This programme is a treatment manual written for mental health professionals. It assumes clinical training, the ability to make a diagnosis, and access to supervision. It replaces neither your clinical judgement nor your professional responsibility. The diagnostic criteria here are reformulated in our own words and never reproduced: consult the original manuals for the letter of the text.
1. The programme at a glance
Indication. Chronic insomnia disorder in adults: difficulty falling asleep, prolonged night-time waking, waking too early, or unrefreshing sleep, on at least three nights a week for three months, with daytime consequences.
Reference model. Cognitive behavioural therapy for insomnia, recommended ahead of any medication by the American Academy of Sleep Medicine (Edinger et al., 2021), by the European Sleep Research Society (Riemann et al., 2017, 2023) and by the American College of Physicians (Qaseem et al., 2016). It rests on Spielman's three-factor model, on the two sleep-regulating processes described by Borbély, and on Harvey's cognitive model.
Format. Eight individual sessions of 50 minutes, weekly then spaced out, followed by two booster sessions at one month and three months. Eight sessions is the upper end of the evaluated dose: most trials use four to eight sessions, and there is no reason to do more. This is the shortest programme in the series, and that is a merit.
Target mechanism. Not obtaining eight hours, but four changes: restoring sufficient sleep pressure, giving the bed back its function, taking effort and control out of the night, and undoing the beliefs that keep the worry going.
| Session |
Focus |
What the session produces |
| 1 |
Assessment, differential, diary handed over |
Sleep diary in place |
| 2 |
Formulation, the two processes, the calculations |
Sleep window calculated |
| 3 |
Sleep restriction and stimulus control |
The five rules, written |
| 4 |
Adjustment, and the difficult week |
Window revised on the figures |
| 5 |
What goes round in the head at night |
Worry period, things put into words |
| 6 |
Beliefs about sleep |
Two beliefs tested |
| 7 |
Sleeping tablets, the day, light |
Withdrawal plan if applicable |
| 8 |
Stocktake, plan, relapse prevention |
Summary sheet written by the patient |
What the patient takes away. Seven printable worksheets, listed in section 29 and downloadable from this page: sleep diary, my calculations, my window and my rules, what goes round in my head, what I believe about sleep, my day, my plan.
What sets this programme apart from sleep hygiene advice. Sleep hygiene on its own does not treat chronic insomnia: that is established, and yet it is what most of these patients have already been given — often several times over. The two components that do the work are restriction of time in bed and stimulus control. They are counter-intuitive, they are unpleasant for ten days, and they work. A programme that replaces them with common-sense recommendations is not this treatment.
2. Before you start
Who this programme is for
This text is addressed to psychologists, psychiatrists, psychotherapists, mental health nurses and doctors trained in cognitive behavioural therapy. It assumes that you can conduct a diagnostic interview and that you know the sleep medicine presentations that have to be ruled out before treating.
It is not written for patients. It contains the calculations, the thresholds, the precautions and the criteria for non-response, and a patient would mostly find in it fresh grounds for worrying about their sleep — which is exactly the mechanism to be undone.
The prior question: is this really insomnia?
This is the most important decision in the manual, and it is taken in session 1.
This programme treats chronic insomnia disorder, that is, a sleep complaint with daytime consequences, in someone who has the opportunity and the environment to sleep.
It does not treat short sleep without a complaint. Some adults sleep six hours and are perfectly well. The absence of daytime consequences rules out the diagnosis, whatever the number of hours.
It does not treat a lack of opportunity to sleep. A patient who goes to bed at one and gets up at five because their work demands it does not have insomnia, they have sleep deprivation. The treatment is another matter.
It does not treat the three presentations that must be ruled out first: obstructive sleep apnoea, restless legs syndrome, and circadian rhythm disorders. All three are common, they present with the same complaint, and they get worse if restriction of time in bed is applied to them. Section 7 is devoted to them, and it is not optional.
What this programme is not
It is not a set of sleep hygiene tips. Those are useful in prevention and insufficient in treatment: several trials use them as a control condition, precisely because they have little effect.
It is not a drug treatment, and it needs none. It does, however, make it possible to come off a sleeping tablet, which is one of its most useful indications: see section 25.
It is not a treatment for depression, generalised anxiety or post-traumatic stress disorder. Those three are almost always accompanied by insomnia, and the important point is that there is no longer any reason to wait: see section 8.
It is not a relaxation therapy. Relaxation appears in the programme as an accessory tool, and it fails when it becomes an effort to sleep.
How to use it
Read the whole thing before the first session, in particular sections 7, 10, 23 and 24: the differential, how to read the diary, and the conduct of the two active components. Those are the four places where people go wrong.
Every session is described on the same frame: the objective, a step-by-step run-through, what you say, the common mistakes, and the criterion for moving on.
One warning specific to this treatment. It works by making things worse before making them better: the first week of restriction is hard going, and the patient will be sleepier than before. A patient who has not been warned drops out at session 4. Warning them is not a courtesy here, it is a component of the treatment.
3. The clinical picture
What the patient describes
They say they no longer sleep, and they say for how many years. The figure is often long: five years, fifteen years. They have tried herbal remedies, melatonin, one sleeping tablet, then another, an app, and the sleep hygiene sheet their doctor printed out for them.
They describe three things precisely: how long they take to fall asleep, what happens when they wake at three in the morning, and the state of their day. And they describe a fourth thing, more important than the other three and which they do not present as a symptom: the time they spend in bed trying.
The sentences that recur: "I dread the evening", "I have tried everything", "if I do not get seven hours I am done for the next day", and "I fall asleep in front of the television and I am wide awake the moment I go up to bed". That last one is very instructive: it says that the bed itself has changed function.
The three forms of the complaint
Difficulty falling asleep. The patient takes more than thirty minutes, often far more. It is the most frequent form in younger patients, and the one where anticipatory worry is most visible.
Prolonged night-time waking. The patient wakes once or several times and does not get back to sleep. The most frequent form after fifty.
Waking too early. Waking at four with no possibility of getting back to sleep. Should prompt thought of depression, and of an advanced circadian rhythm disorder in older people.
These forms mix and change over the life of the same patient. They do not call for different treatments.
The daytime consequences, which make the diagnosis
This is the part patients describe least spontaneously and that has to be asked about: fatigue, irritability, difficulty concentrating and remembering, low mood, reduced motivation, mistakes, sleepiness at the wheel.
One important point: in insomnia the complaint is fatigue more than sleepiness. A patient who genuinely falls asleep everywhere during the day — in meetings, at the wheel, in front of the television at five in the afternoon — probably does not have simple insomnia: look for apnoea, restless legs syndrome, narcolepsy, or sleep deprivation.
What keeps insomnia going
This is the heart of the model, and it is what the treatment attacks. Five mechanisms, all of them accessible.
Time in bed, lengthened. This is maintaining factor number one. The patient who sleeps badly goes to bed earlier, gets up later, naps, stays in bed at the weekend. They increase their time in bed to catch some sleep, and they get the opposite: the same sleep spread over more hours, hence sleep that is more fragmented and lighter, and weakened sleep pressure. It is the most effective and the most counter-intuitive of the loops.
The bed as a place of wakefulness. By learning. Through spending hours in it awake, worrying, reading, looking at the clock, the bed and the bedroom stop being sleep cues and become cues for wakefulness and anxiety. This is what the patient observes when they fall asleep on the sofa and wake up on the way upstairs.
The effort to sleep. Sleep is the only function that retreats when it is pursued. Everything the patient does in order to sleep — concentrating, breathing in a certain way, counting, forcing themselves — wakes them up. This idea should be laid down early, and it is liberating.
Worry about sleep and its consequences. "I will be incapable of working tomorrow." "I am going to fall ill." "My brain is being damaged." That worry produces arousal that prevents sleep, and it occupies the night.
Monitoring. Looking at the clock, working out how much is left, watching for the sensations of falling asleep, consulting the data from a smartwatch. All of these increase arousal, and the last deserves a mention of its own: consumer wearables measure sleep stages poorly, and consulting them daily creates a fresh worry, described in the literature under the name of orthosomnia. Have them taken off.
Epidemiology, in two useful figures
About a third of adults report insomnia symptoms, and 6 to 10 % meet the criteria for a full insomnia disorder. It is the most common sleep disorder, and one of the most frequent reasons for consulting in general practice.
Chronic insomnia is an independent risk factor for depression: it often precedes it by several months. That is a strong argument in favour of treatment, and it should be given to the patient.
4. The model that guides this programme
The two processes that regulate sleep
To be explained to the patient in session 2, with a drawing. It is the most profitable explanation in the whole programme, because it makes what you are about to ask acceptable.
Sleep pressure. It accumulates the longer one stays awake, and it discharges through sleeping. This is the homeostatic process described by Borbély. Two practical consequences: the longer the waking day, the faster the onset of sleep and the deeper it is; and every nap, every doze in front of the television, every lie-in discharges the pressure and weakens the following night.
The internal clock. It sets the times when sleep is possible and the times when it is not, independently of tiredness. It is regulated mainly by light and by regularity of rising time. One major practical consequence: the rising time is the anchor point of the whole system, and that is why it is fixed in this programme, weekends included.
Spielman's three factors
The most useful model for formulating a case, because it shows the patient why their insomnia has outlived its cause.
The predisposing factors. A constitution: a tendency to wake easily, background anxiety, family history, age, sex. We do not treat these.
The precipitating factor. An event: a bereavement, a birth, an illness, a move, night work, a period of stress. It is often identifiable, and it is often long since over.
The maintaining factors. What the patient has put in place to cope: staying in bed longer, napping, going to bed earlier, cancelling the next day's activities, taking a sleeping tablet, monitoring. These are what make the insomnia last, and they and they alone are what this treatment changes.
This point should be said to the patient: we are not going back to the cause, because it is no longer there. We are going to undo what has replaced it.
The cognitive model
Harvey (2002) described the loop that occupies the night: excessive worry at bedtime, selective attention to signs of insomnia and to bodily sensations, a distorted perception of time spent awake and of sleep quality, and safety behaviours — going to bed early, cancelling, taking it easy the next day — that prevent the discovery that one copes despite a bad night.
That loop explains two constant observations: patients overestimate how long they take to fall asleep, and underestimate how long they sleep. This is not bad faith, it is a well documented perceptual fact, and you need to know it before reading a diary.
The four levers of the programme
Restriction of time in bed. The most powerful lever. Time in bed is reduced to bring it closer to time actually slept, which restores sleep pressure and consolidates the night. Detailed conduct in section 23.
Stimulus control. The bed is given back its single function: you only get into it when sleepy, you do nothing else in it, and you get out of it when you are not sleeping. Detailed conduct in section 24.
The cognitive work. The worries of the night, and the beliefs about sleep.
Disengagement from effort. Stopping trying to sleep.
What the model implies you should not do
Do not aim for a number of hours. The goal is continuous sleep and reduced daytime consequences, not eight hours. The eight-hour rule is one of the chief maintainers of insomnia: it has a patient who needs six and a half hours spending hours in bed.
Do not settle for sleep hygiene. It is what the patient has already been given, and it did not work.
Do not prescribe relaxation as a method of falling asleep. Used in order to sleep, it becomes an effort and fails. It serves to reduce arousal, not to bring on sleep.
Do not start with the beliefs. They yield far better after three weeks of improved nights than before. The order of the programme is not arbitrary.
Do not restrict before apnoea has been ruled out. It is the one mistake in this manual that can harm the patient.
5. The DSM-5-TR criteria, reformulated
The criteria below are a reformulation in our own words, as an aide-memoire. They do not replace the manual: consult the DSM-5-TR (American Psychiatric Association, 2022) for the letter of the text and the application notes.
Insomnia disorder requires the following.
A complaint of dissatisfaction with the quantity or quality of sleep, together with at least one of three elements: difficulty falling asleep, difficulty maintaining sleep with frequent waking or difficulty getting back to sleep, or early morning waking with no possibility of getting back to sleep.
Clinically significant distress or impairment of functioning — social, occupational, educational, behavioural, or in other important areas.
A frequency of at least three nights a week.
A duration of at least three months.
Adequate opportunity to sleep: the difficulty occurs despite adequate conditions and time for sleep.
No explanation by another sleep disorder, by a substance, or by another mental disorder or medical condition that would better account for the picture — while coexistence is accepted and frequent.
The useful specifiers
DSM-5-TR allows comorbidities to be specified — mental, medical, or another sleep disorder — and the course: episodic, persistent, or recurrent.
One important change from earlier versions: the distinction between primary and secondary insomnia has been abandoned. This is not a subtlety of classification, it is a major clinical decision. It means that insomnia associated with depression or with chronic pain is diagnosed and treated in its own right, without waiting for the other condition to be settled.
What these criteria do not say, and what you must assess
They do not mention time in bed, which is the principal maintaining factor and the target of treatment.
They do not mention naps, nor clock-watching, nor behaviours aimed at protecting the following day.
They set no threshold in hours, and that is justified: there is no normal number of hours. Resist the temptation to set one.
Frequently asked questions
The patient says they have already tried everything, including sleep hygiene.
They are right, and that is a foothold rather than an obstacle. Say it frankly: sleep hygiene does not treat chronic insomnia, that is established, and the American guideline goes as far as advising against using it on its own. Then announce that what you are going to ask of them is different, counter-intuitive, and unpleasant for ten days. Patients who have been handed the same advice sheet three times are often the most receptive to a method that looks serious.
Do you really have to reduce the time in bed of a patient who already sleeps only five hours?
Yes, and that is the whole paradox of the treatment — but with a floor. Never less than five and a half hours, whatever the figures, for two reasons: patients underestimate their sleep, sometimes by more than an hour, and below that the sleepiness becomes dangerous. In someone who reports sleeping two hours, the window is five and a half hours and you explain that to them. What you are reducing is not their sleep, it is the time they spend awake in bed.
The patient is exhausted after ten days and wants to stop.
This is the critical moment, and it is anticipated. If they were warned — in session 1, in session 2, in session 3, and in writing — they hold on. If they were not, you will lose them. Show them their efficiency, which has almost always risen even when sleep time has not moved: that is exactly what is expected at this stage, and seeing it changes everything. And if the risk is real — driving, machinery — widen by thirty minutes rather than lose the patient.
What about the weekend?
Do not leave it free: that is where the treatment comes undone. Getting up at seven on weekdays and ten on Sunday shifts the clock every Saturday. Negotiate one extra hour at most, and make it an explicit subject in session 3 — otherwise the patient will assume the weekend does not count. It is also the first warning sign to write into their plan.
The patient falls asleep in front of the television at nine in the evening. Does it matter?
Yes, and it is one of the most frequent and least noticed maintaining factors — the patient does not count it as a nap. That doze discharges sleep pressure at the most costly moment, and it alone explains a great many failures to fall asleep at midnight. Name it, forbid it explicitly, and give them something to do standing up at that hour.
Is polysomnography needed?
Not to diagnose insomnia: requesting it routinely delays treatment and feeds the worry. It is indicated in the face of positive apnoea screening, genuine daytime sleepiness, suspected periodic limb movements or narcolepsy, a complex parasomnia, or insomnia that resists a well conducted treatment. The diagnosis is made at interview and on the diary.
How do I tell insomnia from a delayed phase in a young adult?
By a single question: when they are free of a timetable — holidays, a long weekend — do they sleep well? Someone with insomnia sleeps badly whatever the timing; someone with a delayed phase sleeps very well, but from two until ten. It is the most frequent sorting error, and it changes everything: the treatment goes through morning light and a gradual advance of the timing, not through restriction.
The patient has been taking a sleeping tablet for ten years. Should it be stopped first?
No, the opposite: touch nothing before session 7. The patient needs to have seen their sleep improve through something other than the tablet, or the first bad night of withdrawal will prove to them that they need it. Then draw up the taper with the prescriber, warn about rebound at each step — two to four nights — and remove the as-needed dose before reducing the regular dose.
What do I say to "I need eight hours"?
Give the information once, factually: the need varies from six to nine hours between individuals, it decreases with age, and there is no normal duration. Then test rather than argue, with the correspondence record from session 6: the quality of the night and that of the day track each other far less than they believe, and some of their best days follow bad nights. That record convinces better than any figure you could quote.
The patient checks their smartwatch every morning and worries about their scores.
Have it taken off for the duration of the programme, explaining why: these devices measure sleep stages poorly, and consulting them daily creates worry about figures that do not mean very much — a phenomenon described under the name of orthosomnia. The only record that counts here is their diary, filled in in the morning, without looking at the clock at night.
Should you wait for the depression to be treated?
No, and this is an important change: the distinction between primary and secondary insomnia has been abandoned, and comorbid insomnia is treated in its own right. Better still: treating the insomnia improves the response to treatment of the depression, and residual insomnia after remission is a relapse factor. The only exception is severe depression with major slowing or suicide risk, which comes first.
And in a bipolar patient?
Real caution: sleep deprivation can trigger a manic switch. Do not conduct outright restriction without discussion with the psychiatrist — a wider window, a higher floor, mood monitoring at every session. Regularity of timing is on the other hand a central element of the underlying treatment of that disorder, and you are well placed to install it.
The patient does not fill in their diary.
Make it the subject of the session, without reproach. Without a diary there is no calculation, therefore no decision, therefore no treatment — put it that way. Look for the reason: the form is too complicated, they are trying to fill it in during the night, they want to be exact and look at the clock, or they fill it in from memory in the evening. Simplify, and remind them that what is asked for is rough estimates, not measurements.
How long before it works?
Sleep efficiency often rises from the first or second week of restriction, and that is what to show the patient — even when they do not feel better, since sleep time itself only increases later, through successive widenings. Allow six to eight weeks for a settled result. Announce that timetable in session 1: it is what carries them through the difficult passage.
Eight sessions is few compared with the other programmes.
It is the evaluated dose, and it is a merit rather than a shortcoming: most trials use four to eight sessions, and nothing indicates that doing more improves the result. If the insomnia persists after eight sessions, the answer is not to lengthen: it is to go back over the differential and the application of the rules, point by point.