Project Abstract

In the critical realm of emergency medical services in Wales, rapid sequence induction (RSI) is a lifesaving procedure that requires timely and precise execution by EMRTS personnel. This study introduces a pioneering approach to enhance the monitoring and analysis of pre-hospital anesthesia through machine learning (ML), aiming to predict patient out-comes and identify procedural anomalies during RSI. By automating the analysis of time-series physiological data, specifically ETCO2, SPO2, pulse rate, and blood pressure, this research endeavors to conserve approximately 15 hours per month currently expended on manual data review.Our methodology leverages natural breathing patterns and apnoeic times detected in the ETCO2 waveform to algorithmically determine the commencement and success of intubation. Machine learning models, including logistic regression and convolutional neural net-works, are trained on data fetched via IntelliSpace Corsium web API, which provides minutely time-stamped vital signs alongside discrete non-invasive blood pressure readings. These models are tasked with predicting critical outcomes such as hypercarbia, hypocarbia, hypoxia, hypothermia, and hypotension, as well as RSI attempts and durations.Preliminary findings suggest that ML can effectively discern and predict outcome-relevant patterns in physiological data that are typically overlooked in manual audits. The transition to ML-based audits could significantly improve the quality of emergency medical care by providing real-time insights and predictive analytics. This study not only advances our understanding of ML applications in emergency healthcare but also sets a precedent for the integration of intelligent systems in enhancing patient outcomes in high-stakes medical environments.

Keywords: Machine Learning in Healthcare, Time-Series Data Analysis, Predictive Analytics, Process Automation, Deep Neural Networks, Decision Trees, Random Forest, Logistic Regression, Prediction models, Predictive Analytics in Medicine, ML Driven Process Automation, Data Integration

 Conference Details

Session: Presentation Stream 21 at Presentation Slot 1

Location: GH022 at Wednesday 8th 09:00 – 12:30

Markers: C�cilia Pradic, Joe Macinnes

Course: MSc Computer Science, Masters PG

Future Plans: I’m looking for work