Abstract
Currently, there is a limited understanding of long-term outcomes of COVID-19, and a need for in-home measurements of patients through the whole course of their disease. We study a novel approach for monitoring the long-term trajectories of respiratory and behavioral symptoms of COVID-19 patients at home. We use a sensor that analyzes the radio signals in the room to infer patients' respiration, sleep and activities in a passive and contactless manner. We report the results of continuous monitoring of three residents of an assisted living facility for 3 months, through the course of their disease and subsequent recovery. In total, we collected 4,358 measurements of gait speed, 294 nights of sleep, and 3,056 h of respiration. The data shows differences in the respiration signals between asymptomatic and symptomatic patients. Longitudinally, we note sleep and motor abnormalities that persisted for months after becoming COVID negative. Our study represents a novel phenotyping of the respiratory and behavioral trajectories of COVID recovery, and suggests that the two may be integral components of the COVID-19 syndrome. It further provides a proof-of-concept that contactless passive sensors may uniquely facilitate studying detailed longitudinal outcomes of COVID-19, particularly among older adults.
Introduction
The understanding of COVID-19 progression is often constrained by "snapshots" taken during infrequent clinical visits or through subjective self-reporting. This is particularly challenging for older adults, who may face cognitive barriers to accurate reporting or isolation due to quarantine. Traditional monitoring tools, such as wearables, often suffer from low compliance or technical friction, while cameras present significant privacy concerns. This study addresses the need for continuous, objective, and passive longitudinal data to understand the transition from acute infection to long-term recovery—often referred to as "Long COVID"—within a patient's natural environment.
Methods
The researchers deployed a novel contactless sensor system called Emerald. The hardware consists of a device that emits low-power radio signals, significantly weaker than standard Wi-Fi. Machine learning algorithms analyze the reflections of these waves off the human body to infer physiological and behavioral metrics without requiring any action from the subject.
The system was installed in the rooms of three residents (ages 73, 81, and 88) at an assisted living facility. Over a 327-day monitoring period, the system captured 3,056 hours of respiration data, 294 nights of sleep, and 4,358 gait speed measurements.
Key Results
The longitudinal data revealed distinct "digital fingerprints" for each patient. A primary finding was the sensitivity of Respiration Rate (RR) to the infection's status. In symptomatic subjects, RR was significantly elevated and more volatile during the acute phase compared to the post-COVID baseline.
Beyond respiratory metrics, the study identified a "long tail" of behavioral recovery:
- Gait Speed: One subject showed a statistically significant improvement in walking speed (from 0.52 m/s to 0.62 m/s) over several months, correlating with clinical mobility recovery.
- Sleep Abnormalities: Persistent sleep disruptions were observed in symptomatic patients long after they tested negative for the virus, suggesting that neuropsychiatric effects can manifest or persist well into the recovery phase.
- Asymptomatic Stability: In contrast, an asymptomatic subject maintained a remarkably stable respiration rate (~11.5 BPM) throughout the entire monitoring period.
Discussion
While limited by a small sample size (N=3), this case series provides a proof-of-concept for using contactless sensors to monitor complex disease trajectories. The ability to detect subtle physiological changes, such as a sudden jump in RR prior to a medical emergency, suggests that these systems could serve as early warning tools. Future work should focus on scaling this technology to larger cohorts to validate these digital phenotypes and integrate them into automated clinical alert systems for managing chronic and infectious diseases in aging populations.
Summary updated