If you had a COVID-19 test at Brigham and Women’s Hospital in Boston earlier this year, you might have been approached by a dog-like robot mounted with cameras. In an attempt to find ways to lower exposure to the virus for medical professionals, researchers from the hospital and MIT were testing new ways to monitor patients’ vital signs without needing face-to-face contact.
The researchers started working on a prototype in March, as COVID-19 cases were growing in the area, and masks and other protective equipment were in short supply. “We started to think about how we could protect healthcare providers and minimize contact with folks that might be infectious,” says Giovanni Traverso, who is an assistant professor of mechanical engineering at MIT, a gastroenterologist at the hospital, and the senior author of a new study about the technology. They partnered with Boston Dynamics, a company that makes Spot, a robotic dog that can be controlled remotely with a handheld device.
At a triage tent outside the hospital where patients got coronavirus tests, healthcare workers steered the robot to patients. Four cameras mounted on the robot gather data. An infrared camera measures skin temperature on the face, and an algorithm uses that information to calculate core body temperature. The camera can also track the patient’s breathing rate by measuring the changing temperature of their mask as they breathe in and out. The three other cameras each measure different wavelengths of light, making it possible to track blood flowing through blood vessels and calculate pulse rate and blood oxygen saturation.
These technologies aren’t new, but the pandemic presented an opportunity to combine and test them. “I think collectively, as a community, what we’re recognizing is that there are many tools out there today from the robotics perspective, the sensor perspective, and really integrating them and really evaluating them in the field and seeing how they perform is part of the next step,” says Traverso. “Really, what we did here is translating a lot of the work that other people have done, but integrating it and asking the question, is this something that could work? How well does it work?”
The new paper from the researchers explains that the technology worked: Doctors got a good sense of the patients’ vital signs. A second paper, soon to be released, examines how patients reacted to the robot (the feedback was positive) and whether it’s something that’s actually cost-effective for hospitals to use. The team is now developing robots of its own; the original Spot robot, which can navigate complex terrain, is more complex and expensive than is necessary in a hospital. It’s also continuing to improve the whole robotic measurement system. “We’re really focused on honing and further improving all of the algorithms and cameras to really maximize how we extract that data, ensuring it’s as accurate as possible,” he says. The system would need FDA approval before it could be used in regular care outside an experiment.
The robots could also be used to bring a tablet into a hospital room so a doctor could talk to a patient without ever walking inside (this tech already exists, albeit with less agile robots). Similar tech could also eventually be used to monitor patients from their own homes. “I think monitoring patients at home could be one of the ways that we can start to identify signs and symptoms earlier,” Traverso says.