Chronic insomnia is one of the most common presenting complaints in primary care. The 2016 American College of Physicians clinical practice guideline recommends a specific first-line treatment for it. That treatment is not zolpidem, not trazodone, not melatonin, not magnesium, not a weighted blanket. It is a form of psychotherapy called Cognitive Behavioral Therapy for Insomnia — CBT-I. Almost nobody gets it.
The Qaseem 2016 ACP guideline was unambiguous. CBT-I first, before pharmacotherapy, for all adult patients with chronic insomnia. The American Academy of Sleep Medicine reached the same conclusion. And yet the majority of primary care visits for insomnia in the United States still end with a Z-drug prescription, an SSRI off-label, or a lecture about screen time. There is a well-evidenced, drug-free, durable treatment for the most common sleep complaint on the planet, and most of the people who need it have never heard its name.
What CBT-I actually is
CBT-I is not "cognitive behavioral therapy for anxiety" adapted to insomnia. It is a specific, structured, five-component protocol that targets the behavioral and cognitive maintenance factors that keep chronic insomnia in place. It is typically delivered across 4 to 8 sessions with a trained clinician, though several digital versions now compress or self-guide the same protocol.
The components, roughly in order of evidence strength:
- Stimulus control. The bed is for sleep only.
- Sleep restriction. Compress time in bed to consolidate sleep, then expand.
- Cognitive restructuring. Break the catastrophic thinking about sleep itself.
- Relaxation training. Give the sympathetic nervous system a route down.
- Sleep hygiene. The one everyone knows, and honestly the least important.
Each is a defined intervention with its own evidence base. Together they form the most effective non-pharmacologic treatment for insomnia in the literature. The magic is not in any single component. It is in stacking them.
The evidence, in the aggregate
The single strongest piece of evidence supporting CBT-I is the Trauer 2015 meta-analysis in Annals of Internal Medicine. Pooling 20 randomized controlled trials with over 1,100 participants, the authors reported clinically meaningful improvements across every insomnia outcome that matters. Sleep-onset latency dropped by roughly 19 minutes on average. Wake-after-sleep-onset dropped by roughly 26 minutes. Sleep efficiency — the percentage of time in bed actually spent asleep — improved by around 10 percentage points. Effect sizes on the standard insomnia severity scales were medium to large.
The comparison with sleeping pills is where CBT-I becomes uniquely valuable. Head-to-head trials of CBT-I versus benzodiazepine-receptor agonists (the Z-drugs) show comparable short-term effects on subjective sleep. At six months and one year, CBT-I effects hold. Z-drug effects do not — patients regress toward baseline once the medication stops, and tolerance builds while they are still taking it. Morin, Espie, and colleagues have replicated this pattern across multiple trials since the 1990s, and the durability finding is one of the most consistent in sleep medicine.
Put plainly: CBT-I works about as well as a sleeping pill while you are doing it, and keeps working after you stop. Sleeping pills stop working when you stop taking them.
Stimulus control — the bed is for sleep only
Stimulus control is the oldest CBT-I component, developed by Richard Bootzin in the early 1970s. The premise is classical conditioning. If you spend hours in bed awake, worrying about sleep, scrolling your phone, or watching TV, your brain learns that the bed is a place where you are awake and cognitively active. The bed becomes a stimulus for wakefulness rather than sleep.
The rules are simple and, for most chronic insomniacs, uncomfortable.
- Get in bed only when sleepy.
- Use the bed for sleep and sex only. No TV, no phone, no laptop, no worry.
- If you have not fallen asleep in about 15 to 20 minutes, get out of bed. Go to another room. Do something quiet and low-stimulation. Return only when sleepy.
- Wake at the same time every day, regardless of how the night went.
- No naps.
Meta-analyses of stimulus control as a standalone intervention find it produces meaningful improvements in sleep-onset latency and wake-after-sleep-onset. It is arguably the single most impactful component of CBT-I. It is also the one people bail on fastest, because getting out of bed at 2 a.m. when you are exhausted is miserable for the first two weeks.
Sleep restriction — the counterintuitive one
Sleep restriction is the component that surprises people. Developed by Arthur Spielman in the 1980s, the protocol works like this. If you are spending 9 hours in bed but only sleeping 6, your prescribed time in bed gets restricted to roughly 6 hours. You will feel worse for a few days. Sleep pressure will build. Sleep becomes deeper and more consolidated inside that shorter window. Then, once your sleep efficiency crosses about 85 percent, you get an extra 15 to 30 minutes of time in bed. Repeat until you are sleeping efficiently across a normal window.
This is the component clinicians are most nervous about handing to patients without supervision — because it works by making things briefly worse to make them permanently better, and because people with medical conditions requiring adequate sleep should not be aggressively sleep-restricted without oversight. Done properly, it is one of the most powerful single interventions in behavioral sleep medicine.
Cognitive restructuring — untangling the story about sleep
Chronic insomniacs share a set of predictable thoughts. If I don't sleep tonight I won't be able to function tomorrow. I need eight hours or I'll get sick. I'm broken because I can't sleep like a normal person. These beliefs are not neutral. They drive sympathetic arousal at bedtime, which drives more sleeplessness, which reinforces the beliefs.
Cognitive restructuring is not positive thinking. It is a structured examination of the specific catastrophic thoughts and the evidence for or against them. A trained clinician walks patients through the actual outcomes of poor nights — you did function; you weren't fired; you did not get sick — and helps them build a more accurate model of what happens when sleep is bad. The effect is a lower arousal set point at bedtime.
The cognitive component is harder to isolate in trials than the behavioral components, but the Espie group's work consistently finds it adds real variance on top of the behavioral pieces.
Relaxation training — the one people know about
Progressive muscle relaxation, diaphragmatic breathing, body-scan practices. These are legitimate parasympathetic-facing interventions that give the sympathetic nervous system a route down at bedtime. As a standalone treatment for chronic insomnia, they are weaker than stimulus control or sleep restriction. As part of a CBT-I stack, they add value — particularly for patients whose insomnia is driven primarily by hyperarousal.
The evidence bar here is lower than for the other components. If you find guided body scans genuinely calming, use them. If they feel like homework, they will not help.
Sleep hygiene — the least important component, and the only one anyone gets told about
This is the honest part. Every sleep article ever written by anyone who has not read the CBT-I literature is about sleep hygiene. No caffeine after noon. Dark room. Cool room. No screens. Consistent bedtime.
These are real recommendations. In the CBT-I framework, they are also, by a large margin, the weakest component. Sleep hygiene as a standalone treatment for chronic insomnia has been evaluated in multiple trials. It performs roughly at the level of a placebo control — better than nothing, but well below stimulus control or sleep restriction.
Why does everyone hear about sleep hygiene and no one hears about stimulus control? Because hygiene is easy to write about. It generates listicles. It sells blue-light glasses and smart clocks. Stimulus control makes you get out of bed at 2 a.m. and sit in a chair. Guess which one goes viral.
Sleep hygiene is not useless. It is the least important part of the protocol. Most people who "tried everything" for their insomnia tried the least effective piece and concluded the treatment does not work.
Why almost no one is delivering it
If CBT-I is first-line by two of the most conservative medical bodies in the world, why does no one get it?
- Provider supply. The number of clinicians trained specifically in CBT-I is small. In many states, patients find one or two within driving distance, or none.
- Time cost. A prescription pad takes 90 seconds. Referring, coordinating, and completing six sessions of CBT-I takes months.
- Insurance. Behavioral sleep visits are inconsistently reimbursed and often not surfaced by primary care.
- Pharma inertia. The marketing budget behind zolpidem is enormous. The marketing budget behind CBT-I is zero — no one owns it as a product.
Digital CBT-I — how to actually access it
The provider bottleneck is real, and it is why digital CBT-I platforms matter. These are self-guided or lightly-coached versions of the same protocol, delivered through app or web, evaluated in real randomized trials.
- SHUTi — one of the first digital CBT-I platforms, studied in multiple trials by the Ritterband group. Effect sizes comparable to face-to-face CBT-I in some head-to-head comparisons.
- Sleepio — developed by Colin Espie's group. Recommended by NICE in the UK as a treatment option for chronic insomnia. Multiple RCTs support its effect on insomnia severity and daytime function.
- Somryst — the first FDA-cleared prescription digital therapeutic for chronic insomnia in the United States. A packaged, regulated version of the SHUTi protocol.
These are not "sleep apps." They are structured CBT-I programs. If a trained clinician is not accessible, digital CBT-I is a legitimate first step — the evidence for it is stronger than the evidence for most sleep gadgets by an order of magnitude.
Where objective sleep data fits into a CBT-I protocol
Sleep restriction, in particular, is easier to do well with data. The clinician is calculating your sleep efficiency — time asleep divided by time in bed — and adjusting your prescribed window accordingly. Traditionally that has meant a paper sleep diary, which is fine and cheap and used in every trial.
If you want objective data alongside the diary, a non-contact monitor removes the compliance friction of wearing something. The Withings Sleep Tracking Mat sits under the mattress and captures sleep stages, wake events, heart rate, and snoring across the night without a wearable. In a sleep restriction protocol, having a nightly objective read on sleep efficiency and wake-after-sleep-onset is genuinely useful — you can see whether the window is compressing sleep the way it is supposed to.
A caution. The data is a tool for the protocol, not a substitute for it. Checking a sleep score every morning can, in some cases, make things worse — a phenomenon called orthosomnia. The mat is useful in a CBT-I context because a program is using the data to make specific decisions. Outside that context, it becomes a scoreboard for something you are already anxious about.
What CBT-I is not
- CBT-I is for chronic insomnia — trouble falling or staying asleep at least three nights a week for three months. Acute insomnia usually resolves without treatment.
- CBT-I does not treat sleep apnea. If snoring, witnessed apneas, or excessive daytime sleepiness are in the picture, get a sleep study first.
- CBT-I does not treat restless legs, circadian rhythm disorders, narcolepsy, or parasomnias. Different diagnoses, different treatments.
- CBT-I can run alongside a medication taper, but tapering a Z-drug should be supervised by the prescriber.
The honest bottom line
If you have chronic insomnia and your treatment plan has been "try melatonin, then try Ambien, then try trazodone," you have been offered second-line treatments in a first-line problem. The American College of Physicians and the American Academy of Sleep Medicine both agree on what should come first. Ask your provider for a behavioral sleep medicine referral, or start a digital CBT-I program — Sleepio and Somryst are both FDA-cleared or NICE-approved and evaluable at home.
The five components: stimulus control, sleep restriction, cognitive restructuring, relaxation training, sleep hygiene. In that rough order of impact. Sleep hygiene is not the treatment. It is the least important part of the treatment.
CBT-I is the closest thing sleep medicine has to a well-evidenced, drug-free, durable answer to chronic insomnia. The reason you haven't heard about it is not because it does not work. It is because no one is selling it to you.
