Sleep restriction therapy at home: a two-week worked example

What is sleep restriction therapy?

Sleep restriction therapy is a CBT-I (cognitive behavioural therapy for insomnia) technique that shrinks your time in bed to match how long you sleep, then lengthens it step by step.

The American Academy of Sleep Medicine (AASM) suggests it as a single treatment for chronic insomnia. The 2021 guideline rates that as a conditional recommendation. Full multicomponent CBT-I gets the only strong one. The same guideline suggests against sleep hygiene tips alone.

People with insomnia often spend 8 or 9 hours in bed to get 6 hours of sleep. The extra time in bed is spent awake. Cutting it raises the pressure to sleep, so you fall asleep faster and wake less.

A 2021 meta-analysis pooled 8 randomised trials. It found large gains in insomnia severity (g = -0.93) and sleep efficiency (g = 0.91) against control groups. Six of the 8 trials had a high risk of bias. A 2025 meta-analysis of 9 trials and 1,239 people agreed on the direction. It found the benefit for severity lasted at follow-up but the benefit for time to fall asleep did not.

How do you do sleep restriction therapy? A two-week worked example

To do sleep restriction therapy, you log two weeks of sleep, set a window equal to your average sleep, fix your wake time, and adjust the window weekly by sleep efficiency.

The numbers below come from an invented diary, not a real person. The rules are Stanford’s starting window (average sleep, never under 5.5 hours) and the adjustment steps used in the 2025 Stockholm trial arm: 15 minutes, based on seven nights.

Step 1. Baseline. Over 14 nights, the sample sleeper is in bed from 10:30 pm to 7:00 am.

• Time in bed: 8 h 30 min = 510 minutes.
• Average sleep: 6 h = 360 minutes.
• Sleep efficiency = 360 ÷ 510 = 70.6%.
• Awake in bed: 510 − 360 = 150 minutes a night.

Step 2. Set the window. Starting time in bed equals average sleep: 360 minutes, or 6 h. This is above the 5.5-hour floor. Rise time is fixed at 6:30 am, so bedtime is 12:30 am.

Step 3. Week 1 check. The diary shows these averages.

• Time in bed: 360 minutes. Sleep: 330 minutes.
• Efficiency = 330 ÷ 360 = 91.7%.
• At 90% or more, add 15 minutes. New window: 375 minutes (6 h 15 min), bedtime 12:15 am.

Step 4. Week 2 check.

• Time in bed: 375 minutes. Sleep: 335 minutes.
• Efficiency = 335 ÷ 375 = 89.3%.
• Between 85% and 89%, keep the window. It stays at 6 h 15 min.

What if week 1 had gone worse? Say sleep averaged 300 minutes in a 360-minute window. Efficiency = 300 ÷ 360 = 83.3%. Below 85%, the trial rule cuts 15 minutes, giving a 345-minute window (5 h 45 min).

Wake time stays fixed all the way through. The 2025 trial told people to avoid naps, or to keep a nap under 15 minutes at the same time each day before 3 pm.

What happens in the first week of sleep restriction?

In the first week of sleep restriction therapy, expect shorter total sleep and more daytime sleepiness before things improve.

A lab study of 16 patients (mean age 47.1) measured sleep with polysomnography. Total sleep time fell by 91 minutes on night 1, 78 on night 8 and 69 on night 22, compared with baseline. Daytime sleepiness and reaction-time lapses increased during the acute phase. Self-reported insomnia severity and diary-reported sleep outcomes improved by the end of treatment. The study was small and had no control group.

An Oxford commentary on the 2025 trial says restriction has not been shown to raise serious adverse events. It also says the treatment is associated with a temporary worsening of symptoms and functioning, and that fear of sleep loss is a main reason people quit.

Stanford adds that nighttime wakefulness often drops markedly after about a week.

Which thresholds do the guides use?

Published guides use different efficiency cut-offs, and none of the four below match exactly. The step-size column shows the other split: the original method resets a cut to your average sleep, while the others trim by minutes.

Source	Add time in bed	Keep	Cut time in bed	Step size
Spielman’s original method, as summarised in the 2025 standard-definitions paper	About 90%	85% to 90%	Below 85%	Increase: 15 min. Decrease: reset to mean sleep over the previous 5 days, no sooner than 10 days after starting or changing the schedule
2025 Stockholm trial, restriction arm	90% or more	85% to 89%	Below 85%	15 min
Stanford Health Care	85% or more	80% to under 85%	Below 80%	Increase: 15 to 30 min; reduction size not specified
Sleep Foundation (2026)	Above 90% and still needing more sleep	85% to 90%	Below 85% and not sleepy	15 min

The 2025 standard-definitions paper recommends 85% or more as the cut-off to expand. A scoping review of 52 studies found 9 different ways of calculating the starting window alone. The starting floor also varies: Spielman used 4.5 hours, the University of Washington deck says never below 5, and Stanford says 5.5.

Pick one specified protocol and follow it with your clinician or programme. This worked example combines Stanford’s minimum window with the Stockholm trial’s adjustment rules; the trial did not test that exact combination.

Sleep compression vs sleep restriction: which is gentler?

Sleep compression cuts time in bed gradually instead of in one step, and it is gentler but slower, according to the only head-to-head trial.

The 2025 Stockholm trial randomised 234 adults (mean age 44.3, 73.4% women) to 10 weeks of therapist-guided online treatment. Both groups improved insomnia severity. Compression failed non-inferiority at the 1.6-point margin on the Insomnia Severity Index, improved significantly less (p = .006) and improved more slowly. It had fewer side effects and better adherence. The authors say clinicians should prioritise restriction, and treat compression as a valid alternative when restriction cannot be used.

Here is the same sample sleeper under the trial’s compression rule, which cuts the window by 20% of average time awake in bed.

	Restriction	Compression
Start window	360 min (6 h)	510 − 30 = 480 min (8 h)
Why	Average sleep	20% of 150 min awake = 30 min
After week 1 (sleep 330 vs 350 min)	91.7% efficient, window rises to 375 min	72.9% efficient, cut 20% of 130 min awake = 26 min, window 454 min
After week 2 (sleep 335 vs 355 min)	89.3% efficient, window stays 375 min	78.2% efficient, cut 20% of 99 min awake = 20 min, window 434 min

The compression figures are our arithmetic from the trial’s published rule, with invented diary values. In the trial neither arm had a minimum window.

Compression comes from Kenneth Lichstein’s work in the 1980s. The Oxford commentary says it was meant for people who sleep reasonably well but in fragments, and that the few published studies were in older adults.

Is sleep restriction safe after 60?

No trial tests the full restriction protocol against compression in people over 60, so the evidence for this age group is thin.

A small trial in 39 adults aged 55 and over compared restriction plus sleep hygiene with hygiene alone. It found few differences between groups. The authors suggested the participants had mild symptoms and that hygiene helped by itself. That is one small study from 2000.

The 2025 standard-definitions paper says sleep compression has been found to work in late-life insomnia, and lists chronic pain, chronic disease and frailty as reasons to choose it over restriction. It gives no age-specific restriction rules.

If you are over 60, take sleep medicines or have falls, ask a clinician which version suits you. Sleepiness in the first weeks is the specific worry.

Who should not do sleep restriction therapy?

People with bipolar disorder, an uncontrolled seizure disorder, untreated sleep apnea or a safety-critical job should not start sleep restriction without a clinician.

Two clinical sources list the exclusions.

• A 2024 University of Washington training deck, citing Edinger et al. 2021, says to modify restriction for poorly controlled seizure disorder, for bipolar disorder with past mania or hypomania triggered by sleep loss, for a recent illness, accident or surgery, for excessive daytime sleepiness from an underlying sleep disorder, and for unacceptable occupational risk from sleepiness.
• The Sleep Foundation (2026) lists safety-sensitive jobs such as transport, construction and healthcare, major illness or recent surgery, untreated sleep apnea, uncontrolled seizures and untreated bipolar disorder.

The same deck lists an Epworth Sleepiness Scale score above 11 and active heavy drinking among situations where CBT-I needs care. The 2025 standard-definitions paper says people should be warned about drowsiness and safety risks. Skip driving and machinery when drowsy.

Snoring, gasping or witnessed pauses in breathing need a sleep apnea assessment first; see sleep apnea in women over 50. Alcohol worsens both problems (sleep apnea and alcohol). Our guides on perimenopause insomnia and waking at 3am cover other causes to rule out.

When should you see a clinician?

See a clinician if daytime sleepiness makes driving unsafe or your mood changes sharply (unusual energy, racing thoughts or low mood). A clinician can also offer full CBT-I, the AASM’s strongest recommendation, if you have not improved after 10 weeks. For a guided app route, see the CBT-I Coach app.