In 2017, Kelly Baron and fellow sleep clinicians in Chicago described a patient of 27 whose app said she was sleeping poorly. An overnight lab study found plenty of deep sleep, a larger share than average for her age. She still wanted to know why her tracker said otherwise.¹

Are sleep trackers accurate? Roughly, for telling sleep from wake, and poorly for sleep stages. A 2025 meta-analysismeta-analysis: A study that combines the results of earlier studies on the same question into one overall estimate. Pooling makes the estimate more precise, but it cannot repair the studies it pools: a meta-analysis of surveys is still survey evidence.Full entry in the glossary by researchers in South Korea, pooling 24 studies that compared wrist-worn trackers with a lab sleep recording, found the devices differed from the lab on total sleep, time to fall asleep and time awake, while the authors judged them still useful for tracking general sleep patterns.²

So read a tracker the way the evidence below suggests: trust the weekly pattern of when your nights start and end, doubt the minutes awake after a bad night, skip the nightly split of light, deep and REM sleep, and never take a sleep apnea feature that stays silent as an all-clear.

How a sleep tracker guesses when you are asleep

A sleep tracker never measures sleep directly; it infers it from the body’s side effects. Polysomnography, the lab reference test, records brain waves, eye movements and muscle tone overnight, while consumer trackers work from signals such as movement and heart rate, a 2019 review of wearable sleep technology by Massimiliano de Zambotti and colleagues explains. Three of its authors, de Zambotti among them, disclosed research funding from the tracker maker Fitbit Inc. for work unrelated to the review.³

The weak point follows from that design. People can lie awake in bed for long stretches without moving, and a movement-based algorithm then scores that time as sleep. Newer devices add heart rate and other signals to do better, but each company’s algorithm is private, and the review warns that algorithm updates may change how sleep measures are calculated.³

Picture a night like this. You wake at 3 a.m., lie still, run through tomorrow’s meeting and drift off again 40 minutes later. To a sensor on your wrist, a still body with a calm pulse looks much like a sleeping one, and the app may show an unbroken night.

The lesson is to read every tracker number as an estimate. The quieter you are while awake, the more of that wakefulness the device is likely to miss.

Are sleep trackers accurate? Good at sleep, weak at wake

Independent lab tests agree on the pattern: consumer trackers catch nearly all sleep but miss much of the time you spend awake during the night. The clearest single test is a 2021 US Navy study that ran seven devices side by side against polysomnography.

The study

Moderate evidence

A Navy sleep lab ran seven trackers at once, and waking was the weak spot

Researchers at the Naval Health Research Center in San Diego recorded 34 healthy young adults with polysomnography while they wore or slept beside a set of consumer devices. Scored in 30-second slices, every device caught at least 93 percent of the slices the lab called sleep, but only 18 to 54 percent of the slices the lab called wake. Most devices matched or beat actigraphy, the research-grade wrist monitor, and they did worse on the nights when sleep was disrupted. No device company funded the study or had any role in its design, analysis or writing.⁴

Why does the error lean one way? Each stretch of wake a device misses is scored as sleep, so its totals drift toward more sleep and less time awake than the lab recorded, the same bias the authors found in actigraphy. The main caveat is who was tested: healthy young people with no sleep complaints, sleeping in a lab on a fixed 8-hour schedule, and the results hold only for the device models and software tested.⁴

Two later reviews reach the same verdict. A 2026 meta-analysis from Portugal, limited to healthy adults, found that wearables overestimated total sleep, undercounted time awake during the night, and that no device was consistently ahead of the rest.⁵ A 2024 review requested by the Sleep Research Society concluded that the available evidence suggests consumer devices outperform traditional actigraphy, while naming misclassified wakefulness as a clear limit.⁶

So on a restless night, assume the total on your app is more likely too high than too low. The worse the night, the less the number deserves your trust.

Sleep stages are the least reliable number on your app

The light, deep and REM minutes that apps display are the weakest output of consumer trackers. In the 2021 Navy study, most devices failed to identify 30 to 50 percent of deep sleep and of REM sleep, and the stages they got wrong were most often called light sleep.⁴

1. Lab recording: polysomnography reads brain waves, eye movements and muscle tone
2. Tracker estimate: a guess from movement and heart rate
3. A short awakening while lying still: scored as light sleep on the tracker strip
4. Deep sleep: partly scored as light sleep, the most common mix-up for deep and REM sleep
5. Color key: coral is awake, pale is light sleep, purple is deep sleep, amber is REM sleep. Block lengths are illustrative, not real data

The reason is the same blind spot. A lab stages sleep from brain waves, which a wrist device cannot record; it has to infer stages from movement and, in newer models, heart signals.³ The 2024 review adds that person-specific factors such as skin color or obesity can affect how well the sensors work.⁶

Consider a typical breakfast. Your app shows about half your usual deep sleep after a stressful day, and you start planning an early night to win it back. Given the lab results, the dip may say as much about the device’s labeling as about your night, which is the gap the Chicago patient ran into. The nightly split between light, deep and REM sleep is the app’s guess, not a target to chase.

What a sleep tracker can tell you

Used for patterns rather than verdicts, a tracker can still tell you something real. The authors of the 2025 meta-analysis, despite finding significant differences from the lab on total sleep, time to fall asleep and time awake, concluded that wrist devices can still be useful for tracking general sleep patterns.²

Clinicians who see tracker users give similar advice. Baron and colleagues tell patients that the best use of a tracker is probably to watch their sleep pattern, including how much time they spend in bed, rather than the exact minutes awake and asleep. One of their patients, a man of 69, said watching his tracker data helped while he tapered off a sleeping pill under his doctors’ care, because it showed his sleep improving after a few days; that is one patient’s experience, not evidence that trackers help.¹

The American Academy of Sleep Medicine (AASM), the US professional body for sleep specialists, takes a similar middle line. Its 2018 position statement, still listed as current on its website in September 2026, says consumer sleep technology cannot be used to diagnose or treat sleep disorders, but may improve the conversation between patient and clinician when it is read within a proper clinical evaluation.⁷

A sample week shows the difference between a pattern and a verdict. Suppose your weekly average says 6 hours 50 minutes, your bedtime drifts an hour later on Fridays and Saturdays, and you feel groggy on Mondays. None of those nightly totals is exact, but the drift and the shortfall are patterns worth acting on, starting with the best-tested habits for sleeping better.

Five small habits keep your own data in proportion.

WiserHours

Reading your sleep data

• Look at weekly averages rather than single nights
• Watch when your nights start and end, and how long you spend in bed
• Skip the nightly split of light, deep and REM minutes
• Check the numbers against how you feel and function during the day
• Keep a written sleep diary if you plan to see a doctor about sleep

Find this checklist online:
https://wiserhours.com/sleep/sleep-tracker-accuracy/#checklist-sleep-tracker-accuracy-reading

When the numbers keep you up at night

Tracking can backfire. Baron’s 2017 report described patients so preoccupied with perfecting their tracker data that the data drove them to seek treatment, and the team named the pattern orthosomnia, from “ortho” (straight or correct) and “somnia” (sleep). All three patients were spending too long in bed trying to raise the sleep their device reported, which may have made their insomniainsomnia: Trouble falling asleep, staying asleep or waking too early despite having the chance to sleep, with effects on how you feel or function during the day. It is called chronic when it lasts for months; short spells after stress are common.Full entry in the glossary worse. The authors note they could not tell whether the trackers caused the problems.¹

A randomized experiment shows how feedback alone can color a day. In a 2018 study at the University of Oxford, 63 people with insomnia disorder wore a wrist device and were randomly given made-up feedback the next morning saying they had slept well or badly. By evening, the group told they had slept badly reported feeling less alert and more sleepy and tired than the group told they had slept well, though the two groups did not differ on a test of reaction speed.⁹ One of its authors, Colin Espie, co-founded the digital sleep-therapy company Big Health, as he disclosed in a 2017 Oxford trial report; the experiment tested feedback, not a product.¹⁰ The paper’s full text was out of reach, so only its abstract is used here, and the effect was measured over a single day.

Take a familiar morning. You wake feeling fine, open the app, see a poor score and spend the day feeling foggy. The feedback, not the night, may be doing some of that work. If checking your score makes you more anxious about tonight, judge the night by how you function during the day for a while, and consider whether the tracker is helping. Trouble sleeping that persists for months is a separate question: what insomnia is and when it needs a doctor.

Sleep apnea alerts can prompt a sleep test, never replace one

Some smartwatches, phone apps and bedside devices now offer to assess your risk of obstructive sleep apnea, and some of these features are cleared by the US Food and Drug Administration (FDA) for self-screening or risk assessment. The AASM’s January 2025 health advisory says that even the FDA-cleared ones vary in method and accuracy, most cannot tell obstructive sleep apnea from other breathing problems in sleep, and none can diagnose sleep apnea or confirm its absence.⁸

Myth

If my watch has never flagged sleep apnea, I don't have it.

Fact

A 2025 AASM advisory says self-screening apps and devices, including FDA-cleared ones, cannot diagnose sleep apnea or confirm its absence.

Diagnosis needs a sleep study. A 2017 AASM guideline strongly recommends that questionnaires, clinical tools and prediction algorithms not be used to diagnose obstructive sleep apnea in adults without polysomnography or a home sleep apnea test.¹¹ Symptoms carry more weight than any app: the US National Heart, Lung, and Blood Institute (NHLBI) counts pauses in breathing, loud snoring that happens often and gasping for air in sleep among the signs, and notes that you may not know you have them until someone tells you.¹²

The practical rule is lopsided, like the trackers themselves. An alert is a reason to book a doctor’s visit, while silence from the app proves nothing. If a bed partner hears you gasp or stop breathing, see a doctor whatever your watch says.

Symptoms, not scores, decide when to see a doctor

Symptoms set the pace, not tracker scores. Sleepiness at the wheel and thoughts of self-harm need action straight away; possible sleep apnea, whether a tracker or a bed partner raised it, should reach a doctor within weeks rather than months; and poor sleep that has dragged on for months can wait for your next routine appointment. We grouped these tiers ourselves from NHS, NHLBI, NIMH and AASM advice.

• Now: if you feel sleepy, stay out of the driver’s seat; the NHS tells people not to drive when sleepy.¹³ If poor sleep, or worry about it, ever comes with thoughts of suicide or self-harm, treat those thoughts as the urgent problem rather than the sleep. In the US, the National Institute of Mental Health points to the 988 Suicide & Crisis Lifeline (call or text 988) and says to call 911 in life-threatening situations.¹⁴ In the UK, the NHS says to call 999 or go to A&E if a life is at risk or you cannot keep yourself safe, and to use NHS 111 online or call 111 for urgent mental health help.¹⁵ Elsewhere, call your local emergency number.
• Soon: if a tracker raises an apnea alert, see a doctor within days or weeks; the AASM advisory says anyone who thinks they may have sleep apnea should talk to a health professional.⁸ The NHS advises booking a GP visit if your breathing halts and restarts in your sleep, you make gasping, snorting or choking sounds, or you always feel very tired during the day, ideally bringing whoever has noticed the symptoms. If you drive in the UK, it adds that once sleep apnea with excessive sleepiness is confirmed, you must not drive until symptoms such as feeling very tired are under control, and you may need to tell the Driver and Vehicle Licensing Agency.¹⁶
• At your next appointment: mention trouble sleeping that has lasted for months, or that better habits have not fixed, as the NHS suggests.¹³ A diary helps here: the NHLBI says a record of how long and how well you sleep, and how sleepy you feel by day, can help a provider make a diagnosis.¹⁷

The bottom line

A sleep tracker sketches your weeks well and judges single nights badly. Lean on its weekly patterns and time in bed, hold its stage minutes and bad-night totals loosely, and never take a quiet apnea feature as a clean bill of health. When gasping in your sleep or months of poor nights show up, let those symptoms, not a score, send you to a doctor.