Insomnia: What to Try Before Reaching for Pills

If you have been sleeping badly three or more nights a week for longer than three months, the treatment every major guideline puts first is not a pill — it is cognitive behavioural therapy for insomnia (CBT-I), recommended as first-line care by both the American Academy of Sleep Medicine and the American College of Physicians. Its central insight is counterintuitive: what keeps insomnia going is rarely whatever started it, but the sensible-looking things you do to make up for lost sleep — early bedtimes, late lie-ins, naps, weekend catch-up. Undoing those compensations resolves most cases in four to eight weeks without medication.

What actually counts as insomnia#

In everyday speech "insomnia" covers everything from one restless night to a decade-long problem. The clinical definition is narrower, and three of its elements are the ones people most often skip.

Frequency and duration. Difficulty falling asleep, staying asleep, or waking far too early — at least three nights a week, for three months or more. Anything shorter is acute or adjustment insomnia, which usually resolves along with the situation that caused it.

Adequate opportunity to sleep. This criterion filters out the most common confusion of all. If you go to bed at 1 a.m., get up at 6 a.m. and feel wrecked, you do not have insomnia — you have insufficient sleep, and the fix is a schedule, not therapy.

Daytime consequences. Fatigue, irritability, impaired concentration, degraded performance, and anxiety about sleep itself. People who sleep five and a half hours and feel fine are short sleepers, not patients. What makes it a disorder is the price paid during the day.

The 3P model: why insomnia outlives its cause#

The most practically useful framework is Arthur Spielman's "three P" model. It sorts everything into three buckets, and the third one explains why insomnia does not simply fade away.

Predisposing factors are what you brought with you: a tendency to worry and ruminate, a reactive nervous system, family history, being female, older age. You cannot change them, but they explain why one person sleeps badly for a week after a divorce and another for three years.

Precipitating factors are the trigger: bereavement, job loss, illness, a new baby, a house move, shift work. Here is the paradox worth saying out loud — by the time someone seeks help, the trigger has usually gone. The new job is found, the baby sleeps, the illness passed. The insomnia stayed.

Perpetuating factors are the core of the problem and the target of treatment. They are the reasonable-sounding attempts to compensate for bad sleep that end up sustaining it:

• going to bed early to "make sure I get eight hours" — so you lie awake, unsleepy, alone with your thoughts;
• sleeping in and catching up at weekends — which shifts your body clock and drains sleep pressure for the following night;
• napping — which spends the adenosine pressure that would have made you sleepy at bedtime;
• cancelling plans and living carefully — life narrows around sleep, and sleep grows more important;
• monitoring — the bedside clock, the tracker, the arithmetic of "four hours left if I fall asleep now";
• catastrophic worry about tomorrow's consequences.

The mechanism is physiological. Sleep pressure accumulates in proportion to time awake, so if you spend ten hours in bed and sleep six, those six get spread thin across ten — shallow, fragmented, subjectively terrible sleep. Meanwhile the bed stops signalling sleep and starts signalling anxious wakefulness: ordinary conditioning, running in the wrong direction.

If you have not yet dealt with the basic conditions for sleep, start there — but do not expect them to cure an established disorder.

Читайте также: Sleep Hygiene: 12 Rules That Actually Work

Why CBT-I comes first#

The 2021 AASM clinical practice guideline (Edinger et al.) recommends multicomponent CBT-I as first-line treatment for chronic insomnia in adults. The American College of Physicians put it even more plainly in 2016: all adult patients should receive CBT-I as the initial treatment, and pharmacotherapy should be considered only after CBT-I has been tried and found insufficient. The European Insomnia Guideline, updated in 2023, gives the same recommendation at its highest evidence grade.

It is unusual for every major guideline to agree, and the reason is the evidence base. Trauer et al.'s 2015 meta-analysis in Annals of Internal Medicine pooled randomised trials of CBT-I and found sleep onset latency reduced by roughly 19 minutes, wake time after sleep onset by about 26 minutes, and sleep efficiency up by around 10 percentage points. The broader 2018 meta-analysis by van Straten et al. confirmed moderate to large effect sizes across the standard outcomes.

Those minute counts undersell the treatment, because two other properties matter more:

1. The effect survives the end of therapy and in several trials continues to improve at 6- and 12-month follow-up. Drugs cannot do this; their benefit stops when the prescription does.
2. The biggest change is in perceived sleep quality and anxiety about sleep. People stop dreading the night, and that changes daily life far more than twenty minutes does.

The components, one by one#

CBT-I is a package rather than a single technique, and the components contribute very unequally.

Component	What it is	What you do	Effect
Sleep restriction	Compress time in bed to match actual sleep	Average your sleep from a diary, set a window, extend by 15 min once efficiency exceeds 85%	Strongest single component; consolidates sleep, cuts latency and night waking
Stimulus control	Rebuild the bed–sleep association	Bed only when sleepy; out of bed after 15–20 min awake; bed for sleep and sex only	Evidence strong enough to stand alone as a therapy; removes anxious wakefulness in bed
Cognitive therapy	Dismantle catastrophic beliefs about sleep	Identify thoughts like 'tomorrow is ruined', test them against evidence, drop the effort to sleep	Reduces pre-sleep anxiety; protects gains after treatment ends
Relaxation training	Lower physiological and cognitive arousal	Diaphragmatic breathing, progressive muscle relaxation, body scan — practised in daytime	Supportive; most useful where tension and racing thoughts dominate
Sleep hygiene	Remove obvious obstacles	Caffeine, alcohol, light, temperature, fixed wake time	Ineffective on its own; a maintenance layer within CBT-I

Sleep restriction: the strongest and the hardest#

This is the component that generates the most resistance and delivers the most benefit.

The arithmetic: if you spend nine hours in bed and sleep six of them, your sleep efficiency is 67% — stretched, light, full of awakenings. Sleep restriction shrinks the window in bed to approximately your actual sleep time, with a floor of about five and a half hours that you never go below. For the first nights you go to bed later, get up at a fixed alarm, and accumulate a deficit.

Then the physiology takes over. Sleep pressure climbs, you start falling asleep in minutes, awakenings shorten, and efficiency rises to 85–90%. Once it holds above 85% for about a week, you extend the window by 15 minutes — usually by going to bed earlier — and repeat until you reach a duration that leaves you rested while keeping efficiency high.

Expect the first one to two weeks to feel worse, not better. Daytime sleepiness increases, concentration suffers, irritability rises. That is the protocol working, not failing — but it is exactly where people quit, which is the main argument for doing this with support rather than alone.

Stimulus control#

This component has evidence strong enough that the AASM recommends it as a standalone treatment. The rules are short:

1. Go to bed only when genuinely sleepy — not on a schedule and not to bank hours. Sleepiness (eyes closing, dropping the book) is not the same thing as tiredness.
2. If you are still awake after roughly 15 to 20 minutes, get up and leave the bedroom. Do not check the time; a clock turns waiting into measurable failure.
3. Outside the bedroom, do something quiet in dim light — a book, an audiobook, undemanding handwork. Not work, not a phone.
4. Return when sleepiness comes back. Repeat as many times as the night requires.
5. The bed is for sleep and sex only. No working, streaming, eating or scrolling.
6. Get up at the same time seven days a week, regardless of how the night went.
7. No naps while the programme is running.

The point is re-learning. Right now, walking into your bedroom triggers anxious arousal, because it has been paired with fruitless lying awake hundreds of times. The "if you're not asleep, get up" rule breaks that pairing within two or three weeks.

Cognitive work on catastrophising#

Anxiety about sleep is an engine of insomnia in its own right, and it is often what remains once the sleep itself has improved. The usual targets:

• Catastrophising consequences: "if I don't sleep, tomorrow's presentation is ruined and I'm heading for dementia." Checked against your own history, the record usually shows you coped and the fear was disproportionate.
• Unrealistic expectations: "I need exactly eight unbroken hours." Healthy sleep includes awakenings, and the requirement is a range, not a constant.
• Sleep misperception: people with insomnia systematically underestimate how much they slept, sometimes by an hour or two.
• Trying to sleep. Sleep is the one function that breaks when you make an effort at it. Paradoxical intention — lying down and gently holding the intention of staying quietly awake — removes the performance pressure.
• Attributing everything to sleep: a difficult day gets blamed on the night before, when mood and workload were doing much of the work.

If the worry extends well beyond sleep and runs continuously, it needs addressing on a wider front.

Читайте также: How to Manage Stress: Methods That Actually Have Evidence

Relaxation training#

A supporting component, but genuinely useful where muscular tension and a noisy head dominate: progressive muscle relaxation, diaphragmatic breathing with a lengthened exhale, autogenic training, body scans.

One caveat: learn the technique during the day, calm, and only then use it in the evening. Started straight in bed, relaxation becomes one more instrument for forcing sleep — which is to say, effort, which is what prevents it.

Why medication is second line#

Sleeping pills work. Z-drugs such as zolpidem and zopiclone, and benzodiazepines, do shorten sleep onset and extend sleep duration, and guidelines still list them as legitimate options. The problem is not efficacy — it is three other things.

The benefit lasts exactly as long as the prescription. Meta-analyses show no retained effect after discontinuation, unlike CBT-I where the gains outlive the treatment. The drug manages a symptom while leaving every perpetuating factor untouched.

Tolerance and dependence. Beyond a few weeks of nightly use the same dose does less, and both physical and psychological dependence develop — the belief that sleep is impossible without the tablet becomes a perpetuating factor in itself. Hence the guideline caps of two to four weeks and the preference for intermittent dosing.

Rebound insomnia. Stop abruptly and sleep gets worse than baseline for several nights. It is a withdrawal effect, not a return of the underlying problem, but it is experienced as confirmation — "see, without it I don't sleep at all" — and the prescription resumes. It is common enough, and misread often enough, to deserve its own section below.

Then there are the side effects that matter in daily life: next-morning grogginess and slowed reaction time, memory impairment, and an elevated risk of falls and fractures in older adults, for whom the benefit-to-risk ratio is particularly unfavourable.

	CBT-I	Sleeping pills
Guideline status	First line (AASM 2021, ACP 2016)	Second line, after CBT-I proves insufficient
Time to effect	2–4 weeks	First night
Effect after stopping	Retained, often continues improving	Lost; rebound insomnia possible
Dependence	None	Tolerance and dependence with prolonged use
Side effects	Temporary sleepiness during sleep restriction	Morning sedation, falls, memory effects
Availability	Limited by the number of trained clinicians	High

Over-the-counter options, briefly. Melatonin is a timing signal rather than a sedative: useful for jet lag and circadian phase disorders, but in classic insomnia it shortens sleep onset by only a few minutes. First-generation antihistamines sedate quickly but lose effect within days and carry anticholinergic burden. Magnesium is worth correcting if you are deficient, and is not a sleep medication.

Читайте также: Magnesium: Forms, Doses and When You Actually Need It

Rebound insomnia: what it is and how to get through it#

Rebound insomnia is a temporary worsening of sleep beyond your pre-medication baseline in the nights after stopping a sleeping pill. It is the single most common reason people who want to come off hypnotics fail to.

The mechanism is adaptation. Z-drugs and benzodiazepines amplify inhibitory GABA signalling; with nightly exposure, the system compensates by dialling that signalling down. Remove the drug and the compensation is still in place, leaving the brain transiently more aroused than it was before treatment ever started. That is why the first nights off are worse than the insomnia you began with — and why the experience is misleading.

What to expect. Rebound is usually most intense on the first two or three nights and settles within three to seven days. Drugs with a short half-life (zolpidem, triazolam, zopiclone) produce a sharper, earlier rebound; longer-acting agents produce a milder one that may not peak until several days after the last dose, which makes it easy to misattribute. Higher doses and longer use both extend it. Anxiety, irritability and a "wired" feeling in the evening often accompany the sleep disruption.

What actually works. The best-tested approach is a gradual taper combined with CBT-I, and the order matters. In a randomised trial by Morin and colleagues (2004), older adults with chronic insomnia on long-term benzodiazepines were assigned to supervised tapering alone, CBT-I alone, or both. The combination produced the highest rate of drug-free participants — around 85% at the end of treatment, against roughly 48% for tapering alone — and the advantage held at follow-up. Behavioural treatment does not merely make withdrawal tolerable; it supplies the sleep mechanism the drug was substituting for.

A workable sequence:

1. Start CBT-I two to four weeks before the taper. Sleep restriction and stimulus control need time to build sleep pressure. Beginning both on the same day stacks the two hardest weeks on top of each other.
2. Reduce slowly. A common schedule is a 25% cut every one to two weeks, held longer if a step is difficult. There is no prize for speed, and a stalled taper is better than a reversed one.
3. Expect the step down to cost two or three nights. Plan reductions for a quiet week — not before travel, a deadline, or anything requiring long drives.
4. Judge a step after a week, not after a night. The trap is reading night two as evidence and reinstating the dose. Rebound is loudest exactly when it is about to end.
5. Keep the wake time fixed throughout. The most common self-sabotage during a taper is sleeping in to recover, which drains the sleep pressure needed for the following night.
6. Do not substitute one sedative for another. Swapping a Z-drug for a nightly antihistamine or alcohol re-creates the same dependence with worse sleep architecture.

Rebound versus relapse. These look identical for a few nights and diverge afterwards. Rebound is short, peaks early and improves without anything being done. Relapse is the original insomnia returning, and it does not resolve on its own within a week. If sleep is still worse than baseline three to four weeks after the last dose, the issue is the untreated insomnia rather than withdrawal — and that is a CBT-I problem, not an argument for restarting the tablet.

Worth noting what does not cause this: melatonin has no described withdrawal or rebound syndrome, and neither do the behavioural components of CBT-I. If stopping something makes your sleep worse for a few nights, the something was almost certainly a sedative.

Digital CBT-I when no clinician is available#

The limiting factor for CBT-I is not evidence but supply: trained providers are scarce almost everywhere. Digital programmes — apps and web courses that walk you through the same protocol with a diary, an automatically calculated sleep window, weekly adjustments and cognitive modules — close part of that gap.

Zachariae et al.'s 2016 meta-analysis of randomised trials of internet-delivered CBT-I found significant improvements in sleep efficiency, sleep onset latency and insomnia severity, maintained at follow-up. Effects are on average smaller than clinician-delivered therapy, but clinically meaningful — and the realistic comparison is usually not "app versus therapist" but "app versus nothing."

How to choose one: the programme must actually include sleep restriction and stimulus control, with the window calculated from your diary entries. If it never asks what time you went to bed, how long you took to fall asleep and how often you woke, it is a relaxation product, not CBT-I.

Your first week, concretely#

If you have decided to start on your own and none of the cautions above apply, this is the sequence.

1. Days 1 to 7: keep the diary and change nothing. In parallel, answer honestly: loud snoring with breathing pauses, an urge to move your legs in the evening, falling asleep at the wheel? If so, see a doctor first — CBT-I can wait.
2. Calculate two numbers. Average sleep time and average time in bed. The ratio is your sleep efficiency; below 85% means there is room to compress.
3. Set the window. Length = average sleep time + 30 minutes, floor 5.5 hours. Pick the wake time first — one you can hold all seven days — and count backwards. Example: you average 5.5 hours and must be up at 7 a.m., so bedtime is no earlier than 1 a.m.
4. Apply stimulus control from day one. Bed only when sleepy, up and out after 15 to 20 minutes awake, no screens and no clock-watching.
5. Remove the compensations. No naps, no weekend lie-ins, no early bedtimes to bank sleep.
6. Plan for the hard part in advance. Leave a book and a blanket in another room for 3 a.m., and move long drives and high-stakes work out of this week.
7. Recalculate after seven days. Above 85%, extend the window by 15 minutes. Below 80%, shorten it by 15. In between, hold for another week.

Realistic expectations: days 7 to 10 are harder than your starting point, by the end of week two or three falling asleep becomes fast, and by week four to six the window has widened toward a comfortable duration. If four weeks of accurate adherence changes nothing, see a specialist — do not apply the rules more strictly.

When to skip the self-help and see a doctor#

CBT-I treats insomnia; it does not treat the other sleep disorders that impersonate it. Go straight to a clinician if any of these apply: loud snoring with witnessed breathing pauses, or waking up choking; heavy daytime sleepiness with episodes of falling asleep while driving or in meetings; uncomfortable leg sensations in the evening with an irresistible urge to move; sleepwalking, shouting or acting out dreams; regularly sleeping more than nine hours unrested; insomnia alongside significant depression, an anxiety disorder or chronic pain; or more than a month of sleeping-pill use.

Chronic insomnia is one of the more treatable conditions in medicine, which is precisely the opposite of how it is usually experienced. It does not need medication first, and it does not yield to a stricter routine. What it needs is for you to stop doing the thing that feels most reasonable when you are sleeping badly: trying to make the sleep up.