New technology senses patient activity
By Kari Oakes
Sheila DeMelle was diagnosed at age 17 with multiple sclerosis. For most of her life, her MS didn’t slow her down too much. “My disability was invisible. I knew I had it; I knew the symptoms,” but her symptoms weren’t readily apparent.
But 12 years ago, after the birth of her second child, her symptoms flared. Now, at age 50 years, she wears a leg brace, uses a cane, and has had frequent falls. For these and other symptoms, she receives care at the multiple sclerosis clinic of Brigham and Women’s Hospital in Boston.
Though fatigue is her increasingly frequent companion, Ms. DeMelle, a psychotherapist in private practice in the Boston suburb of Wellesley Hills, said she’s grateful that she has been spared the cognitive impairments that those with MS sometimes experience. “I mostly sit when I’m at my job, and I can still work,” she said.
During one of her routine neurology appointments, Ms. DeMelle was approached about participation in a research study that examines the movements of individuals with MS via wireless technology. She readily agreed to participate, and was soon fitted with an ankle actigraph, and a specialized wireless router was installed in her home.
For 8 weeks, she wore the ankle actigraph while in her home. The router, installed at a central location in her home, tracked her movements. Ms. DeMelle said participating in the study was largely a nonevent for her. On occasion, she’d forget to remove the ankle actigraph bracelet before leaving for work, or she’d forget to don it again when she got home. The device needed charging every few days. The technology being studied in the research, however, didn’t need the anklet. It was there only as a control to check the accuracy of a new way to track human movement and activity.
Dubbed Emerald, the technology doesn’t need a wearable, a cell phone, or any other inputs. It just sits there, and senses movement.
Dina Katabi, PhD, and colleagues at the Massachusetts Institute of Technology (MIT) have built on wireless technology to create a device that resembles a standard wireless router. The radiofrequency technology is used in conjunction with software that uses artificial intelligence (AI) to sense how humans reflect radiofrequency waves, tracking position, movement, and even posture.
In the same way that wireless signals travel through walls within a home for Internet access, the Emerald technology “sees” through walls to track activity, explained Dr. Katabi, the Andrew & Erna Viterbi Professor of Electrical Engineering and Computer Science at MIT.
Speaking in Dallas at the annual meeting of the Americas Committee on Treatment and Research in Multiple Sclerosis (ACTRIMS), Dr. Katabi, a 2013 MacArthur Fellow, gave an overview of the potential of the wireless device, now being developed by a startup company.
All of these metrics are very important first, the health of the individual, as well as progression of various diseases, and also side effects of medications on the patients.
Scientists at MIT have collaborated with physicians and other researchers to investigate how another movement disorder, Parkinson’s disease, affects the daily movements of those living with the disease. Placing an Emerald device in the home of one individual in an assisted living facility, for example, showed care providers that the bulk of the individual’s waking hours was spent in one chair, in front of the television.
This kind of information can highlight the need for interventions to minimize social isolation, but it also can point the way to more targeted therapy, pharmacologic or otherwise, for individuals with all kinds of chronic conditions, said Dr. Katabi. In addition, the large volume of continuous biometric data captured by the Emerald device may show utility in clinical trials of drugs and medical devices.
The trial Ms. DeMelle participated in has thus far enrolled 10 participants, said Zachary Kabelac, PhD, a scientist with Emerald who trained with Dr. Katabi at MIT. “We have recruited and deployed with 10 participants so far with the goal of reaching 20 in total,” he said in an interview. The study’s lead physician is Tanuja Chitnis, MD, medical director of the CLIMB Natural History Study at Brigham and Women’s Hospital, and a strong proponent of the importance of a precision medicine approach to MS research and care.
In addition to deploying the device in the home, the researchers are performing standard clinical assessments including the Expanded Disability Status Scale, the Multiple Sclerosis Functional Composite, Modified Fatigue Impact Scale, and tracking current medications and other variables. “We hope to relate these gold standard clinical assessments with gait and activity captured passively from the home to learn more about the patient’s life outside of the clinic,” said Dr. Kabelac.
Sleep tracking without wires or devices
Properly placed, the Emerald device turns out to be exquisitely sensitive to human movement. In a separate study, Dr. Katabi and her collaborators looked at sleep (Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 1;3. doi: 10.1145/3130924).
With the rise and fall of a sleeping person’s frame, Emerald maps out each breath. And it goes farther: During her presentation at ACTRIMS, Dr. Katabi pointed to jagged indentations in the sine wave depicting respirations in sleep. “That’s not an artifact. Those are heartbeats,” she said.
A trial of the Emerald device used specialized algorithms to detect and analyze breathing and movements during sleep. Deployment of the device, called EZ-Sleep, in eight homes measured a total of 100 nights of sleep. Of these, 30 nights of sleep also were captured with an EEG-based sleep monitor, approved by the Food and Drug Administration.
On average, compared with the EEG data, sleep efficiency was accurate to within 2.8%, and the mean sleep latency, time in bed, and total sleep time errors were 4.9, 3.2, and 10.3 minutes, respectively. Waking after sleep onset, an important measure for individuals with insomnia and other conditions that cause fragmented sleep, was accurate to within an average of 8.2 minutes.
“These results show that EZ-Sleep’s accuracy is comparable or better than medical-grade actigraphy devices,” wrote first author Chen-Yu Hsu, a PhD student in Dr. Katabi’s group at MIT, and his coauthors. “Thus, EZ-Sleep can deliver clinically meaningful sleep parameters without asking the user to wear any sensor or record her sleep data,” they added.
Further, the device was able to detect bed location within the home, discern when participants were in bed and out of bed, and track data for two people within the same room simultaneously, wrote Mr. Hsu and his coauthors.
Those living with MS often have disordered sleep, which also may contribute to the profound fatigue of the disorder. Sleep diaries are commonly used to track sleep duration and efficiency by self-report; however, a recent study found poor concordance between subjective sleep and actigraphic assessment of sleep duration and efficiency in MS (Sleep. 2018 Apr. doi. 10.1093/sleep/zsy061.1033). The objectively measured sleep of those with MS was actually significantly worse than the diary entries would indicate.
What are the ethical implications of remote sensing?
During the presentation at ACTRIMS, clinicians expressed concerns about the privacy issues of a silent, all-sensing box living in their patients’ homes. For Dr. Katabi, though, Emerald offers the promise of frictionless patient assessment. With no forms, no wearable devices, and no paperwork, clinicians get detailed, accurate, timely assessments of how patients are really living with their conditions.
And consent by the patient or a legal proxy can be limited, so that only data about sleep stages, but not about movement within a space, can be captured and analyzed, for example.
Secure transmission and storage of information is another critical aspect of ethical use of remote sensing devices, said Dr. Katabi.
When you are sending information from the home to wherever it’s stored, everything is secure; everything is encrypted and maintained in secure storage.
The American Medical Association, keeping pace with increasing use of monitoring technologies, has included guidance for use of remote sensing and monitoring in its code of medical ethics. Opinion 1.2.9 calls for an initial determination of the appropriateness of remote monitoring, along with a thorough explanation to the patient or surrogate of how the technology will be used. Informed consent should be obtained before implementation of any such monitoring, according to the opinion.
“Collectively, physicians should … support research into the safety, efficacy, and possible nonmedical uses of remote sensing and monitoring devices, including devices intended to transmit biometric data and implantable radio frequency ID devices,” wrote the AMA authors.
As an individual with MS, Ms. DeMelle sees the utility of data capture without the burden of paperwork. If she were started on a new medication, she said, she’d welcome the ability to give her providers accurate feedback about her movements and sleep without checking off endless boxes every day. And she’d like to know more about her disease course, if prognostic clues could be gleaned from information about her sleep, activity, or gait.
Ms. DeMelle said she’s curious to see her personal results, but that she participated in the study “for the common good.” The ability to track patient movements in real time could be of benefit in research, but potentially also in daily clinical practice. “That’s evidence. That’s real data.”