Learn › CBT-I

How to treat insomnia without medication

The non-drug treatment for chronic insomnia is CBT-I — cognitive behavioural therapy for insomnia. In the American Academy of Sleep Medicine's 2021 guideline it is the only one of the six behavioural treatments assessed that earned a strong recommendation, and no insomnia medication carries a strong recommendation either in the AASM's separate 2017 drug guideline. CBT-I is not a list of tips: it is a structured programme, usually six to eight weeks, that changes how long you spend in bed, what you do when you can't sleep, and how you think about sleep. It does not work for everyone, and no honest programme can promise you a particular outcome — but it is the method with the best evidence behind it. This page walks through what it involves, what the first two weeks honestly feel like, who should talk to a doctor before starting, and when to stop.

Why sleep hygiene advice keeps failing you

If you're reading this at 2am, you have almost certainly already tried the list: cooler room, no screens, no caffeine after noon, blackout curtains, a wind-down routine. And you're still awake. That isn't a personal failure and it isn't because you did the tips wrong.

Sleep hygiene is supporting context, not treatment. The AASM's 2021 guideline went as far as issuing a conditional recommendation against using sleep hygiene as a stand-alone therapy for chronic insomnia. That's a guideline position, not a marketing line. The tips aren't harmful and they're worth keeping — they just don't touch the two things that keep chronic insomnia running.

Those two things are: first, you're spending more time in bed than your body can currently fill with sleep, so the gap shows up as wakefulness; and second, after months of lying there frustrated, your bed has quietly become a cue for being awake and anxious rather than a cue for sleep. A better mattress and a darker room don't address either. CBT-I addresses both directly.

What CBT-I actually asks you to do

CBT-I is multicomponent, meaning several techniques run together. Each one has a specific job:

Sleep restriction — temporarily shrink your time in bed to match how much you're actually sleeping. This raises sleep pressure and pulls fragmented sleep back into one solid block. It is the engine of the whole programme, and it is also the part with real contraindications, covered below.
Stimulus control — rebuild the association between bed and sleep: the bed is for sleep and sex only, you get up when you've been lying awake and frustrated, you keep one fixed wake time, and you don't nap during the programme.
Cognitive work — dismantle the beliefs that keep arousal high. "If I don't get eight hours I'll ruin tomorrow" is doing real damage at 3am, and it can be examined and loosened.
Relaxation and counter-arousal — practical downshifting for a nervous system that has learned to switch on at bedtime.
Sleep education and hygiene — the background conditions. Useful alongside the above; not a treatment on its own.

The part that does the heavy lifting: matching time in bed to sleep ability

Sleep restriction therapy is where most of the improvement comes from, and it's also the part that sounds backwards. You're exhausted, so being told to spend less time in bed feels like cruelty. Here's the logic: if you're in bed nine hours and sleeping six, three hours of lying awake is being trained into your night. Compress the window and the sleep tends to consolidate.

Before you work out a window, read the safety section below. Sleep restriction is a real intervention with a real contraindication list, and it isn't safe for everyone to run unsupervised — that check belongs before the arithmetic, not after it.

In the original protocol from Arthur Spielman, who developed the technique, you start by keeping a sleep diary for one to two weeks — rough estimates, not stopwatch precision. Your prescribed time in bed is then set to your average nightly sleep from that diary, with a floor so the window can never get dangerously short. Spielman's original protocol allowed the window to go as low as about four and a half hours; modern clinical protocols typically floor at five to five and a half hours, and self-help guidance often uses six. The rule of thumb is simple: the less supervision, the more conservative the floor should be.

Quitude never prescribes less than 5 hours 30 minutes in bed, whatever the arithmetic says. That floor was raised from five hours precisely because Quitude is unsupervised self-help — nobody is reviewing you week by week, so the extra caution is worth more than the extra sleep pressure. If you are doing this on paper instead, decide your floor before you start and hold it, especially on the mornings when a tighter window looks tempting.

A worked example. You average five and a quarter hours of sleep across eight hours in bed, and you have to be up at 6:30am. Setting time in bed to your average sleep would put you under the floor, so the floor wins: your window becomes 1:00am to 6:30am, not later. Floors are not a formality — they are the part of the protocol that stops an ambitious reader from doing something unsafe.

How the weekly review works

Sleep efficiency is the share of your time in bed that you actually spent asleep — total sleep divided by time in bed. It's the dial the whole method steers by, and it is reviewed weekly on a week of diary data, never night by night. One bad night is noise; seven nights is a signal. Adjusting your window every morning based on last night is the most common way people make this harder than it needs to be.

You will meet two conventions online, which is why it's worth knowing which one your programme follows. Spielman's original rule expanded the window when efficiency went above 90%, treated 85–90% as a hold zone, and tightened below 85%. Later adaptations lowered the expansion threshold to 85% for older adults, whose achievable efficiency is typically lower — that age split is a modern modification, not part of the 1987 protocol, and it is often misattributed to it.

Quitude uses the Spielman convention, reviewed weekly against seven days of diary data:

Sleep efficiency of 90% or above — your window widens by 15 minutes.
85% to just under 90% — hold. Nothing changes this week.
Below 85% — your window tightens by 15 minutes, and never below the 5 hours 30 minutes floor.
Quitude does not apply age-stratified thresholds — every user gets the same rule.

Before you start: who this isn't safe for, and when to stop

Sleep restriction deliberately makes you sleepier in the short term. That is the mechanism, and it is also a genuine hazard while it lasts. Talk to a doctor before starting — rather than starting and seeing how it goes — if any of the following apply to you:

A seizure disorder — sleep loss can lower the seizure threshold.
Bipolar disorder — sleep loss can trigger mood episodes, including mania or hypomania.
Untreated or suspected sleep apnoea: loud snoring, witnessed pauses in breathing, waking unrefreshed despite plenty of time in bed. That needs assessing first, because it is a different problem with a different fix.
A parasomnia — sleepwalking, night terrors, or acting out dreams. Sleep loss is a known trigger for episodes, and episodes can cause injury.
Narcolepsy, or any condition that already causes excessive daytime sleepiness. Restriction adds to a sleepiness burden you are already carrying.
A safety-critical job or commute — driving for work, operating machinery, clinical shifts. Increased daytime sleepiness is genuinely dangerous in these settings, and the timing of the programme matters.
A raised risk of falls, which includes many older or frail readers. Night-time trips out of bed while sleep-deprived are the specific concern.
Pregnancy, or another significant medical or psychiatric condition you are being treated for.

The driving rule, the nap exception, and when to stop

Whether or not you drive for a living: during the first weeks, do not drive or operate machinery when you feel sleepy. That applies to every reader, not only professional drivers — the school run and the motorway commute are where sleepiness from a self-run programme is most likely to hurt somebody. If you would not trust yourself behind the wheel, don't get behind the wheel.

The programme says no naps, and there is a good reason for it: naps bleed off the sleep pressure the method depends on. But safety outranks the protocol. If you are too sleepy to be safe — especially before a drive — a short nap early in the afternoon is the right call, even though it costs you some progress that week. Take the nap, get through the day, and pick the programme back up tomorrow.

Because nobody is monitoring you, you have to hold the stop rule yourself. Add 30 minutes back to your window and speak to a clinician before continuing if any of these happen: you fall asleep unintentionally during the day; you nod off while driving or at work; or your mood becomes elevated or agitated, or drops significantly. None of those are the treatment working. Sleepiness that is dangerous rather than merely uncomfortable is a signal to stop, not to persist, and adding time back is a sensible adjustment rather than a failure.

If your mood drops badly — if you feel hopeless, or you find yourself thinking about harming yourself — treat that as urgent rather than as something to manage alongside a sleep programme. In the US you can call or text 988 for the Suicide and Crisis Lifeline; elsewhere, contact your local emergency services or crisis line. Sleep can wait; that cannot.

Stimulus control: what to actually do at 2am

The rule people find hardest, and the one that matters most: if you've been lying awake long enough to feel frustrated — roughly 15 to 20 minutes, judged by feel rather than by the clock — get up. Leave the bedroom if you can. Do something quiet, dim and undemanding. Go back only when you feel genuinely sleepy, not merely tired of sitting up.

You may do this three or four times in a night at the start. That is normal and it is the training happening, not the programme failing. Every time you get up, you stop the bed teaching your brain that this is the place where we lie awake.

Don't clock-watch. Turn the phone face down. Nothing good has ever come from knowing that it is 3:47.

What the first weeks honestly feel like

Weeks one and two are the hard part. Sleep restriction causes daytime sleepiness — that isn't a surprise side effect, it's the mechanism at work, because you are deliberately building sleep pressure. Most people feel noticeably groggier before they feel better, and when improvement comes it usually shows up as sleep that is more solid before it shows up as sleep that is longer.

That extra sleepiness is expected, and it is not a sign that you are doing it wrong. It is also a real safety issue while it lasts, and it has a limit — the stop rules above are the limit. Uncomfortable is part of the deal; dangerous is not, and the answer to dangerous is more time in bed and a conversation with a clinician, not more willpower.

For many people who stick with it, nights start to consolidate through the middle weeks: fewer awakenings, less time staring at the ceiling, and the window widening again as efficiency holds up. By the end of a six-to-eight-week programme, a common pattern is a shorter but far more reliable night, plus a set of skills that tend to keep working after the programme ends.

No honest programme can promise you a specific outcome or a specific timeline. CBT-I doesn't work for everyone, and how quickly things shift varies a great deal between people. What is fair to say is that it is the method with the best evidence base for chronic insomnia, and that when it works, the effect tends to hold up over time rather than fading when you stop.

How this compares with sleeping pills

The AASM's 2017 pharmacologic guideline made 14 recommendations about insomnia medications, and every single one was graded weak, with evidence quality mostly low or very low. Several common over-the-counter options — melatonin, valerian, diphenhydramine — received weak recommendations against use for chronic insomnia. That same guideline positions medication as mainly for people who cannot take part in CBT-I, who still have symptoms after it, or who need a temporary adjunct to it.

On durability, a network meta-analysis with a median follow-up of about 24 weeks found long-term remission in roughly 41% of people treated with CBT-I versus 28% with medication, and fewer people dropped out of CBT-I (about 21% versus 39%). Combining the two showed no clear advantage over CBT-I alone. The straightforward reading: medication works while you take it and acts faster in the short term; CBT-I teaches skills that, for many people, persist after the programme ends.

The AASM published a clinical practice guideline on combination treatment in April 2026, and it speaks directly to this. It conditionally suggests CBT-I plus medication over medication alone, and conditionally suggests against combination treatment over CBT-I alone. Both recommendations are conditional and rest on low-certainty evidence, so hold them loosely — but the practical implication is useful if you are already taking a sleeping pill: current guidance leans towards adding CBT-I to what you are on, rather than treating it as a choice between the two.

The behavioural-first position is long-standing. The American College of Physicians recommended CBT-I as the initial treatment for chronic insomnia in all adults in its 2016 guideline, and it is the ACP — not the AASM's 2021 guideline, which assessed only behavioural and psychological treatments and made no medication recommendations at all — that frames short-term medication as a shared decision when CBT-I alone hasn't worked. None of this means stopping anything you are currently prescribed. If you are on a sleep medication, do not change or stop it on your own: talk to your prescriber, who can advise on whether and how to taper alongside a behavioural programme.

How to start this week

First, the check: if you have a seizure disorder, bipolar disorder, a parasomnia, possible untreated sleep apnoea, narcolepsy or existing daytime sleepiness, a safety-critical job or commute, a raised risk of falls, or you are pregnant, talk to a clinician before you change your sleep schedule. With that done, you can begin the groundwork tonight without committing to anything drastic:

Start a sleep diary. One to two weeks, rough estimates only — roughly when you got into bed, when you fell asleep, how long you were awake in the night, when you got up. Fill it in the next morning, not at 3am; chasing exact numbers works against you.
Pick a fixed wake time you can hold seven days a week, weekends included, and anchor to it. This one change alone does real work.
Stop going to bed early to catch up, and skip naps while you build sleep pressure — with the one exception above: if you are too sleepy to be safe, especially before driving, take a short early-afternoon nap.
From tonight: when you are lying awake and frustrated, get up. Return when sleepy.
After one to two weeks of diary data, set your time-in-bed window — never below the floor you committed to, which is 5 hours 30 minutes in Quitude — and review it weekly rather than nightly. Read the safety section above before you set that window, not after.
Know your stop rules before you need them: unintentional sleep during the day, nodding off driving or at work, or a significant change in mood means add time back and speak to a clinician.
If your sleep problems persist, see a clinician anyway — insomnia can also be a symptom of something else that deserves a proper look.

A note on what this page is

This is general information about a well-established treatment method, not medical advice, and it cannot account for your particular circumstances.

It is also worth being precise about where the evidence sits. CBT-I is the method with strong guideline support, and the trial evidence behind it comes from clinician-delivered programmes and from digital programmes that were tested in their own right, such as Sleepio and Somryst. Quitude is not one of those. It is self-help software that delivers the published protocol — the diary, the window, the weekly Spielman titration, stimulus control and the cognitive tools — and it has not itself been through clinical validation. Use it as a way to run the method, not as evidence that this particular app has been tested.

If your sleep problems persist, worsen, or come with daytime symptoms that worry you, please see a clinician.

Frequently asked questions

How long does it take to fix insomnia without medication?

Most CBT-I programmes run six to eight weeks. The usual pattern is that the first one to two weeks feel harder — more daytime sleepiness, because you are deliberately building sleep pressure — with nights starting to consolidate through the middle weeks for many people. Sleep tends to get more solid before it gets longer. It doesn't work for everyone, and how quickly things shift varies a lot from person to person, so treat any specific timeline as a rough guide rather than a promise about your own nights.

Can I do CBT-I on my own, without a therapist?

In principle, yes — the protocol is well documented, and self-help and digital versions have been studied and are widely used, including in primary care. Doing it alone means you are the one enforcing the schedule through the uncomfortable first fortnight, and you are also the one holding the stop rules. Get medical advice before starting rather than running it solo if you have a seizure disorder, bipolar disorder, suspected or untreated sleep apnoea, a parasomnia such as sleepwalking or acting out dreams, narcolepsy or another cause of excessive daytime sleepiness, a safety-critical job, a raised risk of falls, or you are pregnant.

Is sleep restriction therapy safe?

For most healthy adults it is well tolerated, but it isn't risk-free: it deliberately increases daytime sleepiness in the early weeks. Don't drive or operate machinery when you feel sleepy during that stretch — that applies to everyone, not just professional drivers — and if you are too sleepy to be safe before a drive, take a short early-afternoon nap even though the protocol says no naps. It is contraindicated without medical advice for seizure disorders, bipolar disorder, untreated or suspected sleep apnoea, parasomnias, narcolepsy or existing excessive daytime sleepiness, safety-critical work, pregnancy, and a raised risk of falls. Stop rules matter as much as the start rules: if you fall asleep unintentionally, nod off driving or at work, or your mood becomes elevated, agitated or noticeably lower, add 30 minutes back to your window and speak to a clinician rather than pushing through. Protocols also set a floor on time in bed — Spielman's original allowed about four and a half hours, modern clinical protocols typically use five to five and a half, self-help guidance often six, and Quitude never goes below 5 hours 30 minutes.

What if I'm already taking sleeping pills — do I have to stop first?

No, and you shouldn't stop or change a prescribed medication on your own; some sleep medications need a supervised taper. Talk to your prescriber about starting a behavioural programme and let them advise on timing. The AASM's 2017 guideline positions medication as mainly for people who cannot take part in CBT-I, who still have symptoms after it, or who need a temporary adjunct to it, and its April 2026 combination guideline conditionally suggests CBT-I plus medication over medication alone. In other words, if you are already on a pill, adding CBT-I is the direction current guidance points — the two are not in competition in the way people often assume.

Does melatonin help with chronic insomnia?

The AASM's 2017 pharmacologic guideline gave melatonin a weak recommendation against use for chronic insomnia in adults, alongside valerian, diphenhydramine and L-tryptophan. Melatonin has clearer uses for circadian problems such as jet lag or shifted sleep timing, which is a different issue from insomnia. If your problem is lying awake for months despite adequate opportunity to sleep, the behavioural route has substantially better evidence behind it.

What is sleep efficiency, and is 85% the goal?

Sleep efficiency is the proportion of your time in bed that you actually spent asleep. CBT-I uses it to decide, week by week, whether to widen your sleep window. Spielman's original protocol widened the window above 90%, treated 85–90% as a hold zone, and tightened below 85%; later adaptations lowered the widening threshold to 85% for older adults, which is a modern modification rather than part of the 1987 protocol. Quitude follows the Spielman convention on seven days of diary data — widen 15 minutes at 90% or above, hold between 85 and 90%, tighten 15 minutes below 85%, never below the 5 hours 30 minutes floor — and applies the same thresholds to everyone regardless of age. Treat it as a dial the programme adjusts weekly, not a score to chase every night.

Try the program free

Quitude runs the CBT-I protocol from your own sleep diary — the program is free, there is no account, and your data stays on your device.

Start the CBT-I program