When sleep goes wrong, it usually goes wrong in patterns. The system breaks in identifiable ways that medicine has names for, and most of those names point to causes that respond to specific interventions. The popular framing of insomnia as a generic problem to be solved with sleep hygiene tips and possibly a pill misses how much variation exists underneath the surface. Different sleep problems have different mechanisms and different best treatments.
This page covers the clinical territory. What insomnia and sleep apnea actually are, and the circadian rhythm disorders that don’t quite fit either category. We also attempt to look at when sleep medicine has something to offer that you can’t replicate with self-directed practice, where I believe the medical system gets it wrong, where pharmaceutical incentives shape what gets prescribed, and how to be a more informed patient when you do enter the clinical pathway.
The mechanisms behind normal sleep are in Sleep & Circadian Rhythm Basics. The protocols and tools for self-directed sleep improvement are in the Cheat Sheet. What this page does is the part in between: when self-directed approaches aren’t working, when something is genuinely wrong, when a clinical evaluation is the right next step, and what to expect when you take it.
A note before the rest: this page is here to help you navigate the medical system rather than replace it. If sleep is significantly affecting your life, see a clinician. This isn’t a substitute for proper assessment, but there are things you can do to improve your chances of living a healthier life.
Most people have occasional sleep problems. Stress before a big event, jet lag, a sick child waking the household, caffeine too late, or a stretch of poor decisions may cause a disturbance. These tend to resolve on their own when the underlying cause goes away. They don’t need medical attention; they need basic adjustments to your environment and behavior. The Cheat Sheet covers the practical tools for these situations.
A sleep disorder is more persistent. Typically defined as occurring at least three nights per week for at least three months in the case of insomnia, or as ongoing patterns for the others. It’s not explained by a temporary cause, it affects daytime function (you’re tired, irritable, cognitively slower, less safe to drive), and it doesn’t resolve reliably.
The clinical distinction matters because the interventions that work for occasional sleep problems often don’t work for genuine sleep disorders, and vice versa. People who try to manage moderate-to-severe sleep apnea with sleep hygiene tips don’t get better, because sleep hygiene isn’t the problem. People who try to treat psychophysiological insomnia with sleep medications often get short-term relief and long-term worsening, because medications don’t address the underlying conditioning that maintains the disorder.
Insomnia is the most common sleep complaint and the one most often mismanaged. The standard diagnostic criteria require difficulty initiating or maintaining sleep, occurring at least three nights per week, persisting for at least three months, with daytime consequences (fatigue, mood disturbance, cognitive impairment, etc.). Affects roughly 10-15% of adults at any given time and substantially more episodically.
The dominant clinical framework for understanding insomnia is Arthur Spielman’s 3-P model (also called the Spielman model or the predisposing-precipitating-perpetuating model), developed in the 1980s and still the foundation of clinical sleep medicine. It identifies three factors that combine to produce chronic insomnia:
By the time someone has chronic insomnia, the original precipitating factor is often long gone. The job stress that triggered the initial poor sleep resolved years ago. What’s maintaining the insomnia now is the perpetuating factors: the conditioning, the effort, the time-in-bed pattern, and the worry. Treating the original stressor doesn’t fix things because the original stressor isn’t the active problem anymore.
This is also why “good sleep hygiene” alone usually doesn’t fix chronic insomnia. Sleep hygiene addresses occasional sleep problems and helps prevent precipitating factors from triggering disorders. It doesn’t address the perpetuating factors that maintain a disorder once it’s established.
Cognitive Behavioral Therapy for Insomnia (CBT-I) is the first-line treatment for chronic insomnia in essentially every clinical guideline that exists. The American College of Physicians, the American Academy of Sleep Medicine, the European Sleep Research Society, and others all recommend CBT-I before pharmaceutical treatment.
CBT-I is a structured 6–8 week protocol that addresses the perpetuating factors directly. The main components include:
Meta-analyses consistently show effect sizes comparable to or larger than those of sleep medications, with benefits that persist after treatment ends rather than disappearing when treatment stops. Charles Morin at Université Laval has been the most influential researcher in this space, with decades of trials demonstrating CBT-I’s effectiveness across populations.
The problem: CBT-I is severely underutilized. Estimates suggest fewer than 10% of insomnia patients in primary care receive CBT-I as first-line treatment, despite it being the recommended first-line treatment. Most receive prescription medications instead. Multiple factors drive this: limited availability of trained CBT-I providers, insurance coverage gaps, patient preference for fast-acting solutions, and time constraints in primary care that favor prescriptions over referrals to multi-week therapy programs.
If you have chronic insomnia, finding access to CBT-I should be a priority over starting medication. Several digital CBT-I programs (Sleepio, CBT-I Coach app from the VA, Somryst) make the protocol more accessible than it used to be. Colin Espie at Oxford developed Sleepio and has been the main researcher on digital CBT-I delivery; the evidence base for digital programs is now strong enough that they’re appropriate first-line options for many patients.
Sleep medications are appropriate in some contexts, but they’re inappropriately used in many.
Melatonin is sold over the counter in the US as a supplement (it’s prescription-only in many countries, including Australia, the UK, and most of Europe). The popular framing treats it as a sleep aid; this is mostly wrong about what melatonin actually does.
What melatonin does: It signals biological night to the rest of the system. It’s produced by the pineal gland in response to darkness and tells the body’s various clocks that the dark phase has begun. Its primary role is circadian, not sedative. Endogenous melatonin doesn’t make you sleepy directly so much as enable the conditions for sleepiness.
What melatonin supplementation does well: Modest effectiveness for circadian rhythm disorders, such as jet lag, delayed sleep phase, and shift work. Used appropriately at low doses (0.3–1 mg) at the right circadian time, it can shift the SCN’s phase. Alfred Lewy at Oregon Health & Science University did much of the foundational work on melatonin’s phase-shifting properties.
What melatonin supplementation does poorly: Treating insomnia in people with normal circadian rhythms. The effect on sleep onset is small; the effect on sleep maintenance is essentially nil.
What you should know about commercial melatonin:
The doses sold commercially are dramatically higher than what’s physiologically meaningful. Standard pineal gland production is roughly 0.3 mg over the course of the night. Supplements typically deliver 3-10 mg, and some are sold at 30 mg or higher. The 60-80 mg “for jet lag” recommendation that circulates in some biohacker communities is approximately 200 times the body’s natural production and has no evidence base supporting it. High-dose melatonin can produce paradoxical effects, vivid, disturbing dreams, daytime grogginess, and, at very high doses, hormonal effects (it suppresses gonadotropin-releasing hormone and, at high doses, can suppress reproductive function).
A 2017 Journal of Clinical Sleep Medicine analysis of commercial melatonin products found that actual melatonin content ranged from 17% below to 478% above the labeled amount. Quality control across the supplement industry is genuinely poor.
If you’re going to use melatonin:
Pre-pubertal children should not take melatonin without specialist supervision; the evidence on developmental effects is limited, and there are theoretical concerns about effects on puberty timing. The concern is contested. Many pediatric sleep specialists do prescribe it, but the framework’s caution-when-evidence-is-thin principle applies.
My general position on melatonin: useful in narrow applications, oversold in popular discourse, and the doses commonly recommended are an order of magnitude higher than what makes physiological sense. Never mess with hormone production if you can help it.
Obstructive sleep apnea (OSA) is one of the most consequential and most underdiagnosed conditions in sleep medicine. The full picture is on the Respiratory Health page; the sleep-specific summary here focuses on what’s relevant for the broader sleep system.
OSA is a repetitive collapse of the upper airway during sleep. The airway loses muscle tone, narrows or closes entirely, breathing stops or becomes severely restricted, oxygen saturation drops, the brain partially wakes to restore airway tone, breathing resumes, sleep deepens again, and the cycle repeats. Often dozens of times per hour in moderate-to-severe cases.
The mechanism breaks sleep in two ways. First, the brief arousals fragment sleep architecture, particularly preventing sustained N3 deep sleep. Second, the intermittent oxygen desaturation produces oxidative stress and sympathetic nervous system activation that have cardiovascular and metabolic consequences over time.
Estimates suggest roughly a third of adults with moderate-to-severe OSA are undiagnosed. The rate is substantially higher in groups that don’t fit the stereotypical risk profile. Particularly women (where OSA presentations are often subtler and frequently missed) and lean adults (where the assumption that OSA requires obesity leads clinicians to skip evaluation).
The classical signs, such as loud snoring, witnessed pauses, daytime sleepiness, large neck circumference, and BMI in the obese range, catch obvious cases.
They miss many presentations:
If sleep is unrefreshing despite adequate duration, if you wake with morning headaches or dry mouth, if your partner reports gasping or pauses, if you have hypertension that’s hard to control, or if you have unexplained chronic fatigue, sleep apnea is worth investigating regardless of whether you fit the stereotypical profile.
Formal diagnosis requires polysomnography (overnight in-lab sleep study) or a validated home sleep test. The home tests have become much more accessible in recent years and are appropriate for many patients, though severe cases or atypical presentations still warrant in-lab studies.
Self-screening tools that suggest formal evaluation:
CPAP works. But the industry around it isn’t without problems worth knowing.
The 2021 Philips DreamStation recall is the most consequential recent event. Polyurethane foam used for sound dampening in millions of CPAP and BiPAP machines could degrade and release potentially carcinogenic compounds into the air patients breathed. The recall affected 5+ million devices and is still working through litigation. Internal documents released through discovery suggest Philips was aware of the foam degradation issue for years before the recall.
The SAVE trial, published in NEJM in 2016, raised real questions about whether CPAP reduces cardiovascular events in patients with moderate-to-severe OSA who lack daytime sleepiness. The trial enrolled 2,717 patients and found no reduction in cardiovascular events with CPAP versus usual care. This complicates the “everyone with OSA needs CPAP for cardiovascular protection” framing. For symptomatic patients, CPAP improves quality of life and probably does reduce events, but the case for treating asymptomatic OSA on cardiovascular grounds is weaker than the popular discourse suggests.
These are the conditions where the SCN’s timing is misaligned with the local environment or the demands placed on it. The mechanisms are different from those of insomnia and sleep apnea, and the treatments are different.
The most common circadian disorder. The person’s circadian rhythm is consistently shifted later than the local schedule allows for. Bedtime is naturally 2 AM or later, wake time is naturally 10 AM or later, and forcing earlier hours produces chronic sleep deprivation rather than adapting the underlying rhythm.
DSWPD is heavily genetic and often presents in adolescence. Roenneberg’s chronotype work has documented that what gets called “extreme night owl” behavior in popular discourse has substantial genetic underpinnings rather than being purely a matter of choice or laziness. About 0.2-10% of the population (depending on diagnostic criteria) meets formal DSWPD criteria.
The standard advice of “just go to bed earlier” doesn’t work because the SCN’s timing isn’t easily forced by behavioral effort alone.
For many DSWPD patients, their natural rhythm is shifted in a way that doesn’t fully reset to the standard 9-to-5 schedule even with treatment. Lifestyle adjustments (jobs that allow later start times, partners who tolerate different schedules, recognition that “early to bed, early to rise” isn’t the only valid pattern) sometimes serve people better than treatment that works against their underlying biology.
The opposite: circadian rhythm shifted earlier than desired. Sleep onset at 7 PM, wake time at 3 AM. Less common than DSWPD; more common in older adults as part of normal aging shifts. Treatment uses bright light in the evening (rather than morning) to delay the SCN.
Most common in totally blind individuals (without light perception), where the SCN can’t entrain to the day-night cycle and free-runs at its endogenous period of approximately 24.2 hours. The person’s sleep-wake rhythm drifts later by ~12 minutes per day, cycling through normal alignment and complete misalignment over weeks. Tasimelteon (Hetlioz) is FDA-approved for this condition.
Rarely occurs in sighted individuals, but possible.
Occurs in approximately 10% of shift workers. Chronic misalignment between the body’s circadian rhythm and the schedule the work demands. Long-term shift work has documented health consequences: increased cardiovascular risk, metabolic dysregulation, mental health effects, and the IARC classification of shift work as probably carcinogenic in 2007 was driven partly by these data.
Treatment is partial at best. The underlying mismatch can’t be fully resolved while the work pattern continues.
Strategies include:
For permanent night shift workers, the approach focuses on maximizing what’s possible. For rotating shift workers, the chronic re-adaptation is genuinely hard on the system; consistent shift patterns produce better outcomes than rotating ones.
Acute circadian misalignment from rapid travel across time zones. The full treatment is in the Cheat Sheet; the summary here:
The body’s circadian system can shift by approximately 1 hour per day under ideal conditions. Crossing 8 time zones eastward produces 8+ days of partial misalignment if the system is left to adjust on its own. Strategic light exposure (bright morning light at the destination, avoidance of evening light), timed melatonin, and consistent meal/exercise patterns at the destination can accelerate adaptation modestly but don’t eliminate the underlying issue.
Eastward travel is harder than westward travel because it requires phase advance (which the system does less easily than phase delay). The “extra” hour going from London to Auckland is metabolically harder than the equivalent westward shift.
Excessive daytime sleepiness with sudden sleep attacks, often accompanied by cataplexy (sudden loss of muscle tone triggered by emotion), sleep paralysis, and hypnagogic hallucinations. Caused by the loss of orexin/hypocretin neurons in the hypothalamus, characterized in detail by Emmanuel Mignot at Stanford in the late 1990s. Treatment involves stimulants for daytime alertness and sometimes sodium oxybate for nighttime consolidation. Often diagnosed years after symptom onset because the presentation can be subtle.
Uncomfortable sensations in the legs at rest, with an urge to move that improves the symptoms. Worsens in the evening and at night. Affects sleep onset and maintenance. Often linked to iron deficiency (ferritin levels, even within “normal” range, many RLS patients improve with iron supplementation when ferritin is below 75 ng/mL). Treatment includes iron repletion when appropriate, dopamine agonists, and gabapentin-class medications for severe cases.
Loss of the normal muscle paralysis during REM sleep, causing patients to physically act out their dreams: kicking, punching, sometimes injuring themselves or their bed partners. Important not because it’s common but because it has strong predictive value: RBD precedes the development of Parkinson’s disease, Lewy body dementia, or multiple system atrophy in over 80% of cases, often by years to decades. If you or a partner notice physical movement during dreams (not occasional twitching, but sustained acting out), this warrants neurologic evaluation, both for management of the immediate sleep issue and for the early window into possible neurodegenerative conditions.
Behaviors during sleep: sleepwalking, sleep talking, sleep eating, night terrors. Most common in childhood and is usually outgrown. When occurring in adults, often triggered by sleep deprivation, alcohol, or medications. Treatment usually focuses on safety (preventing injury, ensuring environmental safety) and addressing triggers. Z-drugs in particular are known to occasionally produce complex sleep behaviors with no memory.
The pharmaceutical industry around sleep is substantial. The major sleep medication market generates billions in annual revenue. CBT-I, which works at least as well, generates none of this revenue because it’s a structured therapy rather than a billable medication.
This shapes the field in predictable ways. Medications get more research funding than non-pharmacological treatments. Primary care training emphasizes prescribing because that’s what fits in a 12-minute appointment. Sleep speciality training varies in how heavily it emphasizes CBT-I as first-line. Patient awareness of CBT-I as an option remains low because no major commercial entity has an incentive to promote it.
This isn’t a conspiracy. It’s the predictable result of how the medical system’s incentive structures work. The 2021 Annals of Internal Medicine analysis showed that fewer than 10% of insomnia patients receive CBT-I as first-line treatment despite guidelines recommending it as a documentation of the gap.
The CPAP industry has its own version. The Philips DreamStation recall is the cleanest specific case, but the broader pattern of medical device regulation, incentive alignment between manufacturers and sleep clinics, and the difficulty of long-term independent quality assurance all play out in this space.
The threshold for seeking clinical evaluation:
See a primary care clinician promptly if:
Ask specifically for:
Consider seeing a sleep specialist if:
Behavioral sleep medicine specialists (psychologists trained specifically in sleep) deliver CBT-I and related cognitive behavioral interventions for sleep disorders. The Society of Behavioral Sleep Medicine maintains a directory of qualified providers.
Sleep medicine has gotten substantially better over the last few decades. CBT-I works. Sleep apnea diagnosis is more accessible. Newer medications are more targeted. The system isn’t perfect, but engaging with it is usually better than not engaging when the problem is significant. The framework here is meant to help you be a more informed participant rather than keep you out of the medical system.