Extracorporeal membrane oxygenation (ECMO) is a life-saving intervention that is complex and risky, which makes effective training essential. Traditional ECMO training is resource-intensive, requiring clinical equipment, an intensive care unit (ICU), and qualified instructors. To address this gap, the LAIXR research group developed a virtual reality (VR)-based ECMO simulator. This thesis extends that work by implementing dynamic patient-physiology modelling and one clinical training scenario to improve realism and educational value.

The project was split into three stages: 1) development of a virtual patient monitor to display physiology modelling, 2) scenario scripting and development of a predictive physiology model, and 3) a pilot usability study to assess feasibility and realism.

The virtual monitor generates four synchronised signals, the electrocardiogram (ECG), arterial blood pressure (ABP), photoplethysmogram (PPG), and capnogram, in real time. Models combine literature-based methods with ICU data: a Gaussian ECG, Fourier-series reconstructions for ABP and PPG, and a piecewise capnogram model. Clinically relevant behaviours are included, for example, pulse pressure variation during mechanical ventilation and physiologic delays between signals. The patient monitor layout and logic were based on the Philips IntelliVue MX750 to ensure familiarity and realism.

A predictive model was developed to generate the physiological inputs required by the patient monitor. Using data from 82 VA-ECMO patients, the first stage predicts ECMO pressures and flow from device settings such as pump speed and cannula size. The second stage estimates vital signs from pump speed and the predicted ECMO parameters.

A VA-ECMO console-failure scenario was implemented to train recognition of abrupt pump stop and the restoration of flow with the hand crank. All required interactions were included, and the physiological response to console failure was modelled using fitted functions derived from clinical data and expert review.

The pilot usability study explored the feasibility and perceived realism of the simulated ECMO and vital parameters. Participants indicated that the prototype was usable and realistic, and they identified priorities for richer interactions, broader scenarios, and expanded model coverage.

Overall, this work delivers a VR-ECMO simulator that integrates data-driven, dynamic physiology modelling. The prototype demonstrates feasibility and provides a foundation for future development toward scalable, immersive ECMO education.