OSA-IVC

Just-in-time immersive video coaching for CPAP, powered by 6G-enabled event-driven personalization to empower patients in real time

CPAP therapy can be highly effective, but many patients struggle at home with mask leaks, discomfort, and low nightly use and support is often generic and arrives too late.

This project bridges that divide through an explicit, step-by-step workflow which provides just-in-time immersive video-coaching videos at the point at which an adherence-critical situation arises, to enable the patient to respond on the spot. First, an intervention pipeline uses routine CPAP telemetry and short patient feedback, which captures device-level therapy quality such as use patterns, leaks, and residual events and patient feedback, but cannot fully explain why problems happen in the real world e.g., stress, sleep fragmentation, hypoxemia, or context-related discomfort) to improve the level of decision making for next intervention and decide the most appropriate support action (e.g., troubleshooting, education, motivation, or a technician callback); this is the care-decision backbone, and future sensor streams can be added later to refine decisions without changing the overall process. Next, a video selection and delivery stream maps the identified need to the best short clip and delivers it right away fast, targeted “what-to-fix/ what-to-do” guidance rather than long content. When the library is not sufficient, a video adaptation/generation stream supports creating or adapting a short video using templates, adds it back to the library for reuse, and captures engagement signals (opened, watch-through, quick comprehension feedback) to improve future matching. Finally, a 6G-PATH assets support the implementation of the end-to-end pipeline on Fraunhofer FOKUS (Berlin) testbeds with a hospital edge node at Charité to demonstrate realistic healthcare deployment and prioritized delivery during “moments that matter,” using a multi-vendor 5G SA environment based on Open5GCore with capabilities such as NEF API integration and slicing. Success is measured with practical indicators time-to-delivery (targeting ~60 seconds after a trigger), smooth playback, and patient engagement/understanding supported by secure, role-based access and auditable handling of health data.