HI-AI-R: AI in mechanical ventilation

Intensive care medicine, especially the treatment of critically ill, ventilated patients, is characterized by an inherently high degree of complexity. Large amounts of heterogeneous real-time data are continuously generated— from ongoing physiological parameters and ventilation curves to laboratory diagnostic findings. This immense amount of data poses major challenges for staff. The need to make time-critical and patient-specific decisions in highly dynamic situations is becoming increasingly challenging and carries the risk of suboptimal therapy and information overload.

Establishing intelligent clinical decision support systems is an essential component in overcoming this challenge.
 

We address this complexity by researching and developing state-of-the-art, AI-supported decision support systems. These systems are based on advanced learning strategies and a data-driven training approach and are capable of merging clinical data in real time, identifying causal patterns, and structuring information processing. The goal is to derive and provide precise, patient-specific therapy recommendations for ventilation therapy. This intelligent assistance strengthens the decision-making confidence of clinical staff, which contributes significantly to improving treatment quality and patient safety.

The current leading project in this research area is the BMFTR-funded joint project HI-AI-R (“Holistic inference system for AI-based assistance in the control of intensive care ventilators”).