The wearable technology market was a $20 billion dollar market in 2015 and is on pace to reach $70 billion by 2025.1 Whether glasses, shirts, watches, footwear, or skin patches, the largest sector for this market is healthcare. Multiple factors are driving this growth, but we may be witnessing the powerful convergence of new technology, improved health awareness, rising healthcare costs, and decreased access to face-to-face care. As the general population grows up with personal virtual connectivity, the comfort level for medical virtual connectivity is rising as well.
The use of wearable devices in sleep medicine is not a new concept. In fact, the utility of wrist actigraphy in patients with circadian rhythm disorders and insomnia is well established in the sleep literature. Wrist actigraphy does not, however, give feedback to the user, but rather provides useful information to the provider for both diagnostic and treatment purposes. The modern wrist actigraph now is embodied in relatively inexpensive devices such as Fitbit, Jawbone UP, among others. The new concept in these modern devices is the actionable feedback given to the user regarding their sleep behaviors. Already there are many patients who open up their smartphone apps to show their doctors their data, in order to get assistance deciphering the reams of data provided by such devices. Moreover, some devices go too far in telling users about “deep stage” versus “light stage” sleep, as sleep stages are determined only by electroencephalographic information, something not provided by wrist-worn wearables.
Various studies over the past few years have addressed the accuracy of fitness/activity trackers compared to research-grade counterparts.2, 3 In general, fitness trackers are more accurate at some measures than others, and there is variability in the accuracy among different trackers. For example, the original Fitbit overestimated sleep by >1 hour and underestimated calories burned, but was extremely accurate in counting steps, even at vigorous levels of exercise.2
Researchers from the Sansom Institute at the University of South Australia took 21 (healthy, ambulatory >18) subjects and had them wear 7 different consumer devices and 2 different medical actigraph watches for 48 hours, under free-living conditions.3 Concurrent measurements allowed the researchers to compare each device to the medical device, but also to compare the consumer devices directly to each other. The devices chosen were the Fitbit One*, Fitbit Zip, Jawbone UP*, Misfit Shine*, Nike Fuelband, Striiv Smart Pedometer, and the Withings Pulse*. All 7 measured physical activity parameters, four (marked with *) measured sleep parameters.
The research grade devices selected were the BodyMedia SenseWear Model MF (validated for total daily calories burned and total sleep time) and ActiGraph GT3X+ (validated for step count and moderate-vigorous physical activity). The researchers compared the consumer wearables to their research grade counterparts in terms of the following parameters: steps, sleep duration, total daily energy expended (TDEE), i.e. calories burned, and moderate-vigorous physical activity (MVPA).
The strongest performers across all categories were the Fitbit One, Fitbit Zip, and Withings Pulse. Price did not reflect product validity, since the most expensive device, the Nike Fuelband, was one of the worst performers. The devices with the highest correlation coefficients for measuring sleep duration were the Fitbit One and the Pulse, 0.92 for both. Of note, all 4 devices over-estimated amount of sleep compared to the reference device, from 22-47 minutes.
What does this mean for those who are using fitness trackers to measure sleep? The current devices on the market are: 1) able to distinguish sleep from wakefulness better than they are able to distinguish among the different sleep stages; 2) tend to overestimate sleep quantity when compared to an external reference device, but 3) may be helpful in gauging and modifying one’s own behavior. In the author’s view, this makes current wearable devices useful much in the same way that peak flow meters are helpful to asthma patients. The wearer gains information that may result in changes in their behavior. The usefulness of self-help devices, whether worn or not, however, is optimized by interactive guidance from a healthcare provider.
Over the next ten years, as the wearable industry matures, the sensors that measure our physiology will be smaller, less invasive, and more versatile, and the data collected will be compiled and communicated quickly to both ourselves and our doctors. Polysomnography, the current gold standard of sleep testing, will not be exempt from these advancements. Sleep medicine, even as a younger specialty, has changed rapidly over the past ten years, with the number of sleep centers at one time expanding and now contracting, due to reduced reimbursements and perhaps to increased utilization of home sleep testing. We will be forced to shift our thinking and practice again with the next phase of improved wearable monitoring devices. But even with improved diagnostic tools at our fingertips, we will have only addressed one aspect of the sleep care bottleneck. Expanding diagnostic services to more people may be the easier hurdle to overcome; providing treatment will be the larger challenge, and one that may be solved only by patients actively participating in self-care. Wearable devices may play a key role in both sides of the care continuum.
1Harrop, P et al., Wearable technology 2015-2015: technologies, markets, forecasts, IDTechEx, Feb 2015.
2Takacs J, et al., Validation of Fitbit one activity monitor device during treadmill walking. J Sci Med Sport 2013; 17(5): 496-500.
3Ferguson, T et al., The validity of consumer-level, activity monitors in healthy adults worn in free-living conditions: a cross sectional study. Int J Behav Nut Phys Act. 2015; 12:42 DOI 10.1186/s12966-015-0201-9.