Wearables Under the Microscope: How New Tech Will Disrupt Patient-Physician Dynamics
There’s no evidence yet that smartwatches and gadgets will improve or save lives, but some physicians see other benefits.
Earlier this year a young woman in her twenties turned up at the California-based practice of Jacqueline Shreibati, MD, with a runaway heartbeat. “Every [other] doctor had said that she was fine, it was just anxiety, but she kept feeling her heart racing,” Shreibati said, recalling that day. “She was tracking it with a heart rate sensor on a [watch]. Her doctors kept telling her, ‘Please stop focusing on your heart rate, let’s focus on treating your anxiety.’ She was persistent—she said no, there’s something wrong.”
Shreibati, an acknowledged advocate of wearable technology—in addition to her work at San Mateo Medical Center, she is also the Chief Medical Officer for AliveCor—continued: “She was having these abrupt increases in her heart rate with changes in position from sitting to standing, and there was no corresponding drop in her blood pressure.” Based on data from the woman’s watch Shreibati diagnosed and then confirmed postural orthostatic tachycardia syndrome, or POTS, a condition often missed in young women. “She provided the information for us, basically right there in that clinic visit. . . . It was one of those very satisfying moments in which the patient did 75% of the work in terms of her diagnosis.”
While Shreibati says she would have found POTS eventually, she attributes the rapid diagnosis to the instant availability of data from the watch.
Welcome to the wearables boom—think Apple Watch, or AliveCor’s KardiaBand, or Omron Healthcare’s blood pressure monitor. Wearables are changing the patient-provider relationship and creating new challenges for health data management. Despite the lack of clinical trial evidence supporting a cardiovascular benefit for these devices, some physicians already find them to be useful and have begun building them into their practice. Others are far more cautious, fearing that the hype far exceeds value and that too many of their colleagues are diving in too soon.
Caution and Skepticism
Every clinician who spoke with TCTMD supported the idea of patients taking more control of their health. But some had reservations about how well wearables actually facilitate this goal.
Mamas A. Mamas, BMBCh, DPhil (Keele University, Stoke-on-Trent, England), offered a cautionary tale, noting that he himself wears an Apple Watch when he goes for a jog. “I was running once,” he said, “and my heart rate shot up to 180, and I thought, ‘What the hell is going on here?’” Mamas determined that his pulse rate was actually 100 and did not seek medical attention. But he suspects that somebody else would have rushed to a clinic, needlessly worried.
Indeed, Mamas said he “often” has young patients referred to him whose heart rates have jumped up while exercising. In these cases, he continued, “What do you do? Do you say to these patients, ‘Oh, it’s likely to be nothing,’ or do you investigate?” Mamas came down on the side of investigating further since the patient has sought his counsel, but added that in most cases this amounts to “unnecessary stress” as younger individuals are at very low risk of developing A-fib.
“The wearable depends on the person actually wearing it and charging it and using it, so that for some people is either a finance or convenience impediment,” Paul A. Friedman, MD (Mayo Clinic, Rochester, MN), pointed out in an interview. The devices do represent a “huge advance,” Friedman stressed. “I’m not in any way discrediting that functionality or capability, but I want to highlight that with one test you’re at the hospital with your doctors and in 10 seconds you get an answer. With the other one, you have to buy something and then you have to wear it and recharge it.”
In addition to the convenience concerns Friedman raised, other clinicians said they fear that their colleagues have begun to embrace wearables too quickly.
“There are very prominent physicians that are lending their credibility” to widespread deployment of wearables, Venkatesh Murthy, MD (University of Michigan, Ann Arbor, MI), observed to TCTMD. “We have almost no evidence of benefit, and they’re sold directly to consumers.”
We have almost no evidence of benefit, and they’re sold directly to consumers. Venkatesh Murthy
The Apple Heart Study, results of which were released at the American College of Cardiology (ACC) 2019 Scientific Session, represents the most earnest assessment of wearables’ clinical utility to date. Of more than 400,000 participants in this observational trial, 2,161 (0.52%) received an irregular rhythm alert while wearing an Apple Watch. A subset of this group received an ECG patch, and an even smaller group returned those patches for analysis. A-fib was detected in 34% of the people remaining in this final analysis.
Apple, Murthy stated, “ended up essentially recruiting whoever declared interest. There wasn’t really a lot of filtering for ‘Who are the people where we are most likely to see a signal?’ This mindset of ‘We’re just going to go big and let the numbers sort it out’ is a change from what we’ve done before.”
Other cardiologists have raised similar concerns. As reported by TCTMD at ACC 2019, session panelist Patrick O’Gara, MD (Brigham and Women’s Hospital, Boston, MA), responded to the study results by saying: “When you’re designing a trial where one of your goals is to see if you can scale things up to these massive numbers of patients, . . . how do you weigh the tension between pragmatism, scalability, and precision? It seems to me that there is a great enough deal of hype around this kind of technology.”
Murthy contrasted what’s happening with wearables while the evidence base is still thin to the slower rollout of statins, whose use was initially restricted to a small group of patients and then expanded as more evidence of efficacy emerged.
“Executives at Apple have said this is the ‘ultimate guardian of your health.’ Simple phrases like that, which are not supported by data, are problematic,” Murthy said.
The upcoming HEARTLINE trial—which will assess whether A-fib alerts from the Apple Watch have a measurable effect on MI, stroke, and mortality—aims to provide just such evidence. Unlike the Apple Heart Study, HEARTLINE will be an RCT. And while all comers were welcome in the Apple Heart Study, HEARTLINE will be restricted to participants aged 65 and older.
“One has to really be optimistic that the questions will be answered,” Murthy stated, “but I also think it’s not too much to ask that the marketing be constrained to the evidence” in the meantime.
“Wearables are the future of medicine, no doubt about it,” Mamas emphasized. “But they should be targeted to the right patients.” He argued to TCTMD that wearables should be “classed as a medical device” and regulated accordingly.
The Changing Patient-Provider Relationship
While monitoring for atrial fibrillation has received the most attention so far, Shreibati pointed out that “there’s a lot of development on the clinical application of looking at the pulse” for early detection of diabetes, sleep apnea, or high blood pressure. Indeed, Shreibati said the “holy grail” is a wrist or ear wearable that can predict hypertension much earlier than can be done today.
There are also different ways that people could depend on their wearables, Shreibati noted. “There are folks who are identifying conditions either when they don’t have symptoms or [when] they do have symptoms but a doctor is not involved. . . . They’re doing their own investigations outside of the touch points of their physician.” And then, in another group, there are patients “who have a relationship with their doctor and they’re using wearables to monitor or track health conditions in between clinic visits,” she continued.
That second scenario is familiar to many physicians: patients who trust the expertise of their clinicians and keep them apprised between visits. It’s more of a partnership than in earlier days, in which doctors ordered the tests and had all the equipment, but the main terms of engagement have not changed.
Wearables are the future of medicine, no doubt about it, but they should be targeted to the right patients. Mamas A. Mamas
The first scenario, however, hints at a more seismic shift. A patient can change her diet based on a glucose alert, or decide to stop exercising as vigorously based on an alert from a watch.
C. Michael Gibson, MD (Beth Israel Deaconess Medical Center, Boston, MA), is leading the HEARTLINE study, which aims to recruit 180,000 patients. He predicted fundamental changes in the patient-provider relationship are just around the bend. “The old world was a system where the doctor was a trusted provider. They conveyed in a paternalistic fashion proscriptive information to a relatively uninformed patient,” Gibson observed to TCTMD. “Now the doctor has to earn the trust of their patients, and there’s a conversation.”
For his part, Mamas shared how he frets about the “worried well”—healthy patients who become anxious without cause based on a spurious alert from their wearable, resulting in personal stress and sometimes straining health system resources. To the extent that wearables increase unnecessary visits, this too would change the patient-provider relationship.
Mintu Turakhia, MD (Stanford University, CA), co-principal investigator of the Apple Watch Study, has a ready response to such concerns: “What surprises me about that argument is when we give out a blood pressure cuff we don’t worry about the ‘worried well’ calling us when they have a single reading of 180.” But on this point, Mamas countered that blood pressure cuffs are only given to people believed to have an elevated blood pressure—unlike a watch that anyone with the interest and means can buy.
The Data Deluge
Even if clinical trials show that alerts from wearable devices lead to less disease and death, the challenge of how to implement wearable data into everyday practice remains formidable.
“We have no predicate models for how to make a clinical decision” from patient-generated data, noted Paddy Barrett, MD (Blackrock Clinic, Dublin, Ireland). Moreover, Barrett emphasized to TCTMD, existing health ecosystems “are not designed to accept all of this inbound information.”
Electronic health records (EHRs) are essentially walled gardens, with all diagnostics stored within those walls. “Inbound data” from a patient’s watch is not part of that vision, although Apple aims to bridge this gap with its Health Records product. Besides Apple, the health analytics startup Health Catalyst had its initial public offering (IPO) on July 25, 2019. That IPO may speak to the ferment in this space, but as of today there are no clear protocols for widespread ingestion of patient-generated data into medical record systems.
Imagine a hypothetical scenario in which patient-generated data port seamlessly and with absolute security into any EHR in the world. Even in that case, Barrett highlighted a key conceptual error underlying the concept of personalized digital medicine: the notion that mere provision of this information will be enough in and of itself to stimulate behavior change by patients. He noted that behavioral economists or psychologists could have told clinicians quite early on that the story is not so simple.
Turakhia concurred, saying: “No amount of technology thus far has succeeded at scale in getting people to sustain healthy lifestyles and behaviors.”
For busy practitioners, a large amount of incoming data could drown a practice. Barrett observed that for a “very overwhelmed physician” who has just 8 minutes to spend with a patient, it would be very hard to winnow down a large volume of raw data “without contextual significance or how that might move someone through an algorithm of care.” The crucial challenge now, Barrett said, is to present this information in a way that is meaningful and actionable for individual patients.
And the EHR may not be the best tool for this task. “The EHR is a hospital electronic health record, it is not the transactional history of everything that’s ever happened to a person or patient,” Turakhia said. “That data belongs to the person whose data was captured. So now you have to think about moving away from an EHR to a personal health record, where the person or patient has control of their data and has fairly easy ways to give that data to people who need it. We have to rethink the paradigm of this data distributed model, so that it’s the patient who is in charge,” he concluded.
Turakhia also pointed out that 25 years ago—in 1994—a patent was filed successfully for a wrist wrap to record an ECG. At that time, patients had to fax the record to their physicians, which may have limited uptake. While it’s easy to assume that wearables are a 21st century phenomenon, he argued that this patent shows a longstanding interest in such devices. The obvious difference today is that it much easier to transmit such data to physicians. But then as now, Turakhia said, “patients want their data.”
One solution might be to present the same data differently to patients and clinicians, Turakhia proposed. In his view the challenge boils down to effective user design, much like what has been done with weather service data. What a consumer sees is different than what a meteorologist observes, but in both cases the information is useful.
Another open question is who would be liable for diagnosis made by a medical wearable. Physicians can be sued for malpractice or have their licenses revoked, but how do you sue an algorithm? No one TCTMD approached had a definitive answer, but many agreed that in order for wearable data to truly become part of healthcare delivery, liability will have to be addressed.
Leslie Saxon, MD (University of Southern California, Los Angeles), acknowledged this challenge but didn’t seem particularly concerned about it. “Does it open up potential liability issues? Yeah—any disruption does,” Saxon stressed, saying she is much more worried that health systems will be too slow to embrace the potential of wearables for bureaucratic reasons. The Centers for Medicare & Medicaid Services (CMS), for example, has not yet developed codes to account for diagnostics by medical wearables.
An Exciting and Uncertain Future
The evidence base regarding the effectiveness for wearables continues to build, albeit more slowly than sales of the technology, estimated in the tens of billions of dollars.
Even if there were strong proof of the efficacy, as long as wearables remain fairly expensive—the most affordable Apple Watch is still almost $300—many people will be priced out. And if most of the data in wearable trials come from relatively affluent people, the results will not be representative of the health of the broader population anyway.
This is another puzzler for clinicians, even those active and interested in this space. Gibson noted that the HEARTLINE trial will offer people credits for completing tasks, which they can then redeem so that the cost of a watch becomes free. While this promises to increase accessibility of the devices in the HEARTLINE context, broader strategies for preventing a new healthcare digital divide from forming will be necessary.
Turakhia agreed this is an important consideration. “The pace of smartphone adoption has exceeded the adoption of almost any other device in America, such as the television or microwave,” he pointed out. Smart devices will remain at the center of many people’s lives, which in turn might translate into better access to wearables across all socioeconomic groups, Turakhia commented.
We have to rethink the paradigm of this data distributed model, so that it’s the patient who is in charge. Mintu Turakhia
There’s also the hope that studies like HEARTLINE will chart a new course for clinical trial conduct. “We’re going to cut a lot of the costs out of doing large randomized trials,” Gibson said. “It won’t be nurses or doctors enrolling patients, the patients will be enrolled using an app.” Gibson said it can cost as much as $50,000 to track a patient in a traditional clinical trial, an amount that he feels could drop substantially.
Gibson expressed confidence that the eventual data for HEARTLINE regarding MI, stroke, and mortality will be robust. “The outcomes will be ascertained from the CMS claims database. This is really important because no patients will go missing,” Gibson said. “This is more real-world patients [than the Apple Heart Study], with a broader range of risks, and that makes the results much more generalizable.”
Saxon noted that cardiologists have long relied on wearables in other forms, such as Holter monitors. “You think about the limitations of the traditional model . . . [of] a doctor prescribing a very expensive wearable for a very limited amount of time that may or may not result in a diagnosis,” she said.
The debate over wearables and their place in medicine—and society—is likely to continue.
At one point Gibson himself was a skeptic about the value of wearables until he heard from a relative who’d undergone ablation after being diagnosed with atrial fibrillation, he said. “Then he called me and said, ‘My pulse is feeling irregular, am I back in A-fib?’” Relying on data from the relative’s watch, Gibson determined that he was indeed experiencing a new episode and arranged for a successful cardioversion. “It kind of flipped me over to being a believer. Firsthand, I found it to be beneficial,” Gibson admitted.
Along similar lines, Saxon said she welcomes patient contact spurred by wearable data, calling it a “great trade-off in terms of my time because the level of conversation I’m having is so much higher.” Saxon estimated that on a given weekend she has 10 or so clinical consults with patients based on wearable data. These, she said, are much more satisfying than simply telling people to go to the emergency department and waiting for a report.
There’s no going back for Saxon, who said: “I’m telling everyone to get a watch.”
- Shreibati is the Chief Medical Officer for AliveCor.
- Turakhia led the Apple Heart Study and is the Executive Director of the Stanford University Center for Digital Health.
- Saxon is the Executive Director of the University of Southern California Center for Body Computing.