7.8 C
New York
Friday, April 19, 2024

What If Doctors Are Always Watching, but Never There?

Every doctor is trailed by ghosts. One of my youngest was a 7-year-old boy I’ll call Oliver. He had an upset stomach, and his mother booked an early morning appointment with their family doctor so that he could still make it in time for class. He seemed to have a routine tummy bug, but something about Oliver’s appearance made his family doctor call the hospital where I was working at the time. He asked if we could monitor the boy for a few hours. We found him a bed.

Oliver looked furtively at his mother as I stuck a needle in his arm to draw blood and hang up a drip. His initial blood readings were not bad at all, so we focused on other, sicker, children, and the nurses popped in to take routine observations every four hours. In modern wards, to prevent infection and protect privacy, patients like Oliver, who might be contagious, are put in isolation rooms—out of sight—with their observation chart hooked on plastic rails beside the door.

We first noticed something was wrong when his chart showed the dreaded “seagull sign.” When plotted on the same graph using the same scale, systolic blood pressure (marked on charts with a birdlike chevron) should, in healthy people, remain above the dot that signifies the pulse rate. Nurses are taught a simple rule: When this pattern reverses—when it looks like the poop is landing on the bird—it is a sign that the shit is about to hit the fan.

By that time it was already too late. When I drew blood a second time at around 6 pm, Oliver seemed tired and gray and was unflustered by the needle. I ran down the corridor to process the sample; the blood-gas machine showed that his blood was becoming acidic—an ominous sign. He was even sicker than he looked.

Of the millions of children who suffer with viral infections of the gut every year, very few experience serious complications. But Oliver was one of those who did. He had developed myocarditis, a meaty inflammation of the heart. Despite the frantic flurry of activity of the next few hours—the bedside chest X-ray that showed his heart was blimping up like a balloon, the high-flow oxygen mask we muzzled him with, the medicine we trickled into his veins—Oliver’s heart beat more and more stiffly. His illness was so serious that it was decided he should be transferred to the nearest specialist hospital. He was wheeled out of the ward to the intensive care unit, to wait until the ambulance arrived. It was there that Oliver’s heart stopped. Despite all the efforts to revive him, Oliver died.

One week later, the entire team who’d been involved in Oliver’s care gathered around a table in a cramped, windowless room to debrief. The child’s death left us shaken, with one question hanging like a dagger in the room: What could we have done to notice his decline sooner—while he was still just edging toward the precipice, and not after he had started the steep, downward slide?

This question stayed with me for years, as I started seeing patients in my own clinic. My desire to find better ways of really seeing what was happening with them was sharpened by the coronavirus pandemic. I wondered: How can we safely assess and monitor all these patients we now consult remotely, in their own homes? I set out to discover ways that technology might help me work more safely in the community, which led to a new piece of equipment that was initially developed for the Formula One racing circuit, and which is currently being piloted in intensive care to see if it picks up early signs of decline in children.

This new system, which continually monitors and collects patient data, has recently gone wireless. It is being tested on patients in a hospital in Birmingham, England, but it and similar remote systems might be used in patients’ homes in the future. The more I read on the subject, the more I realized that remote patient monitoring could change medicine radically: hastening medical responses and improving health outcomes; remapping the zones of health care; but also perhaps transforming how doctors like me think, in ways we might not so readily welcome.

Close observation of patients has been a universal duty of all doctors throughout time. For millennia, medical practitioners used their senses to assess a patient’s condition. Even now, we doctors are trained to recognize the hard-candy breath of sick diabetics, the glass bottle clonking sound of an obstructed bowel, and the cold, clammy feel of skin when a patient’s circulation is shutting down. But the systematic recording of numerical observations is a surprisingly recent phenomenon.

In the late 1800s, instruments were designed to measure a standardized set of health indicators. These are the four main vital signs: heart rate, respiratory rate, temperature, and blood pressure. It was just before the turn of the last century that these vital signs, also known as observations, were first documented systematically. By World War I they were used routinely. Studies of these charts revealed that people basically never died when these vital signs were normal; hearts don’t stop out of the blue. But for the better part of a century, the art of interpreting these so-called obs charts was, to the untrained, as mysterious as reading tea leaves.

Then, in 1997, a team based at the James Paget University Hospital, in Norfolk, England, developed an early warning system with which a nurse could quickly turn vital signs into a score. If the score surpassed a threshold, it was a signal to call for a doctor’s assistance. Such systems were steadily rolled out for adult patients, but it was not clear if they would work in children, whose physiological responses to illness are different from those of adults.

Heather Duncan knew about about early warning systems for adult patients in 2000, when she was working in South Africa as a general practitioner with a keen interest in children’s health. Ordinarily, observations taken in a hospital aren’t connected to earlier ones made in primary care clinics. But Duncan tried to link these two datasets—from the community and the hospital—to create a more meaningful, continuous story of what was happening to patients. She took the trouble to scrutinize the records of her sickest children, plotting their vital signs from the time they were first recorded in primary care to their discharge or death in the hospital. “I noticed children were having cardiac arrests or intensive care admissions, and that actually there were missed opportunities where we should have acted further,” she remembers.

Her nagging feeling that more could be done for such children was later corroborated by the UK’s Confidential Enquiry into Child Deaths, which found that more than a quarter of children in National Health Service hospitals were dying of avoidable causes. In 2003, Duncan completed a fellowship in critical care at Toronto’s Hospital for Sick Children, where—together with Chris Parshuram, a pediatric intensive care doctor—she developed the Pediatric Early Warning System, or PEWS, a bedside scoring system designed for sick children.

Duncan now works as a consulting pediatric intensivist in Birmingham Children’s Hospital. I caught up with her on Zoom last October. Duncan was working from home, wrapped up against the English autumn in an oversized, cream fleece, her hair pulled back into a loose bun, and wearing blue-rimmed specs that matched her eyes. She speaks with a genteel South African accent and has a calming manner, surely an asset working in such a stressful specialty. Her hospital had adopted the PEWS score in 2008 and seen a drop in the number of children dying after suffering a cardiac arrest—from 12 in 2005 to no deaths in 2010.

These impressive results, however, didn’t satisfy Duncan. There were still patients who spiraled downward, unnoticed, despite having observations recorded on a PEWS chart. What if, she wondered, patients could have their observations recorded not just when nurses found time to record them but continuously and automatically?

In 2009, Duncan happened to meet Peter van Manen, who was then managing director of McLaren Electronics (now called McLaren Applied), originally founded to design custom engineering components for Formula One sports cars. Both were speaking at the same health care conference. “The organizer of the meeting was a big F1 fan,” Duncan explains, “and so he had got McLaren to come in and talk about how they deal with stress and trauma, and that sort of thing, in their cars.”

 McLaren Applied supplies a standardized “electronic control unit” to Formula One teams to monitor wear and tear on cars during races. In his talk, van Manen explained how more than 120 sensors are scattered throughout a car’s chassis to monitor its 25,000 parts. These sensors convert this primary feedback into data that is sent back to the garage. Over the course of an average 2-hour race, this enormous amount of data is interpreted in real time, providing engineers with updates on the condition of the engine, tires, and fuel. Sitting behind van Manen on the stage, Duncan thought: If this could be done for thousands of parameters in Formula One cars, then why not for a handful of vital signs in our sickest patients in hospital?

Duncan collared van Manen after his talk and asked if they could collaborate. Her idea was to modify the platform that McLaren had been using for its race cars to work for humans. They tried out the idea in her pediatric ICU in Birmingham, attaching electrodes to a few dozen kids as part of a pilot study; over the next decade Duncan and partners from academia and industry worked to perfect the system. Data from all the patients on her ward could be viewed simultaneously by clinicians from a central viewing station. They used machine learning, using data from real patients to train the algorithm, to help predict the trajectory of vital signs for various subgroups of patients. This allows doctors to intervene hours earlier than with standard care. What’s more, the data can be stored for long periods and used for research and learning. Duncan named her team’s new automated system the Rapid (or Real-Time Adaptive Predictive Indicator of Deterioration) Index.

Duncan’s team soon extended the use of their kit to patients in the pediatric cardiac wards, since heart disease is a common cause of rapid decline in children. The downside was that being continuously connected to machines restricts movement, which isn’t great for antsy kids, and you risk losing a lot of data when wires are accidentally tangled or pulled off.

So in 2014 they reached out to companies to codevelop wireless biosensors. One of these is worn “like a little wrist watch,” Duncan explains. This connects to a finger clip probe that records oxygen saturation levels. Another sensor is stuck to the chest to record an electrocardiogram and respiratory rate. A sticker in the armpit measures temperature, and an upper arm sensor records blood pressure. These sensors transmit data by Bluetooth to a platform where it can be monitored continuously. When the measurements fall outside normal parameters, a warning appears immediately on a tablet or smartphone dashboard anywhere in the hospital, alerting staff.


Over three years, Duncan’s group monitored more than 7 million minutes of data from 982 patients at Birmingham Children’s Hospital. Last year, they published the first tranche of results in Nature Scientific Reports. Twenty-nine of those children had significant deterioration, and the Rapid Index identified the declines an average of 52 hours earlier than the PEWS score using hand-recorded observations. The Rapid Index not only caught deterioration sooner, it also picked up 20 percent more significant problems.

These seemed like big wins, but it is still early days. And a 2018 meta analysis compiling the results from 27 randomized controlled trials of wireless monitoring technologies used to track various chronic health outcomes concluded that there is still no statistically significant benefit from using remote monitoring sensors.

Duncan is the first to admit that wireless monitoring also comes with drawbacks. “The thing I worry about is the false alarms,” Duncan says, a problem they encountered during the study. “I really worry about that.” She fears that her kit could distract clinicians by crying wolf every few minutes. But she is confident that, with time, the system can be tweaked and used efficiently by trained staff.

So far, the Rapid Index has still only been piloted in Birmingham, and it could take time for wider adoption. Though Duncan’s team estimates that Rapid will save hospitals money, Gillian Gresham, an assistant professor at the Cedars-Sinai Medical Center who does research on wearable technologies, said more trials are needed to determine the real efficacy and cost effectiveness. “I am not sure that we are at the point of widespread implementation of Rapid index,” she said.

Still, in Duncan’s vision, wireless monitoring throughout the patient journey might help kids like Oliver. In that future world, a child would check in with his parents at the primary care clinic, where a nurse would put wireless sensors on him to start data collection. By the time he arrives at the hospital, doctors would have hours of data trends on which to base their treatment plans. If he unexpectedly starts to deteriorate, the system should be able to recognize it in advance of any irreversible damage. In short, this system might have saved Oliver’s life.

Remote patient monitoring technologies also have another potential: to uncouple patients from their health workers, allowing theoretically limitless distance between the two. In a 2012 TED Talk on his collaboration with Duncan, Peter van Manen said, “With wireless connectivity these days, there is no reason why patients, doctors, and nurses always have to be in the same place at the same time.” Duncan already has plans to expand the use of her kit outside the hospital.

“I think that, in 10 years’ time, everybody will be wirelessly monitored, whether that’s at home, en route in between primary and secondary care, or in hospital,” Duncan says. “It’ll be so normal, and it’ll connect to freeware on their phone.” I asked Duncan if she’s worried that this technology might be overused by the well-to-do “worried well” and underused by people who really need it. “That’s a very good public health question,” she says. “But I have not noticed a socioeconomic divide.”

It’s possible that such technology might actually help bring medical care to underserved communities that doctors now struggle to reach. One third of Australians, for example, live in rural and remote areas, where the life expectancy is four years less than it is in cities. A nonprofit called Integratedliving provides telehealth monitoring of vital signs for older Aboriginal and Torres Strait Islander people. Participants record their own vital signs, and the data is then transmitted to an automated platform, which prioritizes readings for clinical review according to the degree of abnormality. A study of the project showed that not only was the program less costly than in-person care, it also led to more timely and accurate diagnoses. Moreover, most participants found using the system reassuring, and gained insights into their own health and how to manage it.

Wireless systems like Duncan’s could provide a second pair of eyes—in fact, hundreds of pairs of eyes—installed in the homes of patients who are sick, but not sick enough to be in the hospital. Remote consultations would bolster the monitoring. “If a midwife has a population of 40 women at home that she or he is monitoring,” Duncan says, “then they would still do video calls intermittently to touch base.”

But while remote monitoring technologies will extend the frontiers of medicine into domestic, private spaces, will they also, paradoxically, push patients and health care workers further apart? With health care systems desperate to save money, this kind of innovation could give managers an excuse to load health care providers with more patients, or to cut nursing staff, hoping fewer people could do the same work by relying on digital tools. Duncan insists that her kit must assist, and not replace, human caregivers, but it is hard to see how this can be guaranteed. In a system where costs are measured and metered, it’s unlikely that the time saved by using technology would be allocated to help care workers commune in invaluable but unprofitable ways, with their patients.

In other words, remote patient monitoring may mean that doctors are always watching, but never there. I may be guilty of nostalgia; one could argue that remote monitoring is simply a predictable, and welcome, next step in finding safer ways to keep an eye on patients. But while the invention of the observation chart punted the doctor from the bedside to the foot of the bed, remote patient monitoring kicks us out of sight. Duncan says her team has tried to prevent this by requiring that patients get regular physical checks. “I wouldn’t set up a system that’s fully automated, that didn’t have somebody at least checking in once a day, if not twice a day, on these patients,” she says.

Remote patient monitoring could grow in popularity in tandem with remote consultations, a trend bolstered by pandemic protocols. I wonder how this will affect doctors psychologically. Might it change the way we think and how we practice? A fascinating experiment from 2015 suggests a discomfiting answer. Min Kyung Lee and colleagues from Carnegie Mellon University conducted a study with 46 lay volunteers from diverse ethnic backgrounds. The participants were asked to make medical decisions for strangers. They were told to choose between treatments that were either more effective but more painful, or less effective but less painful. Half of the participants advised people who were sitting in the same room, face to face, and the other half advised people remotely, using a video link.

The results depended on how the study subjects formed their sense of self, or what social psychologists call “self construal.” In Lee’s study, the subjects who had “independent” self-construal (defining themselves based on their intrinsic individual attributes, in isolation from the social reality around them) prescribed in exactly the same way both in person and remotely. But those who had “interdependent” self-construal (that is, those who see themselves through their relationships) were much more likely to recommend risky or painful treatments when dealing with people remotely rather than face to face.

It seems that people in the second category—people who rapidly take in social information and who rely on it to make decisions, a trait we usually treasure in our caregivers—are thrown off by telemedicine, where the social cues available are limited. People in this group also reported fewer feelings and emotions toward the video-call patients. The researchers concluded that, for those of us who create our sense of self from the people and environment around us, teleconsultation increases the “experiential distance” between medical provider and patient.

Granted, the study subjects were lay people, not health care providers. But reading it made me wary of a future that is rushing to meet us—one where remote monitoring of patients in the community, with intermittent advice by video conference, is routine. In such a world, will doctors like me, who relish in-person clinical encounters, feel more psychologically distant from our patients? And might it make me, unconsciously, treat my patients more unfeelingly?

I worry about what remote monitoring of patients will mean for a core but underappreciated job of a physician: namely, to bear witness. Probably the most famous painting of a sick child is The Doctor, by Luke Fildes. Painted in 1891, it is a masterwork of chiaroscuro; in the center of a dark room, two central figures are lit by a single oil lamp. We see a child in the middle of the frame, being observed by a visiting doctor, seemingly deep in thought. The child lies limply on a makeshift bed. A ceramic jug and bowl nearby suggest the parents have tried to calm a fever with a home remedy. The child has dropped some tissue paper, which remains, pathetically, on the floor. The doctor’s cup of tea, with a teaspoon absently resting in it, is getting cold on the side table. The first morning twilight shines faintly through a square window into the gloom; perhaps the doctor has been here, watching, all night.

The doctor has clearly crossed into another world; he wears a smart suit, but his patient is a child in a laborer’s humble cottage. In the background the parents worry in the shadows. The father stares intently at the doctor, and the mother, burying her face in her elbow, cannot bear to watch. Only the doctor can—must—look at the child, in the clear beam of the lamp. The scene was inspired by the compassionate care Fildes’ own son, Philip, received from their family physician. Philip died of typhoid fever on Christmas morning in 1877, when he was just a year old.

The painting is often used to demonstrate the archetypal 19th-century doctor, but in reality the scene would have looked a bit different. For one, the doctor would not be sitting there empty-handed; even in the 1870s, it was standard for doctors to carry some instruments, such as early versions of the stethoscope, a blood pressure machine, and thermometer. Fildes deliberately omitted these items to focus the viewer on perhaps the most important things a doctor can offer at times of critical illness: the desire to be present, the ability to endure suffering along with patients and carers, and the courage to watch.

I imagine how the painting would look if it were to capture 21st-century doctoring, if remote medicine becomes the norm. There are some additions I can guess: The child would have sensors on their arm, and the parents might be looking at wavy, colored lines on a monitor on the table. But what about the doctor? Where would they be? I can’t say for sure. Perhaps they’d be out of view, scrutinizing the data for this patient—and a hundred others—in a far-off hospital, or in their own home. Or perhaps they’d be there, just as the doctor in Fildes’ painting is, right by the bedside. Certainly, this is Heather Duncan’s hope for her Rapid Index technology—that it should enable, and not discourage, a doctor’s physical presence with their patients. If her vision of patient monitoring comes to pass, the doctor in our hypothetical painting would be smiling at the sleeping child, having been summoned by remote monitors to administer medicines in time to save its life, now packing up their things into their bag, and ready to leave and face the dawn.

Let us know what you think about this article. Submit a letter to the editor at mail@wired.com.

Related Articles

Latest Articles