcased research projects that already bring neural and AI-based models into clinical reality today: an ambulatory video-EEG system for epilepsy diagnostics, previously possible only in hospitals, and AI-driven seizure detection via smartphone-based motion analysis.

Data Warehouse Neurology – Infrastructure for Innovation
The Data Warehouse Neurology integrates clinical data from EEG, laboratory testing, and imaging, supported by clear governance structures, data protection concepts, and technical interfaces. This provides a robust foundation for data-driven research and aims to translate data into clinical innovation – “From Bench to Bedside.”

• Atrial fibrillation prediction: In stroke monitoring, Prof. Meisel demonstrated that AI can detect atrial fibrillation — responsible for about 30% of cases — directly from raw ECG data. A 72-hour AI-supported monitoring period enables targeted follow-up, with heart rate variability identified as a key marker, including for predicting the risk of post-stroke pneumonia. In the future, such parameters may even be captured through wearables like smartwatches.
• Neuroprognostics: Development of the world’s largest “Universal Map of EEG,” a deep neural network that can scale the analysis of physiological and pathological EEG patterns.
• Epilepsy & stroke detection: Video- and AI-based motion analysis to automatically detect epileptic seizures and motor changes in stroke patients.

• Questionable evidence: Many AI studies are overly optimistic, methodologically weak, and not externally validated → reproducibility crisis.
• Over-regulation in the EU: Strict regulation slows innovation → mindset shift needed: “a right to AI” rather than solely “protection from AI”.
• Interdisciplinary gaps: Closer collaboration between computer scientists, clinicians, and other disciplines is essential to ensure clinical relevance and technical quality.
• Operationalization & translation: Prospective “silent” validation, clinical dashboards, and continuously learning systems are required to bridge the gap between research and clinical practice.

Prof. Meisel concluded with the reminder: “Garbage in, garbage out.” Even the most advanced AI is only as good as the data behind it. Self-supervised learning and explainable AI are essential to ensure data quality and confirm that models learn true patterns rather than artifacts.

Workshop:
Mapping Data Quality – How Do We Generate High-Quality, High-Frequency Data?

The workshops focused on understanding how data can be generated that is both high-quality (precise, realistic, reproducible) and high-frequency, and what this means for different clinical settings. Two use cases were explored in depth: loss of consciousness and stroke. Participants discussed:

• the key role of interfaces, interoperability, and data transfer across devices, systems, and care sectors,
• challenges caused by siloed structures between clinical care and research,
• difficulties in patient matching for research cohorts, not least due to the lack of a unified patient health ID in Germany and the EU,
• regulatory requirements for medical devices,
• and the differing needs of professional groups: while some criticize excessive data collection, extensive documentation is legally required in emergency settings.

Take Home Messages

• Data quality is crucial: High-quality, valid data is the foundation for reliable AI models. As Prof. Meisel emphasized: “Garbage in, garbage out.” Explainable AI and self-supervised learning help distinguish true patterns from artifacts and support clinically meaningful outcomes. High-quality clinical and home-based data enable proactive detection and understanding of neurological diseases.
• Practical AI applications: AI already provides concrete value in neurology: prediction of atrial fibrillation from raw ECG data, assessment of post-stroke pneumonia risk, scalable EEG analysis via the “Universal Map of EEG,” and video-based AI for epilepsy or stroke patients.
• Challenges & roadblocks: Key hurdles remain—overly optimistic studies without external validation, strict regulatory requirements, lack of unified health IDs, and difficulties in translating research into clinical practice.
• Infrastructure & interdisciplinarity: Data warehouses, interfaces, and clear structures enable the use of heterogeneous data. Successful implementation requires close collaboration between medicine, computer science, and industry.

2026 promises to be an exciting year – until then, have a great time!

As the year draws to a close, we would like to express our sincere thanks to all participants, partners, and supporters for a wonderful year. We wish you a restful and peaceful Christmas season, happy holidays, and a good start to a healthy, happy, and inspiring new year in 2026!

We look forward to welcoming you on January 28, 2026, at the first Patient Monitoring Roundtable of the new year at BeST (Berliner Simulations- und Trainingszentrum)! Our focus topic will be „Beep! Alarm or just noise?The potential of UX Sound Design in Healthcare”.
Don’t miss it!

The Patient Monitoring Roundtable is organized by INCH Health in partnership with the Institute of Medical Informatics at Charité – Universitätsmedizin Berlin.

A special thank you goes to our sponsors Masimo, Dräger, and Philips, whose support makes the Patient Monitoring Roundtable possible.

We also thank Prof. Dr. Christian Meisel for his inspiring keynote, the Berlin Simulation and Training Center, and all participants for their engaged collaboration in developing future scenarios and visions for intelligent, safe, and patient-centered care.