We've written about melatonin dosing, caffeine timing, and the evidence hierarchy for sleep supplements — but the actual top of that hierarchy isn't a supplement at all. It's a structured therapy most people with chronic insomnia have never been offered, and that most doctors don't automatically prescribe even though the guidelines say they should.
What CBT-I actually is
Cognitive Behavioral Therapy for Insomnia is a short, structured program built from five components: sleep restriction (temporarily limiting time in bed to match actual sleep duration, which rebuilds sleep pressure), stimulus control (reconditioning the bed as a cue for sleep rather than wakefulness or worry), cognitive restructuring (addressing anxious or distorted beliefs about sleep itself), sleep hygiene, and relaxation training. Sleep hygiene — the part most people already know about — is consistently the weakest single component on its own, which is the whole argument of our piece on sleep hygiene tips versus real treatment; it's the combination, especially sleep restriction and stimulus control, that does most of the work. A typical course runs four to eight sessions, delivered in person, over the phone, or through a structured self-help or app-based program, usually with a sleep diary tracked throughout so the restriction window can be adjusted week to week as sleep efficiency improves.
How it compares to medication
The largest pooled analysis of the field — 87 randomised controlled trials comparing 118 treatments in 3,724 patients against 2,579 untreated controls — found significant effects across essentially every sleep outcome measured.1 The biggest was on insomnia severity itself (Hedges' g = 0.98, a large effect), followed by sleep efficiency (g = 0.71), Pittsburgh Sleep Quality Index score (g = 0.65), wake after sleep onset (g = 0.63) and sleep onset latency (g = 0.57). Number of awakenings (g = 0.29) and subjective sleep quality (g = 0.40) moved less.
Here is the result that matters for setting expectations, and that most CBT-I coverage skips: the smallest effect of all was on total sleep time (g = 0.16). CBT-I is far better at making the sleep you get more consolidated, more efficient and less distressing than it is at adding hours to the night. If your goal is specifically "sleep two more hours," this is not the intervention that delivers it — and nothing else in the category delivers it either. The same analysis found face-to-face treatment of at least four sessions outperformed self-help or shorter face-to-face courses, while results were otherwise robust across patients with and without comorbid disease, younger and older patients, and those using or not using sleep medication.
Head-to-head trials against sleep medications, including zolpidem and temazepam, have found comparable short-term results — but a clear divergence after treatment stops. Medication effects end when the pills stop, often with rebound insomnia; CBT-I's gains are typically maintained after the course finishes.
One honest update to that comparison, added October 2026. CBT-I being the better long-term treatment is not the same as CBT-I being a reliable way off a sleeping tablet, and the deprescribing evidence has moved. A 2019 meta-analysis found CBT-I plus tapering clearly beat tapering alone for discontinuing benzodiazepine hypnotics (risk ratio 1.68); a better-conducted 2026 review of thirteen trials found 1.35, with a confidence interval crossing 1, and concluded there is no high-certainty evidence it reliably achieves complete discontinuation. We go through both, and what benzodiazepines and Z-drugs actually do at the receptor, in what benzodiazepines actually do. If you are currently taking one, do not change it without your prescriber — withdrawal delirium is documented after as little as six weeks of use.
What the actual clinical guidelines say
This isn't a fringe opinion among sleep specialists. In 2016, the American College of Physicians issued a formal clinical practice guideline recommending "that all adult patients receive cognitive behavioral therapy for insomnia (CBT-I) as the initial treatment for chronic insomnia disorder," graded as a strong recommendation on moderate-quality evidence — one of the stronger recommendation gradings a major medical body issues.2 Its second recommendation, that clinicians use shared decision-making to decide whether to add medication where CBT-I alone was unsuccessful, is graded only weak on low-quality evidence — the asymmetry between the two is the point. The guideline notes that "the FDA has approved pharmacologic therapy for short-term use (4 to 5 weeks), and patients should not continue using the drugs for extended periods," and that observational data suggest serious adverse effects such as dementia and fractures may be associated with hypnotic drugs. Its own summary line is blunt: CBT-I "provides better overall value than pharmacologic treatment."
Digital CBT-I is closing the access gap — partially
The biggest practical objection to CBT-I has always been access: it generally requires a trained provider, and there are far fewer CBT-I-trained clinicians than people who need them. Fully automated, app-based digital CBT-I programs were built specifically to solve this, and a 2025 systematic review and meta-analysis in npj Digital Medicine pooling 29 randomized controlled trials and 9,475 participants found they work — producing moderate-to-large effects on insomnia severity compared with control groups.3 Two caveats from the same analysis matter. First, fully automated programs were less effective than therapist-assisted CBT-I. Second, meta-regression found that the type of control group moderated the results — and of the 29 trials, 10 used a waitlist control, which tends to flatter an intervention more than an active comparator does. Completion rate, by contrast, did not moderate outcomes.
| Approach | Effect vs. control | After treatment ends |
|---|---|---|
| In-person / therapist-guided CBT-I | Large on insomnia severity (g = 0.98 pooled); face-to-face courses of 4+ sessions beat shorter or self-help ones1 | Gains typically maintained after the course finishes1 |
| Fully automated app-based CBT-I | Moderate-to-large on insomnia severity, but smaller than therapist-assisted CBT-I, and effect size depends on the comparator used3 | Not yet as well established long-term3 |
| Prescription sleep medication | Comparable to CBT-I short-term; FDA-approved for 4–5 weeks' use2 | Effects end at discontinuation, often with rebound insomnia |
| Any sleep supplement, including ours | Best evidence is for sleep onset only; no supplement has been shown superior to CBT-I4 | Not studied as a durable fix for chronic insomnia |
That last row isn't false modesty. A network meta-analysis in the Journal of Clinical Medicine compared melatonin, light exposure, exercise and a range of complementary interventions for insomnia disorder across 40 studies, and found melatonin supported for sleep-onset difficulties specifically — while concluding that "none of the considered interventions received superior evidence to CBT-I, which should be more widely disseminated in primary care."4 That is the honest position for a supplement company to hold, and it's why we'd rather you read this page than our before-bed supplement guide first if the problem is chronic.
The practical read: an app-based CBT-I program is a genuinely reasonable starting point if a trained provider isn't accessible or affordable, but it isn't fully interchangeable with therapist-guided care. If a self-guided program isn't moving the needle after several weeks, that's a signal to look for human support rather than a reason to abandon the approach. It's also worth being specific about what "digital CBT-I" means before starting one: the evidence above applies to structured, multi-week programs built around the same five components as in-person therapy — sleep restriction and stimulus control included — not to generic meditation or white-noise apps marketed loosely as "sleep aids." A program that never asks you to track a sleep diary or narrow your time in bed isn't delivering the mechanism the trials tested.
Why this doesn't get talked about more
CBT-I isn't a pill, so there's no shelf to put it on and no marketing budget behind it the way there is for supplements or sleep medication. It generally requires a trained provider (in-person, via telehealth, or through a structured digital program), which is a real access barrier compared to buying something online — but it's also the only intervention in the entire sleep category with evidence this consistent, this durable, and this explicitly endorsed by a national physician body. Part of the gap is structural: most primary care training emphasizes pharmacologic management, sleep-medicine referral networks are thin outside major metro areas, and a 15-minute appointment is a hard setting in which to start a multi-week behavioral program. None of that changes what the evidence shows — it just explains why the intervention with the best data is often the one patients hear about last, if at all.
What to expect if you try it
CBT-I is not comfortable in the first one to two weeks. Sleep restriction deliberately shortens time in bed to match your actual average sleep duration, which increases daytime sleepiness temporarily while it rebuilds sleep pressure and consolidates sleep into a tighter window; stimulus control means getting out of bed if you're awake for roughly 20 minutes rather than lying there trying to force sleep — the same principle behind the advice in how to fall back asleep after waking at 3am. Both components work by breaking the learned association between the bed and wakeful frustration, and both tend to feel counterintuitive to anyone whose instinct is to spend more time in bed "catching up." That short-term discomfort, more than the therapy's actual difficulty, is one of the more common reasons people quit early — which is also an argument for doing a structured program rather than trying to freelance the components from a summary like this one.
One thing CBT-I cannot fix is a breathing problem masquerading as insomnia. If you snore heavily, wake unrefreshed regardless of hours, or have been told you stop breathing in your sleep, the right first step is the STOP-BANG screening questions, not a behavioural program or a capsule.
Bottom line
If you're dealing with ongoing insomnia rather than the occasional bad night, the intervention with the strongest and most durable evidence isn't melatonin, magnesium, or any other supplement covered in this series — it's CBT-I, and it's what national guidelines say should be tried before medication. Set expectations honestly, though: its large effects are on insomnia severity, sleep efficiency and time to fall asleep, and its smallest effect by far is on total sleep time. Supplements can be a reasonable complement to the underlying sleep drivers we've covered elsewhere, but they're not a substitute for the treatment actually built to fix chronic insomnia at the source. If a trained provider isn't accessible, a well-studied digital CBT-I program is a legitimate next-best option — just don't expect a supplement to close that gap instead.
This article is for informational purposes only and has not been evaluated by the FDA. It is not intended to diagnose, treat, cure, or prevent any disease. Speak with a healthcare provider before starting CBT-I, any new supplement, or changing treatment for a diagnosed sleep condition, especially if you take medication or have an existing health condition.
References
- van Straten A, van der Zweerde T, Kleiboer A, Cuijpers P, Morin CM, Lancee J. Cognitive and behavioral therapies in the treatment of insomnia: A meta-analysis. Sleep Medicine Reviews, 2018;38:3-16. pubmed.ncbi.nlm.nih.gov
- Qaseem A, Kansagara D, Forciea MA, Cooke M, Denberg TD. Management of Chronic Insomnia Disorder in Adults: A Clinical Practice Guideline From the American College of Physicians. Annals of Internal Medicine, 2016;165(2):125-133. acpjournals.org
- Efficacy of fully automated digital cognitive behavioral therapy for insomnia in adults: a systematic review and meta-analysis. npj Digital Medicine, 2025;8:157. nature.com
- Baglioni C, et al. A Systematic Review and Network Meta-Analysis of Randomized Controlled Trials Evaluating the Evidence Base of Melatonin, Light Exposure, Exercise, and Complementary and Alternative Medicine for Patients with Insomnia Disorder. Journal of Clinical Medicine, 2020;9(6):1949. pmc.ncbi.nlm.nih.gov
Related reading: Sleep Hygiene Tips vs Real Treatment: What the Evidence Hierarchy Actually Looks Like, The Science of Deep Sleep, Melatonin: The Real Dose-Response Evidence and ADHD and Sleep: Why They So Often Overlap.