Intensive care units (ICUs) treat highly vulnerable patients with high mortality and attendant high costs. Many patients require breathing support via mechanical ventilation to survive and enable recovery. However, the application of mechanical ventilation can be very variable and difficult, with significant added risks to patients, as well as significant added costs. The main reason for this difficulty is the inability to directly measure patient-specific condition and response to therapy without invasive, time consuming testing. This research uses novel, clinically validated models of human lung mechanics with clinical data to identify validated metrics of patient condition and responsiveness to therapy. These will be used to track condition and guide treatment. The results will clearly define for the first time the evolution and time course of respiratory disease in the ICU, and test/prove the ability of engineering model-based methods to optimise mechanical ventilation to improve patient care, and outcomes and reduce cost.