< Project Overview >
Sleep services in hospitals across the UK conduct sleep studies, known as polysomnography, to investigate and treat sleep disorders. They collecting a variety of information about patients during their sleep.
The range of data collected is used to analyse sleep patterns in patients including brain, eye, muscle and heart activity as well as oxygen levels, airflow and respiratory data. The study involves attaching up to 27 sensors to a patient, within in hospital laboratory setting, to measure these physiological signals.
Currently each polysomnography study is manually analysed by a sleep physiologist, taking an average of 2-3 hours. This is a very time consuming process and to date, no accurate automated analysis has been successful.
The development of an IoT platform, with potential to be used in the patient’s home, could significantly increase the convenience of patients and reduce NHS costs. These benefits are particularly pertinent to children as avoiding hospital admission would significant improve the quality of their experience.
This project brings together a multidisciplinary team, sharing knowledge between academic and clinicians to develop machine learning methods and a secure, ethical IoT platform to capture and analyse autonomously or semi-autonomously data sets.
The project utilises the clinical data and expertise from Sheffield Children’s Hospital alongside the data extraction and automation techniques of the University of Sheffield and University of Oxford teams.
The system will enable connection of a range of Bluetooth sensors to a central hub, either a phone or dedicated WiFi-connected device. Data will be recorded and securely transmitted to a cloud or on-premises server where strictly controlled access will be granted to authorised individuals.
Machine learning methods will provide clinicians with powerful tools to analyse the sleep patterns not only at hospitals but also in home environments.
The project aims to make the software publicly available and develop a ‘Framework for future IoT applications – in hospitals and home based’ to be used as a base for future projects.
Professor Lyudmila Mihaylova at the University of Sheffield