Skip Navigation

ALS News

field with sun shining

Jun 15

The Measure of a Man

ALS Headlines
Moving beyond simple fitness tracking, wearable devices may soon offer opportunities for monitoring health and bring vast amounts of new data to the study of human diseases.
Since the first Fitbit was introduced in 2009, it’s become more common to see people wearing fitness trackers. Where early devices mainly counted steps, more recent versions also measure heart rate, figure out whether the wearer is exercising, and attempt to capture sleep quality. But the future of wearables, some scientists believe, lies not in tracking fitness but in monitoring health. Devices such as a Garmin tracker or an Apple watch or their successors could provide early warning of disease, aid diagnosis and treatment, and contribute to a deeper understanding of human health.

“It’s only now in 2017 that a lot of these things are kicking in,” says Eric Topol, professor of genomics at Scripps Research Institute in La Jolla, CA. Researchers are just beginning to figure out what aspects of physiology can be measured with device worn on or carried with the body and how those measurements relate to health, he says. And they’re busy developing new devices, from glucose-monitoring contact lenses to sweat sensors that may provide whole new levels of data. “Eventually, wearable sensors will have a significant role in medicine,” Topol says, “but up till now, it’s been pretty marginal as to what kind of technology has been available.”

In a study designed to show the kind of information that could be garnered from wearables, Stanford geneticist Michael Snyder provided the 43 participants with a Basis watch, which measured heart rate, skin temperature, and steps, and which used software and an accelerometer to decide whether the wearer was walking, running, or biking. People taking airplane trips were asked to wear a pulse-ox monitor on their fingers. Snyder himself wore up to seven devices at a time, including monitors for galvanic skin response and radiation exposure.

All participants showed a drop in blood oxygen levels while flying, a phenomenon that had already been reported in the literature. But Snyder discovered that the lower levels correlated with fatigue. The study, published in PLoS Biology in January, also found an elevated heart rate in four people who reported feeling ill and turned out to have high levels of C-reactive protein in their blood, suggesting that the watch could detect an inflammatory illness (Li et al., PLoS Biol. 15, e2001402).

While none of these findings are definitive, Snyder says they at least suggest some of the information that can be learned from monitoring people with wearable sensors. “Minimally, I’m pretty sure we’re going to be able to detect illness pretty early.”

Mobile Studies

Wearables will not only give individual users feedback about the state of their health, but they will also provide scientists with an unprecedented source of data about the day-to-day vital signs of potentially millions of people. The Precision Medicine Initiative, a project that will sequence the genomes of one million people to study their health, includes the Participants Technologies Center, headed by Topol, which will offer people in the study sensors to track their vital signs. That part of the initiative has a $207 million, five-year federal grant from the National Institutes of Health to study the use of mobile technologies on 350,000 of the 1,000,000 enrollees. The sensors might include a wristband that can measure blood pressure.

Such a band will provide data that’s never been available before, says Topol. “We don’t even know what normal blood pressure is because we’ve never measured it in the wild in a large number of people, including while they’re sleeping,” he says. “And now, we have technology that can do that.”

There have been disputes recently about what target blood pressure should be—partly, Topol says, because there’s a lack of information. As for the data that does exist, “Those are based on one-off measurements in a doctor’s office. That’s a pretty contrived setting. We don’t even know what that means.”

Though fitness wearables have mostly concentrated on movement and heart rate, with attempts at monitoring sleep, there are plenty of other signs that can be measured. Continuous glucose monitoring has been available for some time, and groups such as Verily Life Sciences, a company owned by Google parent company Alphabet, are working on contact lenses can could measure glucose for diabetics. Last year, the FDA approved a device from AliveCor, of Mountain View, CA, which takes an electrocardiogram to look for signs of atrial fibrillation. Groups such as Beyond Verbal Communication, a company in Tel Aviv, are developing phone apps that listen to people’s voices for signs of depression. Changes in tone or the number of times a person sighs may indicate depression, as might other signs that can be measured by a phone, such as whether the owner leaves the house much or how often she talks to other people.

Phoning It in

A device need not be attached to the body to count as a wearable. One of the most powerful devices for personalized sensing is the smartphone, which comes with an array of sensors, as well as the processing power and communications capabilities that allow data to be collected, analyzed, and shared—both among an individual’s devices and with researchers. Ray Dorsey, a neurologist at the University of Rochester Medical Center in upstate New York, worked with developers at Sage Bionetworks to develop mPower, an iPhone app designed to assess the movement of people with Parkinson’s Disease and compare it to those without the illness. The app asks users to tap in a circle on the phone’s screen for 10 s to assess their motor control, say “ah” into the microphone to see how the disease may be affecting their vocal muscles, and walk a short distance to measure their gait and balance. The researchers can easily distinguish between people with and without Parkinson’s, Dorsey says, and see the changes in movement before and after taking their medication or when they exercise.

The readings provide more objective and much more frequent measurements than having a physician watch the same behaviors during an office visit and rate them on a scale, Dorsey says. “In a traditional clinical trial, I might see somebody every month for an hour in an artificial setting at an arbitrary time,” Dorsey says. “For the 99 percent of the time that they’re outside the clinic, I have no idea how they’re actually doing.”

Neurologists would like to have objective measures of diseases such as Parkinson’s, which don’t now exist. “We don’t have blood glucose levels, we don’t have cholesterol levels, we don’t have CD4 counts, we don’t have viral loads to measure,” Dorsey says. Wearable sensors might provide what’s needed.

Smartphones may also supplement other measurements. Answer ALS, a research partnership trying to find treatments for amyotrophic lateral sclerosis, plans to collect approximately 6,000,000,000 data points from each of 1,000 patients in a study, including sequencing their genomes, proteomes, and other “omics.” The project is rolling out a smartphone app that asks participants to perform various tasks such as tracing an object on the screen to assess motor function. They’ll be asked to speak for as long as they can with one breath because breathing ability is the best predictor of survival in ALS.

One question will be whether patients with different genetic profiles present differently in the clinic, which is not now known, says Jeffrey Rothstein, a neurologist at Johns Hopkins University and director of Answer ALS. Gathering daily measurements with a phone might turn up nuances that aren’t readily apparent in clinic visits four months apart. In one experiment, researchers will show patients a picture of a scene and ask them to describe it. IBM’s Watson computer will perform linguistic analysis on the responses to see if it can pick up any signs of cognitive decline. “Having that information is a completely new dimension for this disease,” Rothstein says. “And to be able to do this slowly over time—having a patient periodically being tested at home—is completely novel.”

Smartphone apps may also recruit many more people into clinical trials. On the first day mPower was released in March 2015, roughly 2,000 people signed up. Eventually, about 15,000 people enrolled in the study, of whom nearly 15% had Parkinson’s. Having objective, high-frequency data should make studies more efficient. “You can conduct clinical trials with smaller numbers of people in shorter periods of time and with greater confidence evaluate the efficacy or lack thereof of novel therapeutics, and that has clear economic value,” Dorsey says.

Phones may also let a wider segment of the population participate in clinical research, Dorsey says, which could improve the science. “The vast majority of clinical trials are done among people who are generally white, who are generally well educated, who generally have substantial means to come to an urban center,” he says. But large segments of the population own mobile phones, so people who can’t easily get to a medical center can still participate in a trial.

iPhone ownership tends to skew toward wealthier people, and mPower and other apps were based on ResearchKit, a software framework created by Apple. “Obviously, we don’t want to have clinical tools and treatments optimized for only people who can afford iPhones,” says Deborah Estrin, director of the Small Data Lab at Cornell Tech, a technology-focused graduate education campus of Cornell University based in New York City. Her lab developed ResearchStack, a framework similar to ResearchKit but designed for Android phones.

Estrin says phones have enormous potential for healthcare, in part because so many people already have them—95% of Americans, according to a recent survey by the Pew Research Center. “These are not new devices that we’re asking people to use for a particular reason. The power of the phone is that people have it with them almost all the time,” she says. That makes constant monitoring easier.

If researchers and clinicians are going to use the data generated by wearable sensors, they’ll need to develop digital biomarkers, Estrin says, to understand what their measurements mean. “You still have to do the work to figure out how you interpret the data that comes off the phone in a way that actually serves as a proxy measure,” she says. That won’t require the years of clinical trials it might take to prove that a drug is effective, but such proxies still need clinical validation, she says. Dorsey agrees; he’s the editor of a new journal, Digital Biomarkers, being launched by Karger Publishers this year.

Turning to Chemistry

While today’s wearables mostly rely on mechanical or electrical measurements of physiology, they may soon be joined by chemical sensors. Sam Emaminejad, a professor of electrical engineering at the University of California, Los Angeles, is developing a sweat sensor. “Sweat is a very complex sample,” he says. “It has a lot of proteins, metabolites, electrolytes—all this molecular-level information.” His focus is on building wearable sensors that can collect and measure those substances, providing data that will then let biologists figure out what the measurements mean in terms of health and whether relating those readings to the types of information being collected by other wearables can give a clearer picture of an individual’s wellbeing.

In a paper published in Proceedings of the National Academy of Sciences in April, Emaminejad and his colleagues showed they could use a wearable sweat sensor to detect high levels of chloride ions that identified patients as having cystic fibrosis (Emaminejad et al., Proc. Natl. Acad. Sci. USA, 114, 4625–4630). They were also able to measure glucose levels, which could be useful in monitoring diabetics and pre-diabetics.

Emaminejad would also like to develop sensors that could gather data from saliva or from the interstitial fluids between cells. He even imagines a sensor that could reside in a toilet and send a urinalysis to a phone. It might be possible to develop a simple sensor—to measure hydration or stress during exercise—in as little as three to five years, Emaminejad says. More sophisticated devices may take a decade.

The potential of wearable health monitors to change our understanding of human health is vast, Topol argues. “Big data will be able to eventually re-characterize the human species’ physiology,” he says. “Just like we map traffic through people’s phones today, we’ll be mapping the physiology of humans through their sensors.”

Sign up for our ALS Alert Newsletter
Sign Up