Insomnia is a learned condition. That's also why it's treatable.
Chronic insomnia stops being about whatever started it and becomes a self-sustaining pattern. CBT-I has stronger long-term evidence than any sleep medication.
What insomnia actually is
Chronic insomnia is defined by difficulty falling asleep, staying asleep, or waking too early at least three nights a week for at least three months, despite adequate opportunity for sleep, with daytime consequences. The 'despite adequate opportunity' part is what separates insomnia from insufficient sleep.
Most chronic insomnia is not the same as the original trigger. Something started it — a stressor, an illness, a schedule change, a baby, a heartbreak. Then the trigger resolved, but the insomnia persisted. What kept it going is a learned association: bed-as-place-of-anxiety, plus daytime behaviors (catching up on weekends, going to bed earlier 'to be safe,' lying awake hoping) that sustain the cycle. This is sometimes called the 3P model — predisposing, precipitating, perpetuating factors.
Why pills are the wrong long-term answer
Sleep medications (z-drugs like zolpidem, benzodiazepines, sedating antihistamines, trazodone) reliably help sleep onset in the short term. Long-term, almost all of them lose effectiveness, build tolerance, fragment sleep architecture, and create rebound insomnia on discontinuation. Major sleep medicine guidelines (AASM, NICE) explicitly recommend cognitive behavioral therapy for insomnia (CBT-I) as first-line for chronic insomnia, ahead of medications.
CBT-I has shown durable benefit at 1, 2, and even 5 years post-treatment in trials. No medication does this.
What CBT-I actually does
CBT-I has roughly five components, often delivered in 4–8 sessions: stimulus control (use the bed only for sleep; if not asleep in 15–20 min, get out), sleep restriction (compressing time in bed to match actual sleep, then expanding gradually), cognitive restructuring (working with catastrophic thoughts about sleep loss), sleep hygiene (the well-known piece, but the smallest contributor), and relaxation training.
Sleep restriction is the part that scares people. It is also the most powerful component. The premise: spend slightly less time in bed than you currently sleep, build sleep pressure, then expand the window once consolidation returns. It feels worse for a few nights and works within a few weeks.
what people get wrong
wrongI need 8 hours of sleep or I'll be ruined the next day.
closerIndividual sleep need varies (7–9 hours for most adults, with real outliers). Rigid '8 hours' beliefs increase performance anxiety about sleep and worsen insomnia. Daytime function matters more than the number.
wrongIf I can't sleep, I should stay in bed and try harder.
closerLying in bed awake teaches the brain that bed is for being awake. Stimulus control says: get up after ~20 min, do something dull in dim light, return when sleepy.
wrongNaps will help me catch up.
closerFor people with insomnia, daytime naps reduce night-time sleep pressure and prolong the disorder. Most CBT-I protocols restrict napping during the active phase.
wrongMelatonin works for insomnia.
closerMelatonin has good evidence for circadian-rhythm disorders (jet lag, delayed sleep phase) and weak evidence for chronic insomnia. It is not a sleeping pill in any clinical sense.
what actually helps
- CBT-I — full course with a trained therapist, or via validated digital programs (Sleepio, Somryst) when in-person is unavailable.
- Stimulus control: bed for sleep only; if awake >20 min, get up.
- Sleep restriction: compress time-in-bed to actual sleep duration, expand by 15 min weekly as efficiency improves.
- Consistent wake time, even on weekends — the strongest single circadian anchor.
- Light exposure within 30 min of waking; reduce evening light in the 2 hours before bed.
- Cutting caffeine after noon and alcohol within 3–4 hours of bed (alcohol fragments sleep even when it eases onset).
- Treating coexisting anxiety, depression, sleep apnea, restless legs — common drivers of 'insomnia' that need their own treatment.