Given the proliferation of smartwatches, activity bands, and other health wearables, it’s probably safe to assume that wearable devices have reached the coveted tipping point among consumers eager for health insights. In the last two years, adoption of health wearables in the United States has more than doubled, and more than 75 million Americans will use an activity tracker by 2021, yet only 15 percent of physicians say they’ve discussed wearables or health apps with patients.
This raises the question of whether those tasked with analyzing the data (i.e., physicians) are truly ready for the wearable revolution. As consumers continue to don their tech-savvy devices, should wearable data flow into the electronic health record (EHR) for physician analysis?
We believe that every bit of data is really valuable—particularly data that reflects the life of a person
– Florence Comite, MD
Telling the patient’s story through data
“We believe that every bit of data is really valuable—particularly data that reflects the life of a person,” says Florence Comite, MD, endocrinologist at Comite Center for Precision Medicine in New York City. “To us, wearables are almost like magic,” she says, referring to each device’s ability to continuously capture data related to heart rate, sleep, glucose levels, and more. This contrasts with a one-time blood pressure reading or single blood draw that doesn’t necessarily indicate a person’s true health status, she adds.
Health data from half a dozen wearable devices flows into the practice’s EHR where Comite analyzes it. She looks for overall trends and correlations between the wearable data and other information (e.g., labs, self-reported symptoms, and tests). In some cases, data flows automatically into the EHR through an application program interface. In other cases, patients email or print their data so Comite can scan it into the record.
Having access to wearable data allows her to tailor interventions for individuals with diabetes and other chronic conditions more effectively, oftentimes reversing these conditions entirely. She also uses the data to predict future health complications and set patients on a healthy trajectory. To her, health data tells a story. “I know how much I’m helping people by taking this data into consideration and interpreting it,” she adds.
Tackling the issue of data accuracy
However, not everyone agrees that wearable data belongs in the EHR—at least not yet. Consider Hawley Montgomery-Downs, PhD, an associate professor of psychology at West Virginia University and researcher who says she was initially optimistic about wearable devices and their ability to monitor sleep.
“When it came out, I was very excited about the potential that it offered,” she says. “If it worked well and gave us the information that we needed, then my opinion was that it was going to revolutionize sleep science.”
However, Montgomery-Downs says intentions for accuracy don’t necessarily translate into reality.
“As a scientist, I look at the marketing, smile, and nod,” she says. “Then I take the device into my laboratory and subject it to all of the rigors of evaluation compared to the gold standard, which is using electrodes and sensors to very precisely determine when someone is awake or asleep and what stage of sleep they’re in.”
Her research found that although some wearables were highly accurate at identifying when someone was asleep, they were just 50% accurate at identifying when that same individual was awake. “These devices tell you you’re sleeping better and more than you actually do,” she says. “And the worse you’re sleeping, the bigger the difference between reality and what the device is telling you.”
For example, they don’t record relatively brief awakenings, says Montgomery-Downs. This makes it difficult to correctly diagnose someone with sleep apnea who wakes up for 10-15 seconds every couple of minutes. Likewise, if someone with insomnia wakes up every 20 minutes but remains motionless, the device will assume that the individual is still asleep. Troublingly, this could give patients and physicians a false sense of wellness, she adds.
Studies show that the inaccuracies don’t just apply to sleep health data. For example, researchers at Stanford found that seven popular wearable fitness trackers had an error rate of more than 20% when it came to measuring calories burned.
It’s not going to get better unless we insist on having high-quality accurate devices
– Hawley Montgomery-Downs, PhD
If consumers continue to purchase wearable devices that are currently available on the market, she fears there will be no incentive for companies to improve their products. “It’s not going to get better unless we insist on having high-quality accurate devices,” she says. “If physicians, in particular, start using these non-medically approved devices in their evaluation and diagnosis, that’s just another nail in the coffin of ever insisting that companies get it right.”
Comite isn’t fazed by the potential inaccuracy of wearable data because she says that much of the data in an EHR isn’t precise. “Even lab data is inherently imprecise,” she says. “The hope is that any data that goes into the record will be screened carefully. It’s in the eyes of the interpreter more than anything else.”