Stimulus control and sleep restriction

1. Introduction

Version followed: Stimulus Control Therapy (SCT) as set out by Bootzin & Perlis (2011, six instructions, building on Bootzin 1972), and Sleep Restriction Therapy (SRT) as set out by Spielman, Yang & Glovinsky (2011, updating Spielman, Saskin & Thorpy 1987). The two are combined into a sleep-scheduling package as in standard CBT-I (Muench, Vargas, Posner & Perlis 2022). A brief 4-session nurse-delivered SRT format (HABIT trial, Kyle et al. 2023) is given as an alternative. Where these sources differ, the difference is named in the text.

These are two brief behavioural treatments for chronic insomnia. They can be used on their own, which the AASM 2021 guideline conditionally recommends for each, or as the behavioural core of CBT-I [source: Edinger et al. 2021 AASM guideline; Div 12]. Neither is a full psychotherapy. Each is a short set of instructions that the clinician explains, tailors, monitors with a sleep diary and helps the patient stick to.

How SCT says change happens. Bootzin treated falling asleep as a learned (operant) behaviour. Cues that reliably come before sleep become signals for sleep. In insomnia those cues are weak, or the bed and bedroom have become signals for waking, worry and frustration, partly through Pavlovian conditioning. SCT restricts what happens in bed and how long the person lies there awake. This rebuilds the bed as a cue for falling asleep quickly and weakens it as a cue for arousal. A fixed rising time steadies the sleep–wake rhythm [source: Bootzin 1972; Bootzin & Perlis 2011].
How SRT says change happens. In Spielman's 3P model, the factors that keep insomnia going include spending too long in bed, worry before sleep and preoccupation with daytime effects. SRT cuts time in bed to roughly the time actually slept. The resulting mild sleep loss builds homeostatic sleep pressure, which shortens sleep latency, reduces night waking and lowers hyperarousal. Fixed bed and rise times strengthen circadian timing. Time in bed is then extended step by step as sleep consolidates. Patients also experience SRT as "active", which tends to reduce nightly worry about sleep [source: Spielman, Yang & Glovinsky 2011; Spielman et al. 1987 abstract].

2. Framework

Stage 0: Assessment and screening (before any schedule change). [source: Spielman et al. 2011; Bootzin & Perlis 2011; Kyle et al. 2023; Leggett slides]

Goals: confirm chronic insomnia and collect 1–2 weeks of baseline sleep diary. From the diary, estimate average total sleep time (TST), time in bed (TIB), sleep efficiency (SE = TST ÷ TIB × 100), workday rise time and the part of the night where sleep is most consolidated.
Screen for: other sleep disorders (sleep apnoea, restless legs, narcolepsy, parasomnias); circadian rhythm problems and shift work; epilepsy; bipolar disorder; safety-critical jobs and driving; fall risk and mobility; pregnancy; dementia or cognitive limits that would stop the person following instructions; current suicidal ideation, using standard risk assessment. These are the exclusions and cautions named in the HABIT trial and in Spielman et al. (2011).
Indication for SRT in Spielman's chapter: SE below 85% (below 80% in older adults). Also consider people with an early weekday rise and long lie-ins on free days, whose weekly SE can look normal.
Move on when the diary is representative and screening has found no contraindication, or has found one that is managed.

Stage 1: Rationale and prescription (session 1).

Teach the conditioning rationale and the sleep-drive rationale, using the patient's own diary.
Set the sleep window (bedtime and fixed rise time), agree the stimulus control rules, and plan for daytime sleepiness, driving and barriers.
Give diaries and an SE calculation grid.
Session length: about 30 minutes in HABIT. Standard CBT-I sessions run 15–90 minutes [source: HABIT protocol v6.0; Muench et al. 2022].

Stage 2: Titration (weekly, usually sessions 2 to 4–8). [source: Spielman et al. 2011; Muench et al. 2022]

Each week, review the diary, compute mean SE for the week and adjust TIB by the rules in §4.
Troubleshoot adherence and reinforce stimulus control.
Spielman's original 1987 protocol adjusted on 5-day averages. The 2011 authors endorse 7-day averages instead.
Expect a rough first 1–2 weeks: more getting in and out of bed and more daytime sleepiness. Improvement then follows over weeks [source: Bootzin & Perlis 2011; Wikipedia summary of side-effect data].

Stage 3: Endpoint and maintenance. [source: Spielman et al. 2011]

Stop extending TIB when the patient has consolidated sleep and good daytime functioning, with no significant deficits. Spielman calls this the point of maximum net benefit.
Do not chase maximum SE. The highest SE usually comes with the shortest TIB, which leaves patients sleepy and less likely to adhere.
The patient then keeps the final schedule alone. Bootzin advised keeping the SC rules even after success [source: Bootzin 1972].

Typical dose and format.

Original SRT: 8 weekly sessions [source: Spielman et al. 1987].
Standard CBT-I: 6–8 sessions, weekly or fortnightly, individual or group [source: Muench et al. 2022].
Brief SRT in primary care: 4 weekly sessions totalling about 65 minutes (two face-to-face, two by phone), delivered by nurses after half a day of training using a manual, booklet, diaries and SE grids [source: HABIT protocol; Kyle et al. 2023].
Digital delivery exists, but no digital protocol was reviewed here [uncertain].
Most patients have not recovered their full desired sleep time by session 8. Titration can continue on their own afterwards [source: Muench et al. 2022].
A standalone session count for SCT was not found in the primary sources [uncertain].

3. Guiding principles

[source: Bootzin 1972; Bootzin & Perlis 2011; Spielman et al. 2011; Muench et al. 2022; Kaplan & Harvey 2013; [synthesis] where marked]

Data before prescription. Every schedule decision comes from the diary, never from an impression of "a bad week". Practically: no diary, no titration change.
Sleep opportunity matches sleep ability. Time in bed should approximate what the person can actually sleep. Practically: shrink the window first, then earn time back.
The rise time is the anchor. Wake time is fixed 7 days a week, with less than 1 hour of variation between workdays and days off. Practically: adjust bedtime, not rise time, unless the night's pattern says otherwise.
Sleepiness, not the clock, opens the bedroom door. Bootzin kept the bedtime instruction vague on purpose, to avoid clock-watching. Practically: within the window, bed only when sleepy. This instruction is often dropped when SRT sets the bedtime.
The bed is for sleep (and sex) only. Practically: all wakeful activity, including worry, moves to another room.
Limit wakefulness in bed, not effort to sleep. Practically: leave bed when sleep is not coming, judged by feel rather than by staring at the clock.
Explain every rule. Bootzin & Perlis stress going through each rule with its reason rather than handing out a list. Practically: link each rule to the patient's own diary and habits.
Expect worse before better, and say so in advance. Practically: name the likely first-fortnight sleepiness and in-and-out-of-bed nights at session 1.
Safety outranks dose. Never restrict below 5 hours of TIB. Screen for conditions worsened by sleep loss. Give driving and machinery advice. Practically: modify or drop SRT, or use sleep compression, where risk is high.
Tailor intensity to tolerance. Use negotiated windows, +30-minute buffers, gradual compression and permitted naps for older or fragile patients. Practically: a schedule the patient will keep beats a stricter one they abandon.
The goal is daytime function, not a number. Practically: end titration at maximum net benefit, not maximum SE.
Monitor mood where sleep loss is risky. Practically: in bipolar disorder, regularise the schedule first, use a higher floor, rate mood each session and pause techniques if hypomania emerges [source: Kaplan & Harvey 2013].
Collaborative stance. [synthesis] Treat the patient as a co-investigator testing the rules for 2–3 weeks rather than obeying orders. This supports adherence to a demanding treatment, which HABIT authors call "a challenging treatment" [source: Kyle et al. 2023].

4. Intervention inventory

The items run in delivery order. A–C cover assessment and shared rationale, D–J sleep restriction and its variants, K–Q stimulus control and its variants, and R–T apply across both. All example wording is my own illustrative phrasing and is labelled "(illustrative)". Handouts and scales mentioned (Consensus Sleep Diary, Epworth Sleepiness Scale, Insomnia Severity Index, Penn and VA handouts) are named here but not reproduced.

A. Sleep diary monitoring and sleep-efficiency calculation

[source: Spielman et al. 2011; Muench et al. 2022; HABIT protocol; Penn CBT-I clinical tools page]

Purpose / when: The data engine for every decision. Start 1–2 weeks before treatment and continue daily throughout. Use a standard form such as the Consensus Sleep Diary, or a graphic diary, plus an SE grid.
Steps:
1. Each morning the patient records, as estimates (not from the clock): bedtime, lights-out, time to fall asleep, number and length of awakenings, final wake time, rise time, naps, and sleep quality.
2. The clinician or patient computes, for each night: TIB = rise time − bedtime. TST = TIB − sleep latency − wake after sleep onset − early-morning awake time in bed. SE = TST ÷ TIB × 100.
3. Average over the week, or over the 5-day window of the original protocol, to get mean TST and mean SE.
4. Teach the patient to use the grid themselves. HABIT gives an SE calculation grid for each week.
Example language (illustrative): "Fill this in within an hour of getting up, using rough guesses. Don't check the clock during the night. A ballpark figure is all we need."
Watch for: Diaries filled in retrospectively before the session. Clock-checking at night to get accurate entries, which feeds arousal. Missing nap or dozing entries. Large weekday/weekend differences, which may hide a circadian problem. Diaries completed most days mean it is working well enough. Gaps mean revisit the purpose and simplify the form [synthesis].

B. Rationale for sleep restriction (3P model and sleep drive)

[source: Spielman et al. 2011]

Purpose / when: Session 1, before the window is set. It builds the case for accepting short-term sleep loss.
Steps:
1. Draw three bars over time: predisposing traits, a precipitating stressor, and perpetuating habits. Show that the insomnia continued after the trigger faded because of the habits, especially spending longer in bed to catch up.
2. Explain sleep pressure building with time awake, and show that long TIB spreads a limited amount of sleep thinly across the night. Use the patient's own diary SE.
3. Explain that a fixed schedule trains the body clock.
4. Present SRT as temporary. The window grows as sleep becomes solid.
Example language (illustrative): "You're in bed about eight and a half hours and sleeping about six. We'll shrink the window to about what you sleep, so that sleep gets deeper and more continuous. Then we earn the time back week by week."
Watch for: The patient hearing "less sleep" as "more harm". Address this with the expected time course (B → D) and the safety plan (I). Rationale understood: the patient can restate why lying in bed longer backfires.

C. Rationale for stimulus control (conditioning)

[source: Bootzin & Perlis 2011; Bootzin 1972]

Purpose / when: Session 1. It makes the rules make sense rather than feel arbitrary.
Steps:
1. List with the patient everything they do in bed or the bedroom besides sleep: phone, TV, reading, eating, working, planning, worrying, lying awake.
2. Use the learning analogy Bootzin & Perlis describe: ask how well they study in bed compared with at a kitchen table, and why. Then generalise: places become cues for what happens there.
3. If the patient feels sleepy in the evening but "switches on" on entering the bedroom, emphasise the Pavlovian version: the bed has become a trigger for arousal.
4. Present the goal: the bed becomes a strong cue for falling asleep fast, and a weak cue for anything else.
Example language (illustrative): "Your bed has had a lot of practice being the place where you lie awake and think. We're going to retrain it so that it means one thing: sleep."
Watch for: Patients who say they "relax" with screens in bed. Explore whether it actually helps them fall asleep (it usually postpones sleep) rather than banning it outright [synthesis]. Rationale understood: the patient links their own habits to their difficulty falling asleep.

D. Setting the initial sleep window (SRT prescription)

[source: Spielman et al. 2011; Muench et al. 2022; Kyle et al. 2023]

Purpose / when: Session 1, once the diary is representative and screening is complete.
Steps:
1. Set TIB equal to mean diary TST. The floor is 5 hours in Spielman et al. 2011. The original 1987 paper is often cited as using 4.5 hours, but this is [uncertain]. HABIT's floor is not stated in the papers [uncertain].
2. Set the rise time. Make it no later than the usual workday rise time, and one the patient can keep 7 days a week. Set it earlier if the best sleep falls in the first two-thirds of the night.
3. Set the bedtime. Count back from the rise time by the TIB. If the main problem is falling asleep or poor early-night sleep, keep the rise time and push the bedtime later. If sleep is scattered with a long latency, night waking and early waking, push the bedtime later and pull the rise time earlier, so the window sits on the most consolidated stretch of sleep.
4. Write it down as a prescription: "Not in bed before X, out of bed by Y, every day."
5. Agree what the patient will do in the newly freed evening hours (see K and Q).
Worked example (from Spielman et al. 2011, paraphrased): The patient averages 5 h 45 min of sleep, rises at 6:30, and the last two hours of the night are light and broken. The prescribed window is 11:00 pm to 4:45 am.
Example language (illustrative): "Your diary says you average about six hours. So for this week your bed is open from 1 am to 7 am, and only then. The alarm goes off at 7 every day, weekends included."
Watch for: Very low reported TST, around 3 hours, as in paradoxical insomnia. The 5-hour floor means you either hold at 5 hours or do not use SRT; consider J. A window the patient flatly refuses: negotiate (see H). Working well: SE climbs toward 85–90% within 1–3 weeks, and sleep latency and night waking shrink [synthesis on timing].

E. Weekly titration of the sleep window

[source: Spielman et al. 2011; Muench et al. 2022]

Purpose / when: Every follow-up session, to lengthen TIB as sleep consolidates.
Steps (standard adult rules, Spielman 2011 and Muench 2022):
1. Compute mean SE for the past week. Use 7 days; the original protocol used 5.
2. SE ≥ 90%: add 15 minutes to TIB. Spielman allows 15 or 30 minutes; Muench et al. use 15.
3. SE 85% to <90%: no change.
4. SE < 85%: remove 15 minutes (or 30) from TIB, never going below the floor.
5. Older adults: move every threshold down 5 points. Increase at ≥ 85%, hold at 80% to <85%, decrease below 80%.
6. Decide where the extra time goes, usually an earlier bedtime. Put it at the start of the night if the remaining problem is early-morning sleep.
7. Write down the new window and repeat next week.
Note on versions: Other guides use different cut-offs, such as 80/85/90% combinations and 15–30 minute steps. Trials vary widely, and only 7% of 88 trials reported all their SRT parameters [source: Kyle et al. 2015 abstract]. Choose one rule set and use it consistently.
Example language (illustrative): "Your efficiency averaged 91% this week. That earns 15 minutes. Do you want it at the start of the night or the end?"
Watch for: Early sleepiness that tempts you to extend before SE has earned it. Hold the rule unless safety requires otherwise. SE stuck below 85% despite good adherence: reassess for another sleep disorder (see §5). Working well: TST stable or rising while TIB grows, and daytime functioning improving.

F. Deciding the endpoint (maximum net benefit)

[source: Spielman et al. 2011]

Purpose / when: When SE hovers near the threshold after increases, or daytime functioning is good.
Steps:
1. Weigh the costs (time in bed, vulnerability to fragmented sleep) against the benefits (sleep quality, daytime functioning).
2. Stop extending when further time in bed adds little sleep or brings back broken sleep, and the patient functions well by day.
3. Hand over the final schedule for the patient to keep.
Worked example (Spielman et al. 2011, paraphrased): A 74-year-old reached a 7-hour window of 11 pm to 6 am. Her SE then settled between 80% and 85%, she functioned well by day, and treatment ended with advice to keep that schedule.
Example language (illustrative): "Adding more time in bed now mostly adds more time awake. You're functioning well, so this looks like your sweet spot."
Watch for: Patients holding out for 8 hours on principle. Revisit how much sleep they actually need, judged by daytime function [synthesis].

G. Sleep compression (gradual variant)

[source: Spielman et al. 2011; Muench et al. 2022; Penn session slides 2014]

Purpose / when: For patients who will not accept or cannot tolerate an abrupt cut: older adults, people at higher risk from sleep loss, or short sleepers with little daytime impairment.
Steps (Spielman 2011 description):
1. Compute DIFF = baseline TIB − baseline TST.
2. Session 1: cut TIB by half of DIFF.
3. Sessions 2 and 3: cut a further quarter of DIFF each time.
4. The original version stopped once target SE was reached. Newer versions then extend, as in E.
Alternative schedule (Penn slides, Perlis): cut by DIFF ÷ 5 each week, by delaying bedtime or advancing rise time, then apply the SRT titration rules. This differs from the Spielman description. Name the one you use.
Full protocol: Lichstein's sleep-compression chapter (chapter 5 of the 2011 primer) is not drawn on [uncertain].
Example language (illustrative): "Rather than a big cut at once, we'll trim your time in bed in steps over the next three weeks."
Watch for: Slower gains. In one older-adult trial reported by Spielman et al., the benefit on night waking lasted to one year only in the compression group.

H. Tolerance modifications of SRT

[source: Spielman et al. 2011; Muench et al. 2022]

Purpose / when: At session 1 or after early drop-out risk appears, to make SRT tolerable without abandoning it.
Options described by the authors:
- Set the starting TIB at mean TST plus 30 minutes.
- Negotiate the starting TIB with the patient.
- Use a gentler, gradual cut (G).
- Use lower SE thresholds, shorter review intervals or bigger upward steps (Muench).
- Allow or prescribe a short daytime nap early on, especially for older adults (see P).
Steps:
1. Offer the standard window first.
2. If the patient refuses or there is a safety concern, offer one modification at a time.
3. Record which one you used, so the titration stays consistent.
Example language (illustrative): "If six hours feels impossible, we could start at six and a half and tighten only if the diary says we need to. Which would you actually stick to?"
Watch for: Too many concessions, so that TIB stays well above TST and nothing changes. If SE has not improved after 2 weeks of a buffered window, tighten it [synthesis].

I. Daytime sleepiness and safety plan

[source: Kyle et al. 2023 (session 1 content); Spielman et al. 2011; Wikipedia summary of SRT side-effect studies; [synthesis] for specific steps]

Purpose / when: Session 1 of any SRT. It is required, because sleepiness, fatigue, irritability, headache and low motivation are common in the first 1–2 weeks.
Steps:
1. Name the likely side effects and their usual time course: worst early, then easing as sleep consolidates.
2. Driving and machinery. Agree that the patient will not drive or operate dangerous machinery when drowsy. Identify the higher-risk times, such as early morning after a cut or a mid-afternoon dip, and plan alternatives: lifts, public transport, swapping shifts [synthesis]. Safety-critical workers (long-haul drivers, air traffic control, heavy machinery) may need a modified approach or no SRT [source: Spielman et al. 2011].
3. Evening sleepiness before the window opens. Plan upright, mildly engaging activity in another room. Avoid dozing on the sofa, since unplanned sleep counts against the night [synthesis consistent with Bootzin's no-nap rule].
4. Track sleepiness weekly. An Epworth score is one option, used by Penn as a core questionnaire.
5. Agree in advance on a "stop rule": drowsy driving, a near-miss or falls mean contact the clinician and widen the window [synthesis; HABIT tracked near-miss driving incidents and falls as adverse events].
Example language (illustrative): "For the first couple of weeks you'll probably feel more tired, which is a sign the sleep drive is building. The non-negotiable part is safety: if you feel drowsy, you don't drive."
Watch for: Severe or worsening sleepiness after week 2, or any near-miss. Widen the window and reassess for sleep apnoea or narcolepsy.

J. Fixed-increment schedule for paradoxical insomnia

[source: Spielman et al. 2011]

Purpose / when: When the diary shows very low reported sleep (around 3 hours) that seems to underestimate actual sleep, and SE does not rise under restriction.
Steps:
1. Start at the 5-hour floor. Never go lower.
2. Add 15–30 minutes every week whatever the SE.
3. Combine with SCT and education about how imperfect sleep perception is.
Example language (illustrative): "Your diary may be underestimating how much you sleep, which is common. So rather than going by the numbers, we'll add a little time each week and watch how you feel during the day."
Watch for: Whether daytime function improves even if diary TST barely moves. Spielman notes that unrecognised light sleep late in the night can still make SRT worthwhile.

K. SCT rule 1: Go to bed only when sleepy

[source: Bootzin & Perlis 2011; Bootzin 1972; Muench et al. 2022]

Purpose / when: Mainly for difficulty falling asleep. Bootzin's original wording was "Lie down intending to go to sleep only when you are sleepy."
Steps:
1. Teach the difference between sleepiness (heavy eyelids, nodding off, yawning, losing the thread while reading) and fatigue or tiredness (low energy without the ability to sleep) [synthesis on the specific cues; the sources call it awareness of bodily sleepiness cues].
2. Present it as something to build over the first few weeks, not an instant requirement.
3. In combined treatment, the SRT window sets the earliest bedtime. Inside the window, the patient still waits for sleepiness. Muench et al. note this rule is often dropped when SRT is used.
Example language (illustrative): "The window opens at midnight, but that's the earliest you can go to bed, not an appointment. Go when your eyes are heavy."
Watch for: Patients who never feel sleepy and stay up very late. Check whether they are doing something arousing out of bed. In bipolar disorder, replace this rule with a regular bedtime (S).

L. SCT rule 2: Bed and bedroom for sleep and sex only

[source: Bootzin & Perlis 2011; Bootzin 1972; Div 12]

Purpose / when: Core element. It removes the cues for being awake from the bed.
Steps:
1. Using the list from C, move every non-sleep activity out of the bed, and ideally out of the bedroom: reading, TV, phone and screens, eating, working, planning and worrying.
2. Build a new pre-bed wind-down routine in another room. After sex, follow the rules and go to sleep.
3. Problem-solve logistics, such as a chair in another room, charging the phone outside the bedroom, or a bedsit with one room. For a single room, designate a chair facing away from the bed [synthesis].
Example language (illustrative): "From tonight, the bed is for sleep and sex. Your phone, your reading and your planning all happen in the living room."
Watch for: Hidden in-bed activities such as podcasts, scrolling and lying there planning. Ask about them specifically. Working well: shorter sleep latency, and the patient reports feeling drowsy rather than alert on lying down.

M. SCT rules 3–4: Get out of bed when not sleeping, and repeat as needed

[source: Bootzin & Perlis 2011; Bootzin 1972; Muench et al. 2022]

Purpose / when: Core element. Rule 3 is for difficulty falling asleep at bedtime. Rule 4 extends it to waking during the night.
Steps:
1. Time criterion. The standard instruction is to leave bed if sleep is not coming quickly. Lying awake more than about 10 minutes without getting up counts as not following the rule.
2. Don't clock-watch. Turn the clock away and leave at the first sign of frustration or wide-awakeness. The "quarter-hour" (about 15 minutes) criterion is an acceptable, workable alternative. Long stretches in bed (about an hour or more) are never allowed, even without frustration.
3. Go to another room. Do something quiet and relaxing in dim light, such as a reading lamp, or TV at a distance. Avoid computers, email and close-up screens, which are brighter and more arousing than people expect.
4. Return to bed only when sleepy, usually after at least 15 minutes up. Stay up as long as you wish.
5. Repeat as often as needed, at sleep onset or during the night.
6. Early-morning waking. Do not start the day at 4–5 am to avoid the bed. If 45 minutes or more remain before the set rise time, follow the rule and go back to bed. Keep the fixed rise time.
7. Prepare the out-of-bed space in advance: warm room, blanket, low lamp, and a pre-chosen boring activity. The VA manual has a "Things to do when awake" handout (listed on the Penn tools page) [uncertain on content].
Variant: Some authors advise avoiding sedentary activity while doing stimulus control, to prevent dozing on the sofa (Muench et al. 2022).
Example language (illustrative): "If you notice you're wide awake, frustrated or just not drifting off, don't wait it out. Get up, go to the lounge, put on the lamp, and come back when you're sleepy. Do it as many times as you need."
Watch for:
- Many trips in and out of bed in the first few days, possibly up to a couple of weeks. That is expected; forewarn the patient.
- Falls risk at night. Check lighting and walking aids, and consider counter-control (Q).
- Patients on hypnotics, where fall risk is raised [source: Leggett slides].
- Working well: fewer trips needed each week, and shorter time to fall asleep after returning.

N. SCT rule 5: Fixed rising time with an alarm

[source: Bootzin & Perlis 2011; Div 12]

Purpose / when: From session 1. It steadies the circadian rhythm and builds sleep pressure for the next night.
Steps:
1. Choose a rise time that can be kept every day.
2. Set an alarm and get out of bed at that time 7 days a week, however the night went.
3. Keep the difference between workdays and days off under 1 hour.
4. Plan morning light and activity to make getting up easier [synthesis].
Example language (illustrative): "However the night went, you get up at 6:30. A bad night is exactly when sleeping in does the most damage to the next night."
Watch for: Weekend lie-ins, which you can spot in the diary. Explore what the patient gains from lying in and what it costs them the following night.

O. SCT rule 6: No daytime napping (and the planned-nap variant)

[source: Bootzin & Perlis 2011; Spielman et al. 2011]

Purpose / when: Sleep debt carried into the night helps the patient fall asleep quickly, which is a success experience. The original SRT protocol also barred lying down outside the window.
Steps:
1. Default: no naps and no lying down outside bed time.
2. Variant for older adults or severe sleepiness: a brief nap of 30 minutes or less, at the same time every day. Bootzin & Perlis note it is irregular napping that sustains irregular schedules. Later SRT modifications permit or prescribe a nap early in treatment.
3. Record any nap in the diary.
Example language (illustrative): "Let's protect your night's sleep by not spending it in the afternoon. If you really need it, one short nap of no more than half an hour, same time every day, and nowhere near the evening."
Watch for: Unplanned evening dozing in front of the TV. It often goes unrecorded, so ask specifically.

P. Choosing a time out of bed and replacing bedtime behaviours

[source: Bootzin & Perlis 2011 (new pre-bed routine; dim light); [synthesis] for structure]

Purpose / when: Session 1–2. It fills the evening hours freed by SRT and supports SCT rules 2–3.
Steps:
1. Map the hours between the old and new bedtime.
2. Choose low-arousal, dim-light activities in another room.
3. Build a fixed wind-down sequence for the last 30–60 minutes [synthesis on duration].
4. Rehearse what the patient will do on waking at night.
Example language (illustrative): "What will you actually do between 10:30 and 12:30? Let's pick two or three things that are pleasant but not gripping."
Watch for: Engaging activities such as work email or gaming. These delay sleepiness and can push sleep later.

Q. Counter-control (staying in bed but sitting up)

[source: Zwart & Lisman 1979; Muench et al. 2022; Kaplan & Harvey 2013]

Purpose / when: For patients who cannot safely or practically leave the bed: fall risk, disability, pain, shared or single-room living. It was originally a comparison condition, and worked about as well as SCT for difficulty falling asleep.
Steps:
1. If not asleep within the time criterion, sit up in bed. Switch on a dim light and do a quiet activity such as reading or listening to something calm.
2. Lie back down to sleep only when sleepy.
3. Repeat as needed.
4. Muench et al. describe a related form in which the patient deliberately stays awake in bed. Kaplan & Harvey had one patient use relaxation exercises in bed instead of leaving it.
Note: The original Zwart & Lisman condition also asked participants to do some activity in bed for at least an hour a day. That was a design control and is not recommended as clinical practice [synthesis].
Example language (illustrative): "Since getting up at night isn't safe for you, when sleep isn't coming, sit up, switch on the little lamp and read until your eyes get heavy, then lie back down."
Watch for: Lying down while still alert, which brings back the old pattern. Working well: shorter time to fall asleep after lying back down.

R. Barriers-and-facilitators planning and progress review

[source: HABIT protocol v6.0; Kyle et al. 2023; [synthesis] for steps]

Purpose / when: At the end of session 1, and in every titration session; session 3 of HABIT emphasises overcoming barriers.
Steps:
1. Ask what will make the schedule hard: a partner's bedtime, children, evening social plans, sleepiness, boredom, pain.
2. For each barrier, agree a specific plan.
3. Identify facilitators: an alarm across the room, partner support, a prepared lounge.
4. At each review, ask what went well before looking at the numbers. Then look at the diary, adherence and the titration decision.
5. Close with the new window and one adherence goal.
Example language (illustrative): "On which nights is staying up until midnight going to be hardest? What will you do at 10:30 on those nights?"
Watch for: Plans that rely on willpower alone. Push for environmental changes.

S. Bipolar-disorder modification

[source: Kaplan & Harvey 2013; Psychiatric Times summary of Harvey's CBT-I for bipolar disorder]

Purpose / when: When insomnia occurs with bipolar disorder, since sleep loss can trigger hypomania or mania.
Steps (as reported by Kaplan & Harvey):
1. Establish regular bed and rise times, on weekdays and weekends, for 1–2 weeks first. This alone improved SE for many patients.
2. Consider restriction only if SE remains low.
3. Do not restrict patients whose baseline TST is about 5.5 hours or less. The authors judged restriction below about 6 hours not warranted.
4. Replace "bed only when sleepy" with a regular bedtime.
5. Rate mood at every session; they used standard depression and mania rating scales.
6. Keep medication steady, in liaison with the prescriber, and have a safety plan.
7. Modify or suspend the techniques if signs of mania emerge.
Note on versions: The Psychiatric Times summary of CBT-I for bipolar disorder describes a 6.5-hour limit. Its wording is ambiguous about whether this is a floor or a ceiling, and the primary trial paper is not drawn on [uncertain]. A conservative reading is never restricting the window below about 6–6.5 hours [synthesis].
Example language (illustrative): "With bipolar disorder we're careful about sleep loss, so we'll start by just steadying your timing. We'll only trim time in bed if we need to, and never below six hours."
Watch for: Reduced need for sleep, racing thoughts, raised energy. Widen the window and contact the prescriber.

T. Maintenance and relapse prevention

[source: Bootzin 1972 (keep the rules after success); Spielman et al. 2011 (keep the final schedule); Muench et al. 2022 (relapse prevention as a session agenda item); [synthesis] for steps]

Purpose / when: The final session.
Steps:
1. Write down the final schedule and the SC rules that will stay permanent.
2. Name early-warning signs: several nights of poor sleep, the urge to go to bed earlier or lie in.
3. Write a "bad patch" plan: hold the rise time, don't extend time in bed, get up when awake, and re-start a diary. If SE falls, briefly re-run a sleep window from the diary using the same rules [synthesis].
4. Agree when to seek a booster.
Example language (illustrative): "Bad nights will still happen. The trap is going to bed early or lying in to make up for it. If you notice that urge, that's your cue to use what you learned here."
Watch for: Patients who drift back into long time in bed once they feel better. Check at a follow-up appointment [synthesis].

5. Handling problems

[source tags per item; [synthesis] marks my own guidance]

Non-adherence to the window (staying in bed longer, lying in).
- Review the diary non-judgementally and re-link it to the rationale (B).
- Ask what made it hard, then use H (buffer or negotiation) or G (compression) rather than abandoning the treatment [source: Spielman et al. 2011 on negotiation and buffers].
- Name the pattern: in HABIT, withdrawal was most often due to finding SRT hard to implement [source: Kyle et al. 2023].
- [synthesis] Run a small behavioural experiment, such as one week at full adherence, and compare the diaries.
Can't stay up until the window opens (evening sleepiness). [synthesis consistent with Bootzin & Perlis 2011]
- Plan upright activity in another room, brighter light earlier in the evening, and no sofa.
- If unmanageable, widen the window by 15–30 minutes or allow a scheduled short nap earlier in the day (O).
Overwhelm, excessive daytime sleepiness or safety events. [source: Spielman et al. 2011; Kyle et al. 2023 (adverse-event monitoring)]
- Stop the drowsy driving or machinery use.
- Widen the window.
- Screen for sleep apnoea, periodic limb movements and narcolepsy.
- In safety-critical occupations, use compression or SCT alone.
SE won't rise despite good adherence. [source: Penn session slides 2014 (non-response list); Spielman et al. 2011]
- Consider an undiagnosed sleep disorder such as sleep apnoea or periodic limb movements.
- Consider medical or psychiatric illness, substance use, or sleep-state misperception.
- Also consider whether the restriction was too severe to comply with.
- Refer for sleep assessment, or switch to the fixed-increment approach (J).
Clock-watching and anxiety about the "rules". [source: Bootzin & Perlis 2011; Muench et al. 2022; Kaplan & Harvey 2013]
- Remove visible clocks.
- Leave bed by feel, not by minutes.
- If monitoring itself raises anxiety, relax the time criterion. Kaplan & Harvey let one patient stop leaving bed after two weeks.
- Address effortful sleep [synthesis].
Fear of losing sleep ("I'll fall apart"). [synthesis]
- Validate the fear.
- Explain the temporary dip and the floor of 5 hours.
- Frame the first two weeks as a time-limited experiment.
- Show the diary evidence that the person already copes with short nights.
Patient wants to stop early. [source: Penn session slides advise against it]
- Ask what is happening.
- Point to the gains so far.
- Adjust the intensity rather than stopping.
Rupture: the patient feels blamed or bossed. [synthesis]
- Acknowledge that the treatment is demanding.
- Reposition yourself as a co-designer of the schedule.
- Return choices to the patient: where to place the extra 15 minutes, which modification to use.
Partner or household conflicts (different bedtimes, or getting up disturbing others). [synthesis]
- Hold a joint session or give a handout to the partner.
- Agree where the patient will go when out of bed.
- Consider separate wind-down spaces.
Falls risk, mobility limits, pain or hypnotic use. [source: Bootzin & Perlis 2011; Leggett slides]
- Use counter-control (Q).
- Light the route.
- Liaise with the prescriber about medication.
Comorbid depression. [source: Leggett slides; [synthesis]]
- Emphasise "bed only when sleepy" and planned daytime activity.
- Monitor mood.
- Screen for suicidal ideation using standard risk assessment, and act on it before or alongside insomnia work. HABIT excluded current ideation or a recent attempt.
Bipolar disorder, epilepsy, parasomnias, sleep apnoea. [source: Spielman et al. 2011; Kaplan & Harvey 2013; Muench et al. 2022; Leggett slides]
- Use the modification in S for bipolar disorder.
- For the others, avoid or soften SRT: use compression or SCT alone, and liaise with the medical team.
- Muench et al. note that SRT usually produces 1–2 hours of sleep loss, not full sleep deprivation, but advise considering compression or dropping SRT where risk is a concern.
Shift workers. [source: HABIT exclusion criteria; [synthesis]]
- Standard fixed-schedule SRT does not fit.
- Treat or refer for circadian assessment. Do not improvise a rotating window.
Dementia or cognitive impairment. [source: Bootzin & Perlis 2011; [synthesis]]
- The patient may not be able to follow the instructions independently.
- Involve carers. Use a regular schedule and daytime activity rather than full SRT.
Pregnancy. [source: HABIT exclusion; [uncertain] no primary guidance found]
- Avoid SRT, or use only a gentle compression with obstetric input.
Relapse after treatment. [synthesis]
- Re-run the diary.
- Restart the window from current TST.
- Hold for 1–2 weeks, then titrate as in E.

Sources

Sources: