Breathe for Science: Detecting Respiratory Disease in Breathing Patterns
CDS researchers use the remote collection of recorded breathing patterns to study the link between respiratory diseases and breathing patterns in the U.S. with hopes of better understanding the potential of contact-free screening of patients with respiratory diseases.
Amid the crisis of the COVID-19 pandemic, people across the globe have begun adapting to life in isolation. More and more, we are relying on virtual and remote ways to perform our typical everyday activities. Now, the possibility to diagnose respiratory diseases using remotely recorded breathing patterns is being investigated.
Prior studies have shown breathing patterns to exhibit diagnostic differentiation between normal patterns of breath and patterns in patients with respiratory conditions. The remote, contact-free collection of these breathing patterns holds great potential for the diagnosis of contagious diseases. Due to this potential, CDS researchers have been investigating the relationship between breathing patterns and respiratory diseases beyond controlled environments to establish whether the relationship between breathing patterns and respiratory diseases survives when collected remotely in an uncontrolled environment. Associate Professor of Data Science and Computer Science Kyunghyun Cho and Ph.D. student William Falcon and have introduced the platform “Breathe for Science”, designed to collect a database of breathing patterns. This platform consists of a pre-recording survey to collect background information about the subject and an automated phone call to record the subject’s breathing pattern.
The researchers have been coordinating with medical researchers and physicians to ensure the potential future use of the platform. The staged release of the collected data, after rounds of analysis and preprocessing, will ensure that the data meets the quality necessary to conduct further research. The platform takes into account scalability and privacy, encoding telephone numbers so that each caller is unidentifiable except for the self-diagnostic information they provide. Regarding scaling, the architecture of the platform is carefully constructed to prevent bottlenecks and spamming.
This platform is not a quick solution to diagnosing COVID-19, and the platform is carefully designed to make this clear. The platform is, however, intended to provide a collection of breathing patterns and self-reported diagnosis to assist medical professionals and researchers to better understand the potential of contact-free screening. The platform is also conducted with the goal of open sourcing and supporting the “Breathe for Science” platform outside of the United States. Due to the importance of localization in medical research, the researchers’ plan is to distribute the code base behind their implementation with intent to facilitate the implementation and adoption of this platform in other locations.
By Mary Oliver
