Cognitive behavioural therapy for insomnia (CBT-I): the method that works better than pills

If there were a treatment for insomnia that works about as well as sleeping pills in the short term and better in the long term; that has no side effects of morning grogginess or dependence; and whose effects last for years after treatment ends — you would probably want to know about it, wouldn’t you? It exists. It is called cognitive behavioural therapy for insomnia, CBT-I for short, and American and European clinical guidelines unanimously place it first: before medication, not after it.

And yet most people with chronic insomnia have never heard of it. They have heard of melatonin, lavender, magnesium and “put your phone down” — but not of the method with by far the strongest evidence behind it.

In this piece I’ll lay CBT-I out ingredient by ingredient: what exactly it contains, why each part works, what the research says, how treatment goes in practice — and when it isn’t enough on its own.

Why CBT-I doesn’t repair sleep, but the conditions for it

First, the key idea everything else follows from. Chronic insomnia is as a rule not maintained by what set it off (stress, loss, illness), but by what has woven itself around it: a learned association between bed and wakefulness, an over-aroused nervous system, catastrophic thoughts about the consequences, and a pile of well-meant habits — going to bed early, napping, lying there “just in case” — which actually thin sleep out. I described this mechanism in detail in my pieces on insomnia and the nervous system and on hyperarousal.

CBT-I is built precisely around these maintaining factors. It doesn’t try to “send you to sleep” — sleep cannot be produced on command. Instead it systematically dismantles everything that obstructs sleep and restores the conditions in which sleep returns by itself. It consists of five pillars, which a therapist adapts to you; let’s go through them in turn.

The first pillar: stimulus control — the bed means sleep again

In long-standing insomnia the bed stops being a signal for rest and becomes a signal for a fight: all you have to do is lie down and your head switches on. Stimulus control rewrites that learned association with a few consistent rules: you go to bed only when you are genuinely sleepy; the bed is for sleep and intimacy only (not for work, screens or ruminating); if you haven’t been sleeping for a while, you get up and come back only when sleepy; you get up at the same time every day, whatever the night was like; you don’t nap during the day.

On first reading the rules seem banal — until you compare them with what insomnia actually does: it pushes you to lie down earlier “just in case”, to stay in bed while wide awake, to treat a nap on a free day as compensation. Each of these behaviours is understandable, and each one further cements the bed–wakefulness association. Stimulus control takes it apart point by point.

The rules are simple; being consistent with them is not — and the effect lies precisely in that consistency: after a few weeks the brain reconnects the bed with sleep rather than with a fight. Of all the behavioural techniques, this ingredient has the longest history of evidence and is effective even on its own; I described how to use it during night waking in my piece on waking at three.

The second pillar: restricting time in bed — the strangest and the strongest

This ingredient sounds cruel the first time you meet it, so it deserves the longest explanation. Someone with insomnia typically expands their time in bed: lies down at ten, gets up at eight, “to catch at least something”. But in those ten hours they may actually sleep six — stretched out, shallow, broken. Sleep restriction (Spielman et al., 1987) does the opposite: it temporarily shrinks time in bed to roughly what you currently sleep on average (though as a rule not below five and a half hours), with a precisely set bedtime and rising time.

Why on earth would you shorten the night for someone who already can’t sleep? Because of three effects that work at the same time. First, a shorter night means more homeostatic sleep pressure builds up during the day — and greater “sleep hunger” means falling asleep faster and deeper, less broken sleep. Second, sleep becomes concentrated: instead of six hours smeared across ten, you get six hours almost in one piece — and sleep efficiency (the proportion of time in bed you are actually asleep) jumps. Third, and this is the psychologically most important part: after years of insomnia you experience, for the first time in a long while, lying down and falling asleep. The trust in your own sleep that insomnia has eaten away begins to come back. Once sleep holds for a few weeks, the window is gradually widened — by fifteen minutes at a time — until you find the amount that is genuinely yours.

To be honest about it: the first ten to fourteen days are demanding, and daytime sleepiness temporarily increases (which is why the method is adapted for occupations where sleepiness is dangerous, and for certain medical conditions). But in the research this ingredient carries a large part of the effect of CBT-I — and it is also the one that is hardest to carry out alone, without support.

The third and fourth pillars: the head and the body

The third pillar is cognitive work. Insomnia is almost always accompanied by beliefs that feed it: “if I don’t sleep eight hours my body will fall apart”, “tomorrow I’ll be useless”, “my sleep is broken for good”. These beliefs are not innocent — every such thought during a wakeful night triggers a stress response and prolongs the wakefulness. In therapy you write them out, test them against the facts and replace them with more accurate ones (“after a bad night I function worse, but I do function — demonstrably, since I’ve been doing it for months”). This also includes work with the fear of insomnia itself, which is often a stronger maintaining factor than all the habits put together — I’ve devoted a whole separate piece to it — as well as techniques for worry, such as giving it a daily “appointment”, which I wrote about in the worry postponement exercise.

The fourth pillar is relaxation — addressing the body rather than the head. Because chronic insomnia is at its core a state of over-arousal, treatment includes techniques that directly lower activation: a lengthened out-breath and slow breathing, Jacobson’s progressive muscle relaxation, the body scan, a calm evening transition. None of them is a “trick for falling asleep” — they are training for the nervous system, teaching it to come down out of readiness. The effect is built through repetition, not through one magical evening exercise.

The fifth pillar: hygiene — the foundation, not the cure

The fifth pillar is sleep hygiene: a cool, dark bedroom, moderation with caffeine and alcohol, daylight and movement, an evening without screens. What matters is where it belongs in the building: hygiene is the foundation, not a load-bearing pillar. Research consistently shows that on its own it does not usually resolve chronic insomnia — if it did, you would most likely have stopped having it long ago. I broke down why that is, and what hygiene can and cannot do, in my piece on sleep hygiene. Within CBT-I it has its place as support for the other four pillars — but on its own it is about as effective as treating decayed teeth by brushing them.

What the evidence says — and how it compares with pills

CBT-I is one of the best-researched psychological treatments there is. Meta-analyses of dozens of randomised trials (for example van Straten et al., 2018) show reliable, clinically meaningful effects: falling asleep faster, less wakefulness during the night, better sleep efficiency and quality. On this basis the American College of Physicians (ACP) issued a clear recommendation in 2016: CBT-I should be the first-line treatment for chronic insomnia in adults (Qaseem et al., 2016); the European guidelines (Riemann et al., 2017) and the American Academy of Sleep Medicine (Edinger et al., 2021) say the same.

And how does it compare with medication? The picture is nuanced. Sleeping pills have their place — in an acute crisis, under severe stress or as a bridge they can make good sense, and nobody taking them needs to be frightened about it. But they have two structural limitations: they work only for as long as you take them, and they change nothing about what maintains the insomnia. CBT-I achieves comparable effects in the short term and better ones in the long term — and above all, in the research its effects last for months and years after treatment ends, because what you get is not a substance but a changed system. Ideally the two don’t exclude each other: many people gradually come off sleeping pills alongside CBT-I, in agreement with their doctor. Always discuss taking or stopping medication with a doctor, not with a blog.

What treatment looks like in practice

A typical course of CBT-I runs to four to eight sessions over six to eight weeks — individually, in a group or a combination of both. The backbone of the treatment is the sleep diary: every morning you take two minutes to note when you went to bed, how long it took (by your own sense) to fall asleep, how many times you woke and when you got up. No clock-watching at night — estimates by feel are enough, and we deliberately don’t chase precision, so that the diary doesn’t become a new obsession. On the basis of the diary the therapist adjusts your sleep window each week, and together the two of you work through whatever is sticking.

The first session is mainly a map: a careful review of your history with sleep — when it began, what set it off, everything you have already tried, what your evening and your night look like, which beliefs you carry about sleep. For many people that review alone is illuminating, because they see their insomnia for the first time as an understandable system rather than a mysterious fault. The ingredients are then introduced gradually, as a rule the behavioural ones first and the cognitive ones after — tailored to each person’s picture, because the insomnia of an early riser with a stretched-out rhythm and the insomnia of a tense perfectionist who dreads the bed do not need the same emphasis.

Two things are worth knowing. First, at the start it often gets slightly worse before it gets better — restricting time in bed exacts its price in the first weeks, so it is worth beginning in a period without exceptional demands. Second, there is also digital CBT-I: structured online programmes that have shown good effects in randomised trials (Espie et al., 2012) and are a sensible option when a therapist isn’t available. My experience, though, is that for people whose insomnia is interwoven with deeper themes — boundaries, anxiety, old stories — a live person brings what a programme cannot: someone who hears what your wakefulness is saying.

What can you realistically expect? Not perfect sleep — nobody has that. The goal of CBT-I is sleep that is good enough and that you trust again: that you lie down without fear, fall asleep within a reasonable time, return to sleep after waking in the night, and feel usable in the morning. The occasional bad night remains part of life — the difference is that after treatment it no longer sets off an avalanche. That is exactly why the effects last: what you have gained is not an external prop but a changed relationship between you, the bed and wakefulness, and that goes with you into stressful periods too, when insomnia likes to try to come back.

When CBT-I isn’t (quite) the right address

CBT-I is meant for chronic insomnia — not for every sleep problem. If you snore loudly with pauses in your breathing, wake with a headache and are irresistibly sleepy during the day despite long enough nights, the first step is a doctor, to rule out sleep apnoea. The same goes for restless legs, markedly shifted rhythms (shift work), pain conditions, thyroid problems and pregnancy or menopause — insomnia alongside these conditions is treatable, but the plan has to be set by someone who sees the whole picture. With more severe depression or trauma, too, sleep is often only one voice in the choir, and treatment has to cover more than the night.

And one therapeutic remark of my own, familiar to you if you read my writing: the protocol dismantles the mechanics of insomnia, but it doesn’t always answer the question of what your body is standing guard against at night. For many people CBT-I is entirely enough. For others, once sleep is in order, a theme appears underneath it for which wakefulness was only the messenger — being overloaded, a relationship, grief. That is when the work continues in psychotherapy; what that looks like with me is described on the My Approach page, and if you’d like to check which of this applies to you, you can write to me.

The most important message of this piece, though, is a simple one: chronic insomnia is not a life sentence and it is not a matter of willpower. It is a maintained state with a known mechanism and a treatment that takes that mechanism apart — proven, without a prescription, and with a lasting effect. If you have been carrying it for years: you don’t have to carry it any longer.