A clinical team used MIT CSAIL’s AI to remotely monitor a COVID-19 patient

A clinical team in Boston this week reported that a device developed by MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), dubbed Emerald, enabled them to remotely monitor a COVID-19 patient’s breathing, movement, and sleep patterns using wireless signals. While Emerald has been deployed in over 200 hospitals, homes, and assistive-care facilities, including Heritage Assisted Living in Framingham, Massachusetts, this is the first time it’s been used to track the progression of the novel coronavirus disease.

The researchers say their study is a small but encouraging step toward treatments that minimize health workers’ exposure while potentially improving health outcomes. Emerald could also allow hospitals and physicians to triage less severe COVID-19 patients by monitoring them in their own homes.

“When doctors have to interact directly with patients to conduct exams or monitor vital signs, each step along the way represents an increased risk that they will get infected,” said Harvard Medical School assistant professor of psychiatry Dr. Ipsit Vahia, who participated in the study. “Given how Emerald can generate important health data without any patient contact, it could minimize the risk that doctors and nurses will catch the disease from their patients.”

Above: The Emerald system monitoring the patient’s breathing.

Emerald was developed over the course of several years by MIT professor of electrical engineering and science Dina Katabi. It’s shaped like a small box and transmits a low-power wireless signal throughout a space the size of a one- or two-bedroom apartment (even through walls), which reflects off of people’s bodies and is ingested by the box’s antennas to extract physiological data. Subtle but detectable electromagnetic field changes enable Emerald’s machine learning algorithms to measure not only movement and breathing, but clinical biomarkers like heart rate, gait, and more.

After obtaining consent from the COVID-19 patient, who’s a resident at Heritage, Emerald was installed in their room. It showed that over time, the patient’s breathing rate declined from 23 to 18 breaths per minute — much closer to their baseline — that their sleep quality improved, and that they were able to walk more quickly around their apartment as they recovered.

Emerald Innovations, the company Katabi cofounded in 2013 to commercialize Emerald, makes clear that the data the device collects only concerns specific traits and that it’s encrypted. Moreover, Emerald requires users to complete a series of movements before it can track them, guaranteeing that it can’t clandestinely follow unwilling participants.

In light of U.S. Food and Drug Administration guidance published in March to expand the use of remote monitoring devices that facilitate patient management, both hospitals and health systems are increasingly piloting AI-enabled at-home solutions. In Kentucky, Baptist Health is using an AI platform developed by Current Health to track about 20 COVID-19 patients. In Washington, Providence has deployed remote monitoring from Twistle to care for more than 1,000 confirmed and suspected COVID-19 patients. And this week, the LSU Healthcare Network in New York and New Orleans plans to begin leveraging AI to remotely monitor cardiac patients vulnerable to COVID-19.

Elsewhere, predictive analytics platform CLEW is working with Sheba Medical Center and the Ichilov Hospital at Tel Aviv Sourasky Medical Center in Israel to remotely diagnose the type of respiratory deterioration characteristic of COVID-19. And Orion Health released a remote monitoring platform that in the future will tap AI to allow providers to identify patients at risk of contracting COVID-19.