CBT-I for Insomnia: The First-Line Treatment Most People Never Try

Most people with chronic insomnia never hear the letters C-B-T-I. They hear about magnesium, blackout curtains, a stricter digital sunset, or a prescription that works until it doesn't. Which is strange, because cognitive behavioral therapy for insomnia is the treatment that sleep medicine guidelines put first, ahead of every sleeping pill on the market, and it has been that way for a decade.

It is also the treatment most likely to be skipped, because it is not a product. It is a structured, usually six-to-eight-week program of behavioral changes, some of which feel counterintuitive the first week and start making sense the third.

This is what CBT-I actually is, what the trials show it does to your sleep (including the parts it does not fix), how the sleep-restriction component works day one, and how to use the sleep data on your wrist during the process without turning it into a source of anxiety.

TL;DR: CBT-I is a multicomponent behavioral treatment for chronic insomnia and both the American Academy of Sleep Medicine (2021) and the American College of Physicians (2016) recommend it as first-line therapy, ahead of medication. It usually combines sleep restriction, stimulus control, cognitive work, relaxation, and sleep hygiene, and a 2015 meta-analysis of 20 randomized trials (1,162 participants) found it cut time to fall asleep by about 19 minutes and time awake during the night by about 26 minutes, with sleep efficiency up roughly 10 percentage points and benefits that held up at later follow-up. Notice what that list does not include: total sleep time barely moved on average, and its confidence interval crossed zero. CBT-I mostly buys you consolidated, efficient sleep, not more hours. The sleep-hygiene-only version of insomnia advice is explicitly not recommended as a standalone treatment by the AASM, which is the single biggest reason "I already tried better sleep habits" is not the same as trying CBT-I. Sleep restriction temporarily reduces total sleep time and measurably increases daytime sleepiness, so it belongs with a clinician if you drive for a living, operate machinery, or have bipolar disorder, untreated sleep apnea, or a seizure disorder. And if the reason you never tried it was access, digital CBT-I has now been shown statistically equivalent to in-person delivery across 15 head-to-head trials.

YouTube: Related video

What CBT-I actually is

Cognitive behavioral therapy for insomnia is a package, not a single technique. It targets the things that keep insomnia going after the original trigger has passed: the habits, the learned associations, and especially the thoughts.

Here is the loop it is designed to break. Something disrupts your sleep for a week or two. You start going to bed earlier to compensate, and lying awake in bed. Your brain learns that bed is a place where you lie awake and worry about lying awake. Now you sleep badly because you are worried about sleeping badly, and the insomnia has a life of its own, long after the original trigger is gone.

CBT-I attacks that loop from several directions at once. The American Academy of Sleep Medicine's 2021 clinical practice guideline is explicit about the pieces and how strong each recommendation is:

• Multicomponent CBT-I: strong recommendation. This is the full package, and it is what the guideline tells clinicians to use under most circumstances.
• Brief multicomponent therapy: conditional. Shorter versions (often four sessions or fewer) are suggested as an option.
• Stimulus control alone: conditional. Rebuilding the bed-sleep association as a standalone treatment.
• Sleep restriction therapy alone: conditional. Compressing time in bed as a standalone treatment.
• Relaxation therapy alone: conditional.
• Sleep hygiene alone: conditional recommendation against. The guideline suggests clinicians not use sleep hygiene as a single-component treatment. Read that line twice, because it is the reason so many people think they have "already tried everything."

Sleep hygiene, meaning the familiar list of cool dark rooms and no caffeine after 2pm, is a supporting ingredient, not the meal. If your only intervention has been hygiene-level changes, you have not tried CBT-I.

What the evidence says it does to your sleep

The most useful trial summary for a person deciding whether to invest six weeks is the 2015 meta-analysis in Annals of Internal Medicine, which pooled 20 randomized controlled trials of face-to-face, multicomponent CBT-I against inactive comparators in adults with chronic insomnia. Mean age was 56, 64% were women, and every program included at least three of cognitive therapy, stimulus control, sleep restriction, sleep hygiene, and relaxation.

Here is what changed by the end of treatment:

That last row is the one worth sitting with. CBT-I is not a way to manufacture extra sleep. It is a way to stop wasting the time you already spend in bed. For a lot of people with insomnia, that is the actual problem: eight hours in bed converting into five and a half hours of sleep. Fixing the conversion beats adding more time to a broken process.

The authors also reported that the changes appeared to hold at later follow-up points, and that no adverse outcomes were reported in the pooled trials. That durability is the difference between CBT-I and hypnotics: there is no tolerance curve, because you are not taking anything.

There is a newer and more surprising data point too. A 2026 secondary analysis in The Lancet Healthy Longevity of adults aged 60 and over with insomnia disorder compared CBT-I against sleep education therapy, an active comparator. Full remission was more common with CBT-I (34% vs 13%), and people who received CBT-I showed a slower pace of biological ageing on one epigenetic clock, DunedinPACE, with a difference of −0.02 (95% CI −0.04 to −0.01). It is a secondary analysis on a small sample and only one of the three clocks moved, so treat it as a signal worth following rather than a settled fact. But the direction is interesting: fixing insomnia might be a longevity intervention, not just a comfort one.

Sleep restriction: the part that feels wrong

Sleep restriction therapy is the component that makes people quit in week one, and the one most likely to change their sleep permanently.

The logic is simple and slightly brutal. If you spend eight hours in bed and sleep five and a half, you have trained your body to be awake in bed for two and a half hours a night. Shrinking your time in bed shrinks your wake window and rebuilds sleep pressure, so the sleep you do get is deeper and less fragmented. Then you expand back out.

In practice, a clinician typically sets your initial time in bed from your sleep diary: your average actual sleep time plus roughly 30 minutes, with a floor to prevent dangerous sleep deprivation. If your diary says you sleep five hours and 45 minutes, your prescription might be six hours and 15 minutes in bed, with a fixed wake time and a bedtime calculated backwards from it. If you wake before your alarm, you stay in bed only until you have hit something close to your assigned sleep window, then you get up.

Then comes the titration. Each week (or two), if your sleep efficiency, the percentage of time in bed actually spent asleep, comes in around 85% to 90% or better, you add back roughly 15 to 30 minutes. If efficiency is poor, you stay put or tighten slightly. This is the part where having objective data helps, and where it can also mislead, which we will get to.

Fair warnings, because this is a real intervention with real short-term costs:

• Your total sleep time will drop before it rises. In a 2014 laboratory study in the journal Sleep, 16 patients with psychophysiological insomnia completing a four-week sleep restriction protocol lost an average of 91 minutes of objectively measured sleep on the first night, 78 minutes on night eight, and 69 minutes by night 22, even as their subjective sleep and insomnia severity improved.
• Daytime sleepiness and vigilance get worse during the acute phase. The same study found increased daytime sleepiness and objectively impaired performance on a psychomotor vigilance task during the acute weeks, recovering by the three-month follow-up. Sleep restriction is not the week to learn to drive a truck.
• Some conditions need clinician involvement first. Clinicians typically screen for bipolar disorder, untreated obstructive sleep apnea, seizure disorders, and safety-sensitive occupations before prescribing sleep restriction, because compressed sleep can destabilize mood and worsen daytime risk. If any of those apply to you, this is a therapy, not a self-experiment.

Stimulus control is the gentler sibling and works on the same principle: bed is for sleep (and sex), nothing else. No laptop, no scrolling, no lying there for 40 minutes mentally replaying a meeting. If you are awake for what feels like 20 minutes or more, you get up, sit somewhere dim and boring, and return when sleepy. It is tedious for two weeks and effective for years.

A realistic CBT-I timeline

People abandon CBT-I because week one feels like it is making things worse. It usually is, briefly. Knowing the shape of the curve helps.

The first two weeks are the price of admission. If you go in expecting that, you are far more likely to finish.

How your tracker fits in (and where it hurts)

If you wear a watch that tracks sleep, you already have two of the three things CBT-I works from: how long you were in bed and how much of it you were asleep. That makes your sleep efficiency a genuinely useful titration input, alongside a written diary. A wearable that gives you a daily sleep score and a rolling view of your sleep trends is doing part of the measurement work, which is what tools like Century are built around: turning wearable data into a score you can actually act on rather than a nightly verdict.

Three rules keep the data helpful instead of harmful:

• Trust trends, not single nights. One bad Tuesday means nothing and should never trigger a protocol change. Look at weekly averages before you touch anything.
• Do not check your sleep score at 3am. This is not a gentle suggestion. Sleep tracker checking in the middle of the night is one of the clearest ways to convert mild insomnia into a chronic problem, and it has a name.
• Give the tracker less authority than your own experience when they disagree. If you slept well and your watch says 61, you slept well.

That last point is the well-documented dark side of sleep tracking. Researchers described "orthosomnia" as far back as 2017: patients pursuing a perfect sleep score so intently that the pursuit itself worsened their insomnia, sometimes treating tracker output as more legitimate than their own experience of the night. If your wearable is making your sleep worse, the orthosomnia guide is worth reading before you start CBT-I, and it is a conversation worth having with your clinician too. It also helps to know how sleep scores are actually calculated, because once you see how much of the score is duration and consistency, it stops feeling like a grade on your character.

If you cannot get a therapist

Access is the main practical barrier, and it is improving. A 2024 systematic review and meta-analysis in JMIR Mental Health pooled 15 randomized head-to-head comparisons of in-person CBT-I against eHealth CBT-I, meaning telephone, video, web-based, and app-delivered versions, with 1,083 participants. Conventional comparisons slightly favored in-person delivery, but the two formats were statistically significantly equivalent for most sleep and non-sleep outcomes. The author's conclusion was blunt: eHealth CBT-I is a clinically relevant alternative, which is a roundabout way of saying you do not have to live near a sleep center to get the treatment.

Practical routes, roughly in order of evidence strength:

• A clinician or sleep psychologist, in person or by video. The gold standard, and the version most trials tested.
• Digital CBT-I programs built on the same components, ideally one with published trial data behind it. Several are prescription-grade in the UK and available directly elsewhere.
• Structured self-help with a diary, if nothing else is available. This is weaker, but still not the same thing as sleep hygiene alone, because it includes restriction and stimulus control rather than just a list of rules.

One companion move that pairs well: if you are currently using something to fall asleep, whether that is melatonin or a prescription hypnotic, talk to your clinician about how it fits alongside behavioral treatment rather than instead of it. In 2026 the AASM weighed in on combination treatment and suggested CBT-I plus medication over medication alone, while suggesting against combination treatment over CBT-I alone. Translation: medication is not the enemy; medication by itself, indefinitely, is the wasted opportunity.

Mistakes that make CBT-I fail

• Starting sleep restriction on a week you need to be sharp. Travel, exams, a new baby, a night-shift rotation: pick a week that can absorb some fatigue.
• Changing five variables at once. If you also start a new supplement, a new workout plan, and a new bedtime routine, you have no idea which change worked.
• Keeping the diary in the bed. The diary is the treatment's instrument panel, not a bedside companion to obsess over at 3am.
• Napping to compensate. Long naps drain the sleep pressure that restriction is trying to build. If you must nap, keep it short and early, and know it will affect the titration math.
• Quitting in week two. The protocol is designed to be uncomfortable before it is effective. Week two is the most common exit.
• Expecting more hours instead of better hours. If you are comparing total sleep time before and after, you will be underwhelmed. Compare sleep efficiency and how you feel at 10am.

The bottom line

Chronic insomnia is one of the most common reasons people see a doctor, and one of the few conditions where the best-evidenced treatment is not a prescription. CBT-I beats sleep hygiene alone, matches medication without the tolerance problem, and the measured benefits in the pooled trials are between 19 and 26 fewer minutes of awake time per night, plus a roughly 10-point jump in sleep efficiency that shows up on your wrist as well as your diary.

Measure your baseline first. If you wear a watch, look at your last two weeks rather than last night: your average time in bed, your average time asleep, and your sleep consistency. That is roughly the intake paperwork a sleep clinician would ask for. Then get the behavioral treatment, use the data to titrate rather than to judge yourself, and give it six weeks before you decide it did not work.

Century AI helps you understand your body with a daily health score, recovery score, and sleep insights — using the watch you already wear.