Remote assessment of Mental Health Patients for Cardiac Safety using Smartwatch ECG

点击翻译按钮获取中文摘要

This machine learning project uses deep neural networks to investigate historical ECG databases containing more than 1 million patients to develop new methods of assessing cardiac risk in mental health patients from smartwatch ECG data.It aims to establish new ECG features that improve the prediction of arrhythmias due to drug interactions and builds on the company's established AI-enabled ECG interpretation software which has demonstrated Cardiologist level performance for arrhythmia detection.We use AI to extract subtle changes from smartwatch ECG signals which indicate the patient is experiencing an adverse reaction to their medication, for which they need immediate treatment. This enables us to create a cloud-based analytics platform where patients can be more effectively managed remotely.We all live in a hectic world where seeing the doctor and getting checkups are more complicated. This, coupled with poor lifestyles and increased stress levels, has increased the prevalence of mental health disorders over the last decade. We believe that the works undertaken as part of this project will allow patients to be more effectively managed outside the clinical environment.With this project, we are working to optimise how patient information is recorded and transmitted to the doctor. This project provides an exciting look into how machine learning can detect subtle but significant abnormalities from smartwatch data and provide that information to the doctor remotely, so they can more effectively care for their patients when they are not in the clinic/hospital---ultimately leading to improved patient safety and significant cost savings for healthcare systems.