Many patients admitted to intensive care units (ICU) require breathing support from a mechanical ventilator. The ventilator pushes oxygen in and out of the lung. This results in a waveform, which changes moment to moment if the flow changes, it’s like a “fingerprint” and certain conditions have a pathological “fingerprint”. Early detection of these abnormalities allows clinicians to detect problems and proactively intervene as appropriate, potentially preventing complications and improving patient outcomes. Unfortunately, these changes are commonly missed in hospital environments.  

This project will detect these changes automatically, through the development of artificial intelligence (AI) solutions to analyse the data from mechanical ventilators, enabling them to detect early signs of patient deterioration. This would alert the clinician, enabling them to provide proactive interventions to manage or prevent the complication from occurring, improving patient recovery and outcomes. This will be the first step in the creation of a smart ICU.