Sleep efficiency is the percentage of a defined time-in-bed or recording interval that is estimated or scored as sleep. The basic formula is:
Sleep efficiency (%) = total sleep time ÷ relevant interval × 100
The word relevant matters. A sleep diary, clinical actigraph, laboratory sleep study, and consumer tracker may estimate total sleep time differently. They may also use different start and end points for the denominator. A percentage without its method and interval is incomplete.
Sleep efficiency describes a ratio. It does not by itself show whether someone slept long enough, slept well, or has insomnia, obstructive sleep apnea, or another disorder.
A simple sleep-efficiency calculation
Suppose a person is in bed from 11:00 p.m. to 7:00 a.m. The time-in-bed interval is 8 hours, or 480 minutes. They estimate that they slept for 6 hours 40 minutes, or 400 minutes.
400 ÷ 480 × 100 = 83.3%
Under that definition, the sleep efficiency is about 83%.
Now suppose the person spent the first 20 minutes reading before trying to sleep and the last 10 minutes awake after deciding not to try again. One proposed insomnia-focused denominator, called the duration of the sleep episode (DSE), excludes those 30 minutes. The denominator becomes 450 minutes:
400 ÷ 450 × 100 = 88.9%
The same night now produces about 89%. Neither calculation should be interpreted without its definition. The first describes sleep as a share of literal time in bed. The second describes sleep as a share of the period during which the person was trying to sleep.
Researchers David Reed and William Sacco proposed DSE because literal time in bed can include reading, texting, talking, or resting before and after any attempt to sleep. Their formula adds sleep onset latency, total sleep time, wake after sleep onset, and time spent trying to sleep after the final awakening. It is a proposed way to make insomnia research and sleep diaries more consistent, not a universally superior denominator for every device or clinical test 1.
The denominator changes the question
| Denominator | Typical start and end | What it includes | Main limitation |
|---|---|---|---|
| Literal time in bed | Getting into bed to getting out of bed | Sleep, wakefulness, and any nonsleep activity in bed | Reading or resting in bed can lower the percentage even when the person was not trying to sleep |
| Duration of the sleep episode (DSE) | First attempt to sleep to the final point at which sleep is no longer attempted | Sleep onset latency, total sleep time, wake after sleep onset, and any final attempt to return to sleep | Requires the person to distinguish being in bed from trying to sleep |
| Polysomnography recording interval | Commonly lights out to lights on | Every scored sleep and wake epoch in the defined laboratory interval | The interval is set by the study protocol and may not match the person's usual night |
| Device-defined sleep period | Set by user input, event markers, or an algorithm | Whatever the device and software count inside that period | The exact start, end, and sleep rules may be proprietary or change with software |
Ask what interval a report uses before comparing percentages. Even research papers have not always defined sleep efficiency consistently 1.
Sleep efficiency is not sleep duration or sleep quality
Several measurements can describe one night, but they are not interchangeable.
| Measurement | What it means | What it cannot establish alone |
|---|---|---|
| Total sleep time (TST) | The total minutes estimated or scored as sleep | Whether that amount met the person's sleep need |
| Sleep onset latency (SOL) | Time from trying to sleep or lights out to first sleep onset | How much wakefulness occurred later |
| Wake after sleep onset (WASO) | Total time awake after first sleep onset within the chosen interval | Why the person awakened |
| Sleep efficiency | TST divided by the chosen interval | Sleep duration, sleep quality, or a diagnosis |
| Sleep quality | The person's overall experience of sleep, including satisfaction and restoration | A standardized physiological measurement |
A person can fall asleep quickly but spend a long time awake later, or take a long time to fall asleep and then sleep continuously. Both patterns can lower the same ratio for different reasons. See the focused guides to sleep onset latency and wake after sleep onset for those calculations.
Sleep quality is broader and partly subjective. Two nights with the same sleep efficiency can feel very different because of total sleep time, pain, nightmares, breathing events, medication effects, caregiving, or the timing of sleep.
Is 85% or 90% a good sleep efficiency?
There is no universal sleep-efficiency percentage that defines healthy sleep for every person and every method. A threshold used in a study, clinic, or cognitive behavioral therapy for insomnia (CBT-I) protocol belongs to that specific purpose. It is not a stand-alone diagnosis or a self-treatment rule.
Consider two examples:
- A person who sleeps 4 hours 45 minutes during a 5-hour opportunity has 95% sleep efficiency. The ratio is high, but the short opportunity may still leave them sleepy and impaired.
- A person who sleeps 6 hours 30 minutes during an 8-hour opportunity has about 81% sleep efficiency. One such night could reflect illness, pain, caregiving, an unfamiliar laboratory, environmental disruption, or ordinary night-to-night variation.
A higher ratio is not automatically healthier, and a lower ratio does not reveal its cause. Deliberately shortening time in bed can raise the percentage without adding sleep.
Insomnia is assessed from persistent difficulty sleeping despite adequate opportunity and circumstances, together with distress or daytime effects. One ratio does not capture those criteria 2. Obstructive sleep apnea requires a comprehensive sleep evaluation and appropriate objective testing, not a sleep-efficiency score 3.
How different methods produce the percentage
Sleep diary
A diary asks the person to estimate when they tried to sleep, how long sleep onset took, time awake during the night, final awakening, rise time, and total sleep time. The Consensus Sleep Diary was developed with expert review and patient focus groups to standardize these entries 4.
A diary records the person's sleep experience in their normal setting. It cannot directly detect brain-wave sleep, and brief awakenings may be forgotten. Its sleep efficiency also depends on whether the denominator runs from bed entry to bed exit or from the first to the final sleep attempt.
Clinical actigraphy
Clinical actigraphy uses a wrist-worn movement sensor and software to estimate sleep and wake over multiple days and nights. A diary, event button, or clinician-defined interval often helps set the sleep period. Quiet wakefulness can look like sleep because movement, not brain activity, is the main signal.
The American Academy of Sleep Medicine conditionally recommends actigraphy when an objective estimate would help evaluate adult insomnia, circadian disorders, or suspected insufficient sleep. Its guideline describes recordings from at least 72 hours through 14 consecutive days, depending on the clinical question. Objective monitoring is not required for the routine diagnosis of insomnia 5.
Polysomnography
Polysomnography (PSG) uses brain waves, eye movements, muscle activity, breathing, oxygen, heart rhythm, and other signals. Trained scoring identifies sleep and wake in short epochs. Sleep efficiency is commonly total scored sleep divided by the lights-out to lights-on recording interval 6.
PSG provides physiological detail that a diary or movement sensor cannot. It usually represents one or a few monitored nights, however, and the laboratory interval or setting may differ from usual sleep. A low PSG percentage does not by itself diagnose insomnia, and a high one does not rule out breathing, movement, or other abnormalities.
Consumer trackers
Watches, rings, phone apps, and bedside devices may combine movement, pulse signals, temperature, sound, or other inputs. Their algorithms infer sleep and may create product-specific sleep periods and scores. The label “sleep efficiency” does not guarantee that two products used the same denominator or sleep definition.
The AASM notes that sleep terms, sensors, populations, algorithms, and validation can differ among products. Software updates can also change an output after a validation study was completed 7. Consumer sleep technology can support a conversation with a clinician, but it should not replace appropriate evaluation or validated diagnostic testing 8.
Why a diary and device may disagree
Disagreement is expected because the methods observe different things.
A person may remember being awake and still for 40 minutes while an actigraph classifies most of that interval as sleep. A PSG can detect brain-wave sleep and wake, but it measures a monitored night under a protocol. A consumer tracker may use a different start time, end time, or software definition from either one.
This is sometimes called a subjective-objective discrepancy. It does not prove that the person's experience is false or that the device is broken. The useful questions are:
- Did both methods use the same start and end points?
- Was quiet wakefulness likely to be scored as sleep?
- Was one result from a typical night and the other from an unusual night?
- Did the device, application, or algorithm version change?
- Is the purpose to understand the person's experience, estimate a multi-night schedule, or test for a physiological disorder?
Compare trends within the same method before comparing one device's percentage with another method's value.
A practical multi-night record
One night is often too little context. If tracking is useful and does not make sleep more stressful, record one or two representative weeks. That is a practical window, not a diagnostic minimum. Clinical actigraphy commonly uses recordings within the 72-hour to 14-day range set out in the AASM guideline 5.
Complete the record in the morning rather than watching the clock through the night. For each night, note:
- when you got into bed
- when you first tried to sleep
- estimated sleep onset latency
- total estimated wake time after sleep onset
- final awakening and when you stopped trying to sleep
- when you got out of bed
- total sleep time
- naps, shift changes, travel, illness, pain, caregiving, alcohol, caffeine, and relevant medicines
- daytime sleepiness, fatigue, concentration, mood, and any safety concern
Use the same denominator throughout the record and label it. If using a device, note its model and software version. The Consensus Sleep Diary offers a standardized structure and keeps final awakening separate from getting out of bed 4.
Review the pattern in this order:
- Sleep opportunity: Was enough time available for sleep, or was the interval shortened by work, caregiving, travel, or choice?
- Total sleep time: Did a high percentage still contain little sleep?
- Source of wakefulness: Did the ratio fall mostly because of sleep onset latency, wake after sleep onset, or time after the final awakening?
- Context and symptoms: Were pain, nocturia, hot flashes, breathing symptoms, leg discomfort, nightmares, a schedule mismatch, substances, medicines, or the environment involved?
- Daytime effect: Was there sleepiness, fatigue, impaired concentration, mood change, or unsafe driving or work?
The goal is to identify a repeatable pattern and its likely cause, not to earn a nightly score.
Improve the cause, not the ratio
A sleep-efficiency percentage does not tell you which intervention fits. Noise calls for a different response from chronic insomnia, pain, circadian mismatch, medication effects, or obstructive sleep apnea.
Seek a clinical assessment when difficulty falling or staying asleep persists despite adequate opportunity and affects daytime life. Chronic insomnia is best addressed with CBT-I, a structured treatment that combines behavioral and cognitive methods. Current guidelines recommend multicomponent CBT-I rather than relying on sleep-hygiene advice alone 29.
A clinician may use sleep-efficiency thresholds within a CBT-I protocol to help adjust a sleep schedule. Do not copy those thresholds into a self-directed time-in-bed restriction plan. Early time-in-bed restriction can increase daytime sleepiness and concentration difficulty. The AASM guideline identifies excessive daytime sleepiness, safety-critical work such as driving or operating heavy machinery, risk of mania or hypomania, and poorly controlled seizure disorders as situations requiring particular caution and clinician monitoring 9.
Pregnancy also calls for an individualized plan rather than a generic sleep-efficiency rule. A randomized trial supports tailored CBT-I for prenatal insomnia, but the intervention was delivered as a treatment program to screened participants, not as unsupervised number-chasing 10.
Get the competing problem evaluated before restricting time in bed if there is loud snoring, witnessed breathing pauses, gasping, marked sleepiness, unusual nighttime behavior, a major mood change, or another symptom suggesting that insomnia is not the only issue. Do not drive or perform hazardous work when too sleepy to stay alert.
The bottom line
Sleep efficiency is total sleep time divided by a clearly defined time-in-bed, sleep-attempt, recording, or device interval. The denominator and sleep-detection method can materially change the percentage.
Use the ratio to describe a pattern, not to grade a person or diagnose a disorder. Look first at sleep opportunity, total sleep time, daytime function, accompanying symptoms, and the measurement method. When persistent difficulty despite adequate opportunity points to chronic insomnia, seek CBT-I rather than shortening time in bed to force the percentage upward.
