CDT Postgraduate Studentship in Healthcare Innovation (RCUK Digital Economy Programme grant number EP/G036861/1) (EU)
Detection of OSA using audio, actigraphy, PPG and demographics
Obstructive Sleep Apnoea (OSA) is a prevalent disorder affecting up to 4% of the general population with severe consequences including increased risk of accidents, stroke and diabetes. The standard method of diagnosing the condition involves an overnight sleep study which records numerous signals. It is argued that such studies take the subject out of their normal environment and lead to unrepresentative results. A diagnostic method based on the signals that can be obtained by a mobile phone i.e. audio, actigraphy, body position, demographic information would be preferable. Data from a night spent at home could be analysed, and only those subjects with a high likelihood of OSA would be referred to hospital for further testing. This type of system would also allow for long-term monitoring, something which is not available currently, in order to see how well different treatment work.
Data collected from home studies using a clinical system, Grey Flash (Stowood Scientific Instruments, Oxford, UK) was used to develop a classifier for a variety of signals in different combinations including audio, actigraphy and demographic information. The model achieved an area under the receiver operating characteristic curve of 0.85.