The Complete Guide to Insomnia

Insomnia is usually described in terms of three patterns: difficulty falling asleep, difficulty staying asleep, and waking too early and not getting back off. They often occur together, and all three count. The definition rests on something less obvious than the symptoms, though, which is that the difficulty happens despite adequate opportunity and circumstances for sleep. Somebody who sleeps five hours because they work nights and have small children at home is sleep-deprived. Somebody who lies awake for two hours with nothing stopping them sleeping is describing something else.

The second half of the definition is daytime impairment. Insomnia is increasingly described as a 24-hour disorder rather than a night-time one, because what brings people to a GP is usually the fatigue, the flattened mood, the difficulty concentrating and the sense that daily function has slipped. Poor sleep that leaves you perfectly functional is not the same clinical problem.

NICE divides it by duration. Short-term insomnia means symptoms lasting less than three months, which is often a response to a stressor and may resolve as circumstances change. Chronic insomnia means symptoms that occur at least three nights a week, persist for more than three months, and impair daytime function.

A modified Delphi consensus on UK primary care, published in BMJ Open in 2025, describes chronic insomnia as affecting between 6.8% and 14.9% of UK adults, and as being roughly twice as common in women. Internationally, a 2025 meta-analysis in the Journal of Sleep Research found a pooled prevalence of 12.4% when insomnia disorder was diagnosed through clinician-administered DSM interviews and 16.3% when based on self-reported DSM criteria, with higher-quality studies reporting a lower figure of 10.7%.

The spread in those numbers is itself informative. Estimates vary enormously depending on the instrument used: studies applying one screening questionnaire produced a pooled prevalence of over 30%, while stricter diagnostic interviews produced half that. When you read that "a third of adults have insomnia", the figure is usually counting symptoms rather than the disorder.

The gap between symptoms and recognition is the more striking finding. A study of 163,748 UK Biobank participants in England compared what people reported with what appeared in their primary care records, and found 29% reporting frequent insomnia symptoms against 6% with a corresponding code in their notes. The authors concluded that only a small proportion of people experiencing insomnia symptoms have them recorded by their GP, while noting that primary care data still provides a clinically meaningful measure. Some of that gap is definitional, since reporting symptoms is not the same as meeting diagnostic criteria. Some of it is that people don't raise it, or raise it at the end of an appointment about something else.

Insomnia often begins with something identifiable: illness, bereavement, a stressful job, pain, a new baby, shift work. Sometimes that trigger resolves and the sleep problem continues anyway. Sometimes the trigger is still there, which is why chronic insomnia so often sits alongside persistent pain, anxiety, depression or difficult circumstances. Both situations are common, and the second does not make the insomnia less real or less treatable.

What the behavioural model adds is an explanation for why sleep problems can outlast or outgrow their original cause. The sequence is recognisable to anyone who has been through it. You sleep badly, so you go to bed earlier to catch up, which means more time lying awake. You stay in bed in the morning hoping to recover the lost hours, which weakens the pressure to sleep the following night. You nap. You start monitoring: checking the clock, calculating how many hours are left, noticing every noise. Lying awake in bed for long stretches can gradually build an association between the bedroom and being awake and frustrated, and trying harder to sleep tends to make it less likely, because effort and arousal work against falling asleep.

This matters because it explains why treatment looks the way it does. Where extended time in bed, conditioned wakefulness and sleep effort are maintaining the problem, treatment has to target those directly, and advice about caffeine and screens does not.

Sleep hygiene is the collection of sensible habits most people have already heard: a cool dark bedroom, limited caffeine late in the day, less alcohol, a consistent wind-down, no heavy meals before bed, morning light, exercise earlier rather than later. There is nothing wrong with any of it, and it genuinely supports good sleep.

What it is not is a treatment for insomnia disorder. NICE positions sleep hygiene as supporting advice, with the components of CBT-I driving the larger sustained benefits. The clearest demonstration is that sleep hygiene is what researchers routinely use as the comparison condition in trials of insomnia treatment, and more active treatments reliably outperform it.

There is a second problem that is less often acknowledged. For somebody with established insomnia, hygiene advice can become counterproductive if it is taken to extremes. Someone desperate to sleep can turn a list of tips into a nightly inspection: no screens from six o'clock, no coffee after breakfast, the bedroom temperature checked, lights dimmed on a schedule, every departure from the routine registered as a reason tonight will go badly. That is sleep effort, and sleep effort is one of the things that can maintain the problem.

The useful way to hold it is that sleep hygiene is the foundation you build on and the thing you maintain once sleep is working. Our own Sleep Reset is built around exactly those foundations: the evening routine, the disruptors worth cutting, the body clock, the bedroom and the morning. For someone who sleeps reasonably but wants to sleep better, that is the right place to start. For someone who has been lying awake three nights a week for six months, the foundations are not the treatment.

Cognitive behavioural therapy for insomnia is recommended as first-line treatment for chronic insomnia in adults, including older adults, ahead of sleeping tablets. It is not general CBT with sleep as the topic, but a specific protocol usually delivered over roughly four to eight sessions, built from several components.

Stimulus control rebuilds the association between bed and sleep. The instructions are simple and hard: go to bed only when sleepy, use the bed for sleep and sex only, get out of bed if you have been awake for a while and return when sleepy, get up at the same time every morning regardless of how the night went, and don't nap.

Sleep restriction, covered in the next section, limits time in bed to roughly the amount you are actually sleeping, then expands it as sleep consolidates.

Cognitive work addresses the beliefs that keep arousal high: that eight hours is mandatory, that tomorrow will be ruined, that you have lost the ability to sleep. These thoughts are not irrational so much as unhelpfully certain, and they generate the anxiety that makes sleep less likely.

Relaxation methods give the body something to do other than brace for another bad night.

Sleep hygiene is included, as the minor component.

Everything runs on a sleep diary, which is the backbone of treatment and of follow-up. Clinical experts advising a NICE appraisal committee estimated that 70% to 80% of people respond to CBT-I and around half of those who respond experience long-term remission. Those are expert estimates given in evidence rather than trial results, but they reflect why the treatment is positioned where it is.

Sleep restriction is the component people find hardest to believe in, because it involves spending less time in bed when you already feel short of sleep. It works by using the body's own sleep drive: compressing time in bed builds pressure to sleep, which consolidates fragmented sleep into a more solid block, after which the window is gradually widened as sleep efficiency improves.

The strongest UK evidence comes from the HABIT trial, published in The Lancet in 2023 and funded by the NIHR. Researchers recruited 642 adults with insomnia disorder from 35 general practices across England and randomised them to either four sessions of nurse-delivered sleep restriction therapy plus a sleep hygiene booklet, or the booklet alone. Two sessions were in person and two by phone, and participants kept a sleep diary, agreed bed and rise times, and calculated their sleep efficiency before each consultation.

At six months, the sleep restriction group scored 10.9 on the Insomnia Severity Index against 13.9 for sleep hygiene, an adjusted mean difference of −3.05 (95% CI −3.83 to −2.28, p<0.0001, Cohen's d −0.74). The researchers reported that the effect exceeded clinically significant thresholds defined by the American Academy of Sleep Medicine, and was larger than estimates from a meta-analysis of CBT-I delivered in primary care. The treatment was judged safe and likely to be cost-effective.

Two details make the result more interesting. The comparison group received the thing most people are actually given, and the therapy was delivered by practice nurses after training rather than by sleep specialists. The participants were not easy cases either: mean baseline ISI was 17.5, median duration of insomnia was ten years, 76% had already consulted a doctor about it and 25% were taking prescribed sleep medication.

Three limitations are worth stating. The trial was open-label, so participants knew which group they were in, which matters for a self-reported outcome. The sample was 76.2% female, 97.2% from a White ethnic background and around half university-educated, with a mean age of 55, so it does not represent everyone. And the trial compared one component against sleep hygiene rather than against full CBT-I, so it shows that brief sleep restriction can outperform the usual advice, not that it accounts for most of the benefit of multicomponent treatment.

Sleep restriction also has costs, and they are predictable. Compressing time in bed produces daytime sleepiness in the first weeks, which matters if you drive, operate machinery or work shifts. It is not suitable for everyone: digital and face-to-face programmes screen people before starting, and NICE advises a medical assessment before referral to digital CBT-I for people at higher risk of other sleep disorders, including sleep apnoea, and for people who are pregnant or have other conditions. This is a treatment to do with guidance and a diary, not something to improvise from a paragraph in an article.

This is where the guidance and the reality diverge. CBT-I is the recommended first-line treatment, and NICE's own appraisal committee heard that even where it is available, people with insomnia are often unaware of it. NICE has also acknowledged that the availability of face-to-face CBT-I in the UK is limited.

There are four realistic routes. NHS Talking Therapies accepts self-referral in England without going through a GP, though what is offered varies by service and not every service delivers the full CBT-I protocol. Digital CBT-I is available free on the NHS in some places: Sleepio is recommended by NICE as a cost-saving option in primary care for people who would otherwise be offered sleep hygiene advice or sleeping tablets, and is available nationally in Scotland but only in some areas of England, where provision depends on local commissioning. Sleepstation is an alternative digital programme commissioned by some NHS services, with self-referral where it is available. Both can also be paid for privately, as can one-to-one therapy with a CBT-I trained therapist.

If you want to know what is available where you live, a GP or practice pharmacist is the quickest way to find out, and it is worth asking specifically about CBT-I rather than about sleeping tablets, because the two lead to very different conversations.

Sleeping tablets work, which is precisely why their limits matter. Each has a defined role, and none is intended as an indefinite solution to a problem measured in months or years.

Z-drugs — zopiclone and zolpidem — are the usual short-term option. NICE recommends them for short-term management, and prescribing guidance restricts them to two to four weeks because tolerance and dependence develop with regular use.

Benzodiazepines such as temazepam are now rarely used for insomnia for the same reasons, with dependence developing after as little as two to four weeks of regular use and a withdrawal syndrome on stopping.

Prolonged-release melatonin 2mg is licensed for adults aged 55 and over, for short-term treatment of primary insomnia characterised by poor sleep quality, for a maximum of 13 weeks. It is not recommended for primary insomnia in adults under 55. Immediate-release melatonin 3mg is prescription-only in the UK and licensed for jet lag rather than insomnia. Melatonin bought from overseas websites is not a UK-regulated medicine and its actual content may differ from the label.

Daridorexant (Quviviq) is the newest option and works differently, blocking the orexin system that promotes wakefulness rather than sedating. NICE recommended it in TA922, published in October 2023, for adults with symptoms at least three nights a week for at least three months whose daytime functioning is considerably affected, and only if CBT-I has been tried and not worked, or is unavailable or unsuitable. Unlike the older hypnotics it is not restricted to a few weeks, but treatment should be assessed within three months and stopped if insomnia has not responded adequately, with regular review after that. Clinical trial evidence showed improvement compared with placebo over 12 months.

The pattern across all of them is that medication is positioned around CBT-I rather than instead of it, and the question worth asking a prescriber is what the plan is for after the prescription ends.

Some sleep problems look like insomnia, or sit alongside it, and need addressing in their own right. This is why assessment matters before treatment.

Obstructive sleep apnoea is the most important to rule out. Loud snoring, witnessed pauses in breathing, choking or gasping, and significant daytime sleepiness all point towards it, and it needs investigating rather than treating as insomnia. This is also why digital CBT-I programmes ask screening questions before accepting people.

Restless legs syndrome produces an urge to move the legs, usually worse in the evening and relieved by movement. The result looks like difficulty falling asleep, but the treatment is different.

Pain, nocturia, reflux and menopausal symptoms all fragment sleep, and treating the underlying cause often does more than treating the sleep.

Depression and anxiety commonly travel with insomnia in both directions. Insomnia is a risk factor for later depression, and depression disrupts sleep, which is why the two are usually addressed alongside each other rather than sequentially. Importantly, insomnia occurring alongside another condition is still worth treating in its own right.

Shift work and circadian problems produce the same complaint from a different cause. Somebody whose body clock is running several hours behind everyone else's does not have the same problem as somebody lying awake at a conventional bedtime.

Wearables now hand millions of people a nightly sleep score. For most people that is harmless, and some find the data genuinely motivating. For others, particularly those already anxious about sleep, it can become part of the problem, and sleep medicine has a name for the pattern: orthosomnia, coined in a 2017 case series in the Journal of Clinical Sleep Medicine to describe patients seeking treatment for sleep problems they had diagnosed from tracker data.

Two issues are involved. The first is accuracy. Consumer trackers estimate sleep from movement, heart rate and breathing rather than measuring brain activity, so their verdicts about light, deep and REM sleep are inferences. In the original case series, clinicians described patients trusting their devices over laboratory sleep studies.

The second is behavioural. Some people trying to improve their score spend longer in bed, which is the opposite of what treatment involves, and extra time awake in bed can make insomnia worse. Add a nightly number you keep failing to improve, and a device marketed as the solution can feed the sleep effort that maintains the problem. Orthosomnia is not a formal diagnosis and the research on it remains limited, mostly case series and cross-sectional work, so this is a risk for some users rather than an inevitable consequence of owning a tracker.

If you are working on insomnia, the sleep diary used in treatment is the more useful instrument, precisely because it records how you slept rather than scoring it.