Rajasi Mills, MS, vice president, SleepIQ Health, Sleep Number, speaks on the practical use of Sleep Number 360 smart beds in measuring sleep metrics and leveraging data to identify symptoms characteristic of COVID-19 and other respiratory illnesses.
Unobtrusive sleep metric data derived from the Sleep Number 360 smart bed identified worsening sleep trends in people positive with COVID-19, signaling its potential use as a preventive, diagnostic platform, said Rajasi Mills, MS, vice president, SleepIQ Health, Sleep Number.
For your study assessing the practical use of the Sleep Number 360 smart bed to surveil and monitor sleep metrics of people with COVID-19, how was it conducted and what types of data were included?
The body's responses to viral infection affects sleep duration, quality, and cardiorespiratory function, and COVID-19 symptoms overlap with other respiratory diseases like flu and SARS as well. Early symptom detection in these conditions is important to encourage diagnostic testing, mitigate the spread of the disease, and enable early treatment. And, we know that smart and connected innovations that collect data over many years, like our Sleep Number 360 smart bed, hold promise for infectious disease monitoring.
The purpose of this study, specifically, was to determine if our 360 smart beds could predict at an individual level the worsening of COVID-19 symptoms, and we wanted to do this using our longitudinal biometric data that's captured in an unobtrusive, real-world manner. We started with an IRB [Institutional Review Board]-approved survey and we asked our sleepers to opt-in to answer questions about whether or not they suffered from COVID-19. We did this from August to November of 2020.
Of more than 9000 sleepers who completed the survey, over 3500 reported results that they did have a COVID-19 test. And by the end of this study, data for 122 respondents who tested positive and who reported the date of symptom onset were used, and data for over 1600 respondents who tested negative who reported testing dates were also used.
I would say, in addition, sleep duration, sleep quality, restful sleep duration, time to fall asleep, respiration rate, heart rate, motion level, all were used and obtained from ballistic cardiography data from all the way from January 2019 to December 2020.
These metrics were measured in the analysis and used to design a symptom progression model. With this, we were able to establish individual baseline biometric signals and detect substantial deviations from that baseline during the illness periods.
Can you speak on findings presented at SLEEP 2021?
Through the use of our 360 smart bed, our data revealed that worsening symptoms in people positive for COVID-19 is associated with a significant increase in sleep duration, respiration rate, heart rate, restful time, and motion. It's also associated with a decrease in sleep quality, but there was no apparent change in the time to fall asleep.
Also, the average duration of predicted symptoms was consistent with the duration of symptoms reported by the users. And using symptom progression algorithms, we detected 50% of the cases by day 0 of the symptoms and 75% of the cases by day 2 with a mean detection delay of 0 days. Some probability peaks for experiencing symptoms that predated COVID-19 really suggest that our model can detect respiratory illnesses that go beyond just SARS-COV-2 and can look at other illnesses as well.
How may use of unobtrusive sleep metrics be leveraged further?
We're also working on opportunities to expand these capabilities to detect symptoms for illnesses such as the common cold, influenza, and SARS. The intersections between health and retail, between technology and health are becoming more frequent.
Our beds are playing a cutting edge role in this evolution and will one day function as a preventive care platform for better overall health. And to that end, we do see a future where our innovations will give health care providers a unique insight into a patient's holistic health through sleep data.
Guzenko D, Garcia G, Siyahjani F, et al. Longitudinal, unobtrusive, and ecologically valid sleep metric estimation from a smart bed to predict the pathology of COVID-19. Sleep. 2021;44(suppl 2):A255. doi:10.1093/sleep/zsab072.649