Monitoring and improving patient-ventilator interactions for pressure support ventilation

Asynchronous interaction between patients and their ventilators is very common. However, the mechanisms leading to asynchrony may be complex. There remains much for the research and clinical communities to understand regarding the nature of patient-ventilator (P-V) asynchrony and its effects on patient outcomes. Furthermore, commercially available ventilators may be deficient in their capacity to monitor and rectify asynchronous (P-V) interaction. The aims of this thesis were to 1) develop technologies that improved monitoring and detection of P-V asynchrony, and improved P-V synchrony using novel cycling algorithms; 2) achieve technical designs that were uncomplicated and clinically meaningful such that their adoption into clinical practice would not be met with resistance or inertia; and, 3) promote change in the nature of ventilator technology such that expert intervention would be less critical during titration and patient monitoring. Four studies were undertaken in furtherance of these aims. The first two were devoted to developing tools for monitoring asynchrony. The others were focussed on developing methods for ameliorating expiratory asynchrony, and comprised a method for measuring respiratory mechanics that fed into two adaptive cycling algorithms. Outcomes include1) the first real-time P-V asynchrony monitoring algorithm implemented in a ventilator device that detected trigger-related asynchronies in a patient cohort with >90% accuracy; 2) the first application of pattern classification to identifying patient-ventilator asynchrony, with the feature sets and classifier able to detect expiratory ineffective efforts and double triggering with >95% accuracy; 3) a real-time and adaptive algorithm for estimating respiratory system mechanics implemented in a ventilator, intended for spontaneously breathing, non-invasively ventilated patients; and, 4) two novel methods for automated cycling that significantly reduced expiratory asynchrony compared with conventional cycling by up to 50% in a patient cohort. This thesis marks important beginnings for new approaches to the treatment of P-V asynchrony. The future for research in P-V interactions needs to encompass both monitoring and intervention in a joint venture to unite cutting edge advances in ventilator technology with common clinical practice. Only by effectively monitoring P-V interaction can the problem of asynchrony be better understood and ameliorated through smarter technology