The wristband tracking devices worn by millions of fitness-conscious Americans do an adequate job of measuring heart rate, but not of estimating energy expenditure, reports a study published online May 24 in the Journal of Personalized Medicine.
Although six of seven tested devices had a median error for heart rate below the acceptable threshold of 5% during a cycle ergometer task, none achieved an error below 20% for energy expenditure across several tasks. Even the most accurate device was off by an average of 27.4%, and the least accurate erred by 92.6%.
The results suggest the need for caution in using energy outlay readings from these units to guide health improvement and weight loss programs, say researchers from Stanford University in Palo Alto, California, and the Swedish School of Sport and Health Sciences in Stockholm, who laboratory tested the accuracy of seven commercial wristband devices. The team, led by Euan Ashley, DPhil, FRCP, professor of cardiovascular medicine, of genetics, and of biomedical data science at Stanford, calls for reference standards for validating consumer health devices and are asking US users to share their device-collected numbers with the research community by uploading them to a Stanford data hub.
Some device users already share their numbers with their physicians, and it has been proposed that measurements from wearable commercial technology be integrated into medical records. “People are basing life decisions on the data provided by these devices,” Dr Ashley said in a Stanford news release. However, the consumer devices do not meet the same standards as their medical-grade counterpart. “It’s hard for doctors to know what to make of heart-rate data and other data from a patient’s wearable device,” according to the news release.
For the current study, the researchers recruited volunteers age 18 y and older from Stanford University and amateur sports clubs. The 60 participants were selected for maximum demographic diversity in age, height, weight, body mass index, wrist circumference, skin tone, and fitness level. There were 29 men with a median age of 40 years and 31 women with a median age of 37 years.
The seven devices tested were Apple Watch, Basis Peak, Fitbit Surge, Microsoft Band, MIO Alpha 2, PulseOn, and Samsung Gear S2.
Participants performed a total of 80 tests, 40 each in two phases. Tasks included sitting, stationary bike cycling, and treadmill walking and running. Two investigators independently rated skin tone at the wrist, which can affect tracking sensors. Subjects wore the units during laboratory assessment with continuous telemetry and indirect calorimetry.
Overall, the lowest device error occurred for the cycle ergometer test (median 1.8%; 95% confidence interval [CI], 0.9% – 2.7%), and the highest for the walking task (median, 5.5%; 95% CI, 3.9 – 7.1%). Error was higher for men, those with greater body mass index, and those with darker skin tone.
When analyzed by unit type, the researchers found that three units achieved a median error rate at or below 5% for heart rate during the walking task: Apple Watch at 2.5% (95% CI, 1.1% – 3.9%), PulseOn at 4.9% (95% CI, 1.4 – 8.6%), and Microsoft Band at 5.6% (95% CI, 4.9% – 6.3%). In the remaining four devices, the median error ranged from 6.5% to 8.8%.
Across all activity modes, the Apple Watch achieved the lowest overall error (2.0%; 95% CI, 1.2% – 2.8%) for heart rate, whereas Samsung Gear S2 had the highest error (6.8%; 95% CI, 4.6% – 9.0%).
In measuring energy outlay, the Fitbit Surge had the lowest median error rate across tasks (27.4%; 95% CI, 24.0% – 30.8%), and PulseOn had the highest (92.6%; 95% CI, 87.5% – 97.7%).
“The heart rate measurements performed far better than we expected, but the energy expenditure measures were way off the mark,” Dr Ashley said in the release. “The magnitude of just how bad they were surprised me.”
The researchers are not sure why the energy measurements are so inaccurate, but it may be that each device has its own proprietary algorithm for calculating energy expenditure, and these algorithms are not well matched to individuals.
“My take on this is that it’s very hard to train an algorithm that would be accurate across a wide variety of people because energy expenditure is variable based on someone’s fitness level, height and weight, etc.,” said coauthor Anna Shcherbina, a Stanford graduate student, in the news release. Although heart rate is measured directly, energy expenditure is measured indirectly through surrogate calculations.
The findings on energy parallel that of a smaller study, published in 2016, which reported that energy expenditure estimates deviated by as much as 43% from the reference standard. Both studies cast doubt on the units’ utility for fostering weight loss.
Similarly, a study reported last year by Medscape Medical News found that wearable technology tracking exercise and diet does not improve weight loss.
“Individuals and practitioners should be aware of the strengths and limitations of consumer devices that measure heart rate and estimate energy expenditure,” Dr Ashley and colleagues conclude. “We encourage transparency from device companies and consistent release of validation data to facilitate the integration of such data into clinical care.”
The take-home message, according to Dr Ashley, is that consumers can rely on fitness tracker’s heart rate measurements. “But basing the number of doughnuts you eat on how many calories your device says you burned is a really bad idea,” the news release states.
This study was supported by Stanford University’s Department of Medicine, Department of Genetics, and Department of Biomedical Data Science. The authors have disclosed no relevant financial relationships.
J Pers Med. Published online May 24, 2017. Full text
For more news, join us on Facebook and Twitter
Tidak ada komentar:
Posting Komentar