Mechanical ventilation (MV) is used to support or fully control patient breathing to treat respiratory failure. Patients with respiratory failure, specifically acute respiratory distress syndrome (ARDS) and acute lung injury (ALI), require MV to fully control their breathing and provide adequate positive end-expiratory pressure (PEEP) to prevent alveolar lung collapse. ARDS patient have high mortality rates up to 60%, with significantly increased daily medical cost. Current practice in using MV to treat respiratory failure is a mixture of clinician intuition and generalised one size fits all approaches, which can lead to poor MV therapy care and ventilator-induced lung injury (VILI), increasing time on MV and cost, while reducing patient outcomes.

Currently, there are growing trends towards individualising care in medicine. In ventilation care, model-based methods are applied to identify patient physiology and respiratory mechanics to assist and guide ventilation therapy. Model-based methods can be utilised to provide greater insight to patient condition and allow more optimal patient- specific care. Optimal, personalised care would increase the quality of MV therapy and decrease the duration of MV. It would also reduce VILI and, as a result, decrease mortality and morbidity along with their associated costs.

Spontaneous breathing (SB) effort is when patients try to breath on top of the supported MV breathe. SB patients are generally transitioned into assisted spontaneous breathing (ASB) ventilation modes, which synchronises with patient breathing and supports breaths based on their demand, resulting in reduced overall work of breathing and increased pulmonary gas exchange. However, SB breaths hinder the accuracy of identified patient-specific elastance. SB efforts can be highly variable and are not measurable without invasive tests, but would be clinically useful to knowIn this thesis, SB is quantified utilising dynamic lung elastance Edrs. The time-varying elastance is the sum of alveoli elastance chest wall elastance and patient elastance generated by the patient demand, and can be utilised to identify SB effort. The Edrs trajectory is used by itself and again with proven basis function models to quantify patient demand and SB effort. The SB effort is identified and validated using measured electrical activity of the diaphragm (Eadi).

Preterm neonates are prone to respiratory failure syndrome (RDS) due to their prematurity, which sees reductions in alveoli growth and surfactant production. Thus, they require MV therapy to assist breathing due to their lack of respiratory development. Neonatal MV is common in the neonatal intensive care unit (NICU) and aims to minimise duration of MV as prolonged ventilation can lead to bronchopulmonary dysplasia or VILI.

Infants are known to exhibit different pulmonary mechanics compared to adults, and are also not widely studied. In this thesis, a first in-depth study on neonatal elastance is presented. The well-validated single compartment model of pulmonary mechanics is used to identify patient-specific elastance in this cohort, which consists of 535,428 breaths over N=10 patients. The model fit was good and was further validated by com- paring with known physiological differences, such as weight, and use of surfactant.

Anecdotally, male infants are harder to ventilate than female infants. The sex differences between male and female infants stems from their fetal growth stage where, female infants are typically more developed than males by 1.5-2 weeks. This thesis quantifies sex differences in neonatal elatances and breath-to-breath variability of each cohort, a first of its kind result in MV for any cohort. Male infants shown higher specific elastance than female infants and thus have lower variability. The results indicate the potential for potentially entirely different approaches to MV care for male and female pre-term infants.

This thesis presents analyses of lung physiology, pulmonary mechanics, and quantification of SB efforts. These outcomes advances the state of the art in modelling and MV, particularly with first of its kind, in-depth analyses and results in NICU MV. The physiology and respiratory mechanics was successfully quantified in this cohort and further validated in subcohort analyses. These results presented would provide good basis for further clinical use of model-based methods in both adult and infant cohorts.