Consumer wearables have transformed the ability to monitor sleep over weeks, months and even years. Yet a surprisingly basic problem remains: two devices may report a measure called “total sleep time” while defining, calculating or quality-controlling it differently.
A multinational, multidisciplinary team of sleep scientists, clinicians, and engineers from academia and industry convened by World Sleep Society has developed a specification designed to address this problem. Published in Sleep Medicine, the proposed framework establishes common definitions and data standards for a core set of Fundamental Sleep Measures (FSMs) generated by consumer health trackers.
The goal is not to make every wearable work in the same way. Manufacturers would remain free to choose their sensors, algorithms, machine-learning approaches, and consumer-facing interfaces. Instead, the specification standardises what key sleep measurements mean when they are exported for clinical, research or public-health use.
The same name should mean the same thing
Measures such as total sleep time, sleep efficiency, wake after sleep onset and sleep regularity are already widely reported by wearable devices. However, identically named measures may not refer to the same construct across manufacturers, limiting their comparability and usefulness in healthcare and research.
The new specification addresses this by defining a common semantic framework for fundamental measures including sleep onset and offset, the main sleep period, sleep duration, wake after sleep onset, sleep regularity, 24-hour sleep including naps, and subjective sleep satisfaction. It also introduces explicit rules for handling missing or uncertain data.
Professor Michael Chee, Director at the Centre for Sleep and Cognition, NUS Yong Loo Lin School of Medicine (NUS Medicine), who led the team, said, “People naturally assume that a measurement called ‘Total Sleep Time’ means the same thing regardless of which device produced it. In reality, manufacturers often use different definitions, algorithms, and quality controls.”
“Without common definitions, it is difficult for clinicians, researchers and healthcare systems to know whether they are comparing like with like,” added Professor Sean Drummond, School of Psychological Sciences, Turner Institute for Brain and Mental Health, Monash University, co-chair of the specification team.
From interesting data to useful health information
Wearables have a major advantage over traditional sleep assessments: they can characterise sleep repeatedly in everyday life and over long periods. But exploiting that information clinically requires more than accurate sensors.
Healthcare systems depend on standardised measurements. Blood pressure, laboratory tests and other clinical measurements are useful partly because clinicians understand what a reported value represents. Wearable sleep data currently lack an equivalent level of consistency. This becomes particularly important as consumer-generated health information begins to enter electronic health records (EHRs).
The specification therefore focuses on the underlying data architecture required for interoperability. It defines not only sleep measures but also their metadata, provenance, quality indicators, time boundaries, and rules for dealing with incomplete recordings. Its objective is to allow data produced by different devices to be interpreted consistently by EHRs, research databases and population-health systems.
Focusing on measures people can understand—and act upon
The framework deliberately emphasises fundamental and interpretable aspects of sleep rather than proprietary scores or detailed estimates of sleep stages.
This distinction matters clinically. A person can take practical steps to provide adequate opportunity for sleep, adjust when they attempt to sleep, or establish a more regular schedule. By contrast, consumers have much less direct control over proprietary measures such as a “deep sleep score,” whose definitions and algorithms can vary among devices.
The specification also distinguishes behavioural intention from physiology. A person’s decision to try to sleep is not the same thing as the physiological onset of sleep, and the framework requires those concepts to be represented separately.
Similarly, it recommends replacing ambiguous terminology where necessary. Rather than using “Sleep Efficiency” when a wearable cannot independently establish that a person was actually in bed attempting to sleep, the specification defines a reproducible “Sleep Percentage” based on sleep occurring between physiological sleep onset and offset.
Standards without suppressing innovation
An important principle of the project is that interoperability should not require manufacturers to disclose proprietary algorithms or build identical products.
The specification does not prescribe which sensors should be used, how algorithms should be constructed or made public, how sleep stages should be classified, or what manufacturers should display in their apps. Instead, it establishes common backend meanings for a minimum set of measures. In other words, manufacturers can continue to compete on hardware, algorithms, design and user experience while still communicating fundamental sleep information in a common language.
Built jointly by academic, clinical, and industry experts
The specification, developed by a 16-member international team of academic scientists working with industry counterparts, built consensus on multiple complex issues over months of iterative discussions.
The cooperation of industry scientists from five companies that compete in the marketplace is unprecedented in this field, and was invaluable in ensuring the specification is not merely an academic exercise but a practical framework that can be readily incorporated into products.
“The collaboration demonstrates the value of the World Sleep Society mission, convening diverse stakeholders across sleep science and medicine to turn emerging challenges into opportunities,” noted Allan O’Bryan, Executive Director of the World Sleep Society.
Building the infrastructure for longitudinal sleep health
The authors see standardisation as an important step toward a larger goal: moving wearable sleep information from isolated proprietary ecosystems toward data that can be meaningfully used across healthcare and public-health systems.
Long-term wearable data could eventually allow clinicians to consider not simply how someone slept last night, but whether they routinely obtain adequate sleep, maintain a stable schedule, or show sustained changes in sleep patterns alongside changes in health. For such applications to become scientifically credible, however, a basic requirement comes first: everyone must mean the same thing when they report the same sleep measure.
“Wearables have given us an unprecedented ability to observe sleep in everyday life. Standardisation is what can turn those observations into information that clinicians, researchers and health systems can reliably understand and use,” Prof Chee concluded.
