The potential of telemedicine in respiratory illness has so far failed to live up to expectations, in part because of poor objective measures of symptoms such as cough events, which could lead to early diagnosis or prevention. Continuous monitoring and analysis of cough events is currently feasible and affordable using generic readily available sensors such as those embedded in current smartphones. This brings along important research challenges in using a smartphone as a medical device, namely the necessity to deal with noisy inputs in mobile environments as well as battery consumption issues related to continuous sensor monitoring and computing. Smartcough implements robust signal processing and efficient machine learning algorithms to overcome those challenges, featuring:

  • Continuous cough monitoring for your Android smartphone
  • Robust cough detection enabling continuous monitoring in noisy environments (e.g., when carrying your phone in your pocket/bag)
  • Efficient implementation of cough detection algorithms leading to 48h+ autonomy running in background
  • High accuracy (>93%) in noisy environments
  • Self-calibration and autoadaptive power threshold to improve accuracy and efficiency

Benefits of the Research 
The research has shown that the app can be efficiently used for the assessment of efficacy of therapies and deterioration of a condition, and as a reliable tool for diagnostic triage. The amount of available application opportunities connects the civic value of the project with an immediate benefit to the Scottish economy. For instance, the identification of factors that help predicting the occurrence of COPD exacerbations would significantly lower the treatment costs of the diseases, which totals around £100 million a year for NHS Scotland, with over 122,000 bed days. Work productivity loss is also an issue in respiratory illness, with an estimate of approximately £2176 on average per patient per year due to reduced working hours in COPD patients in the UK.

The app can use contextual information on the user or environment (e.g. pollution, temperature, activity) to provide feedback on the health status of the user and how the environment can have impact on it.

Collaboration with External Partners
The Smartcough project currently involves a multidisciplinary team comprising academic partners with technical and clinical profiles. This multidisciplinary inter-academia team is complemented by a business partner (Cirrus Logic) with extensive expertise in the development of advanced audio signal processing (ambient noise cancellation, robust signal acquisition) and power management solutions (efficient signal processing algorithms and energy control systems). A civic partner (Chest Heart and Stroke Scotland, CHSS) is also involved to help in the evaluation of the developed technology from the users’ perspective.

The team are currently interested in further validation of the system by running clinical studies where respiratory patients are monitored using the app. They are keen to collaborate with NHS and industry partners in the medical and wellbeing sectors to explore further opportunities for the app.

Researcher Background
Dr. Pablo Casaseca is a Senior Lecturer in Signal & Image Processing within the AVCN Research Centre at UWS. His research interests are signal & image processing and analysis, statistical signal and image modelling, pattern recognition, data mining, and computer vision, with applications to different fields, especially telecommunications, Unmanned Aerial Vehicles, and Biomedical Engineering.

Contact Details
Dr Pablo Casaseca