This article is adapted from The Great Indoors: The Surprising Science of How Buildings Shape Our Behavior, Health, and Happiness, by Emily Anthes, published by Scientific American / Farrar, Straus and Giroux.
For more than half a century, writers and futurists, from Ray Bradbury to the creators of The Jetsons, have been conjuring up intelligent, high-tech homes that essentially run themselves. In their versions of the future, our homes would be much more than shelter: The houses of tomorrow would cook, clean, and care for us. They’d wake us up, make us breakfast, and then tidy up afterward.
Tomorrow is here. In homes across the world, smart thermostats glow, autonomous vacuums spin, and intelligent speakers stand at attention. Programmable shades rise with the sun, and connected refrigerators monitor our supply of milk. We can rely on smart flowerpots to water the plants, smart pet feeders to dispense kibble to the dog, and smart locks to let the maintenance worker in—all while we’re out of the house. By 2023, more than half of American households, and one-sixth of those around the world, are expected to have smart home devices.
The first wave of products aimed to alleviate the hassles of daily living, but now Big Brother is getting an MD. The smart mattresses that keep us comfortable are also collecting reams of data about our sleep quality, heart rate, and respiration, while some smart thermostats are monitoring the air quality in our homes. Many companies now sell smart pill bottles that light up, chime, or fire off text messages when patients miss a scheduled dose. Google has patented an optical sensor that would enable smart mirrors to monitor our cardiovascular health by detecting subtle changes in skin color, and Amazon has patented a system that could prompt our smart speakers to order cough drops when they overhear a sniffle. (It would, presumably, order those lozenges from Amazon.)
These advances, and the ones still to come, mean that our homes are becoming more intimately involved in our health care than ever before. Even as these technologies move into the mainstream, there’s one specific segment of the market that’s ahead of the curve, and it isn’t the one you might expect. It’s not the tech-savvy millennials who are pioneering the most advanced versions of these products—it’s senior citizens. It’s in the senior care community where scientists and engineers are putting all the pieces together, experimenting with integrated health-monitoring systems that are truly turning homes into medical devices. The potential upside is enormous. There are more than 700 million people on the planet who are 65 or older, a figure that is expected to double by 2050. Moreover, our lengthening life spans mean that seniors are living longer, spending more years in a phase of life that is often accompanied by increasing frailty and diminishing cognitive function. In theory, at least, in-home sensors, cameras, trackers, and monitors could help keep many of these older adults healthy and independent, allowing them to age safely in their homes, even in the face of illness and infirmity. These systems provide a glimpse of the future of smart home health monitoring—a preview of both the promise and the risks of letting our buildings play doctor.
Marilyn Rantz, a gerontological nurse and researcher at the University of Missouri, has dedicated her career to improving the experience of aging in America. Typically, as seniors’ needs increase, they get shuttled off to a series of residences that provide escalating levels of care. These moves, however, can be disruptive and dangerous. Transferring elderly men and women from one home to another, especially against their will, can cause confusion, sleeplessness, loss of appetite, anxiety, and depression. Seniors aren’t too keen on them either; in survey after survey, they report that they want to stay in their homes and remain independent as long as they possibly can.
In 1996, Rantz and some of her colleagues at the University of Missouri’s Sinclair School of Nursing decided to establish a new model for senior housing that would allow people to “age in place.” They partnered with the Americare Corp., which owns and operates senior living facilities, to create what they came to call TigerPlace, a nod to the University of Missouri’s feline mascot, and launched a health-care agency, which they called Sinclair Home Care, to administer it. TigerPlace would be a complex of apartments where adults could live out the entirety of their golden years. It would be an independent living facility, but an adaptable one. If a resident’s needs changed, Sinclair could bring in additional aid, like physical therapy, wound care, assistance with bathing and dressing, and, when the time came, hospice care.
In late 1999, while Rantz was in the midst of trying to get TigerPlace up and running, her elderly mother fell and shattered her shoulder. She was home alone and lay on the floor for eight hours before someone came to her aid. She never really recovered. “She was dead well within six months of that fall,” Rantz told me. “It was a classic example of what happens to people after falls and fractures.”
More than one in four older adults fall every year, and those who spend long periods of time lying on the floor have an especially poor prognosis. Rantz wondered if she and her colleagues could devise a technological system that would automatically detect falls at TigerPlace so that no one would have to suffer her mother’s fate.
Rantz collaborated with Marjorie Skubic, a professor in the department of electrical engineering and computer science. In a series of focus groups, Skubic and Rantz learned that while seniors were open to the idea of fall detection technology, they were wary of wearables, like pendants, bracelets, and clip-on devices. These gadgets require the user to do a lot of work, and they can stigmatize the wearer as ill or frail. So Skubic and Rantz decided to focus on developing sensors that could be deployed in seniors’ homes. “We said, ‘Let’s just make it so they don’t have to do anything,’ ” said Skubic. “You mount them in the environment, there’s nothing for them to wear, you don’t have to charge anything.”
Working with several colleagues and a large team of students, Skubic spent several years playing around with fall detection technologies in her cluttered university lab, eventually deciding that the most promising method relied on a piece of video game technology: the Microsoft Kinect, a now-discontinued motion-sensing camera that worked with the company’s Xbox gaming consoles. The Kinect contained a depth sensor that measured how far away objects were from the device. From this depth data, Skubic and her team could extract the silhouettes of human figures, track their movements over time, and detect when they took a tumble. When they were satisfied with what they’d built, they took it to TigerPlace.
I visited TigerPlace in the summer of 2018. The sprawling, one-story brick building, which opened in 2004, is located just a few miles from the university’s campus and has a lot of luxe touches: a grand piano in the lobby, brocade curtains, and so many chandeliers that I lost count. Every apartment has its own screenedin porch, which opens directly to the outside, and a large communal courtyard features fountains, bird feeders, and pots filled with herbs.
The residents range in age from their mid-60s to their late 90s, and the majority of them have at least one chronic disease; the typical tenant takes more than a dozen medications. They get comprehensive health exams every six months, though they retain their own personal doctors, who stay in close touch with TigerPlace nurses. The approach seems to be working; the average length of stay has been increasing since TigerPlace opened and has surpassed the average tenure at senior living facilities nationwide. Some residents stay for more than a decade, all the way up through their very last hours of life.
Researchers from the University of Missouri are constantly testing out new ideas at TigerPlace. Skubic’s team launched a fall detection study in 2013, installing the depth-sensing cameras in a dozen TigerPlace apartments, and the research team subsequently deployed the system in additional apartments at TigerPlace and at several other senior living facilities throughout the state. But from the beginning, the researchers’ goals had been much grander than simply responding to crises. Sensors installed in the home, they thought, offered them an opportunity to detect subtle changes in seniors’ health and behavior and to diagnose small problems before they ballooned into big ones. If they could identify folks at risk of falling, they could send in physical therapists to help them improve their strength and their balance; if they suspected that residents were developing pneumonia, they could start them on antibiotics and fluids.
So Skubic’s team developed gait-monitoring software that could identify residents who were likely to fall, as well as a hydraulic bed sensor, an array of long, flexible, water-filled tubes that sits under a mattress. The sensor monitors nighttime restlessness and measures heart rate and respiration by picking up on the minute movements of the heart and lungs. In addition, simple motion sensors mounted in the apartments provide clues about tenants’ daily movements and routines, tracking each time they enter and exit a room, for instance.
Together, these sensors perform what amounts to a constant virtual checkup, notifying the nurses if they detect any worrisome behavioral changes. A sudden uptick in nighttime restlessness might indicate pain; more trips to the bathroom might be an early sign of a urinary tract infection. A woman who’s spending more time in bed might be fatigued or depressed; one who begins leaving her apartment in the middle of the night might be in the early stages of dementia.
When the nurses receive an alert, they go chat with the patient in question. “We don’t go in and say, ‘Hey, you had an alert last night,’ ” said Kari Lane, the director of operations at TigerPlace. “But we go in and use careful questioning and relationship building to figure out how they are feeling and determine if anything is going on that we need to pay attention to. It’s kind of like detective work.”
In one recent case, the staff got an alert that one woman had been more restless than usual at night. When a nurse went to check on her, she admitted that she’d been sleeping poorly—and that she’d developed a funny tingling in her hands. That set off alarm bells; tingling hands can be a sign of dehydration and an electrolyte imbalance. The nurse called the woman’s doctor, who ordered lab work that confirmed the diagnosis. “And they were able to push fluids to avoid a hospitalization,” Lane said. Over the course of a single nursing shift, the woman had been evaluated, diagnosed, and treated—all without leaving home.
The system has helped the nurses catch pneumonia, delirium, heart failure, hypoglycemia, and other problems. It can detect changes weeks before significant health events, like falls, ER visits, and hospitalizations—before people report symptoms to nurses and, some-times, before they even perceive them. In a 2015 study, the research team found that TigerPlace residents who’d had the sensors installed were able to live in their apartments for 1.7 years longer than those who’d chosen to forgo them.
TigerPlace has provided powerful proof of concept, and the team is working to build upon what they accomplished there. One of Skubic’s former graduate students has formed a company called Foresite Healthcare to commercialize the technology. Foresite has developed its own depth-sensing camera to replace the Microsoft Kinect and has managed to bring the fall detection false alarm rate down even further, to one every three months. The company has installed its system in dozens of senior living facilities in multiple states.
Skubic and Rantz are thinking through how they might adapt the illness detection system for private homes. It’s not as easy to translate as it might seem, a lesson that Skubic learned firsthand when she installed one of the sensor systems in her parents’ South Dakota home. Shortly thereafter, her 96yearold father was diagnosed with pneumonia. When Skubic found out, she went back and looked at his data. The bed sensors made it clear that, over the previous six weeks, her father had become more restless. In fact, his restlessness had been significant enough that Skubic had received alerts about it. But she didn’t know how to interpret them—or what to do in response. Restlessness itself didn’t necessarily portend a problem, so she had just filed it away as an anomaly.
The only reason Skubic’s father ended up seeing a doctor was that his wife of more than 70 years insisted—she knew that he didn’t seem like his usual self. “So she basically was doing what we want our sensor system to do,” Skubic said, laughing.
The experience was instructive. “I mean, I’ve been working in this for 10, actually 15 years, and I didn’t know how to interpret the alerts,” Skubic told me. “We need to a do a lot better job of turning all this data into useful information, and we need to really make it simple.”
Skubic and her colleagues are working to translate their wonky, dataladen alerts into plain English. They’re exploring the possibility of delivering them through commercially available smart speakers, like the Amazon Echo or Google Home. “It’s set up as a conversational system so that you can ask questions like ‘How did I sleep last night?’ or ‘What is my fall risk?’ ” Skubic told me. “It also will support the use of family members, so that a family member could say ‘How did my mother sleep last night?’ Or ‘Does my mother have fall risk?’ ”
The bigger challenge is figuring out what comes next. At TigerPlace, the pathway for the nurses is clear. But what should an elderly woman do if she’s living on her own, in her own home, and Alexa tells her that she hasn’t been sleeping well lately? “The question becomes ‘Is it bad enough that I need to go see a doctor?’ ” Skubic said. “ ‘Or is this just something where I need to change what I eat or I need to go to bed earlier at night? Or am I taking the wrong medications?’ So somebody has to be able to look at this in a holistic way and give people some guidance.” Routing all of the alerts to the relevant physicians doesn’t seem like a reasonable solution, given how overloaded most doctors already are, so the team has been contemplating other solutions, like creating a network of nurses or “care coordinators” who can help consumers review and interpret their alerts.
While many seniors are legitimately excited about these technologies, some have expressed concerns that the devices will be intrusive, unreliable, or downright anxiety-provoking. Not everyone wants to see the excruciating minutiae of their body’s inner workings charted out for them in a daily e-mail. Even those who are open to the technology have concerns about privacy. “They would say things like, ‘I’m OK if my daughter sees it, but I don’t want my son to see it because I don’t trust my son’s wife,’ ” Skubic told me. “They viewed it as their data, which I thought was an interesting perspective and actually an encouraging perspective.” And although the fall detection cameras show the body only in silhouette, some seniors have asked not to have them installed in their bedrooms or bathrooms.
That’s understandable, but it comes with trade-offs. The more that developers limit the scope of the surveillance, the more the system will miss. That caused some misunderstandings when the TigerPlace researchers installed a few of their sensor systems in private homes on a trial basis. “Someone would call and say, ‘Well, I had a fall, and you guys didn’t call me,’ ” Lane told me. “Well, that fall wasn’t where the sensors are installed. Because they tried to install the fall sensors where the person felt they were most at risk for falling and also where the person was OK with them being installed for privacy reasons.”
Likewise, other senior care facilities that are interested in adopting the technology don’t always understand its limitations. “They want it to be plug and play,” Lane said. But getting the most out of it requires staff education and, often, a culture change, a commitment to being less reactive and more proactive. After all, the technology isn’t magic, and it’s not the sensors themselves that are keeping TigerPlace residents healthy. “The technology that we use here is wonderful, but it still needs nurses behind it,” Lane said. “It’s just another tool, like a stethoscope, to help us make informed choices.”