CBT-I is the first-line insomnia treatment and nobody offers it

The treatment with the strongest evidence for chronic insomnia isn't a pill. It's a behavioral program called CBT-I, and the American College of Physicians recommended it as the first-line treatment for adults with chronic insomnia back in 2016. The American Academy of Sleep Medicine says the same thing. Almost nobody gets offered it, because a twelve-minute appointment ends faster with a prescription than with a four-week protocol.

So you'll likely have to run it yourself. Good news: the core of it is three rules and some arithmetic, and it costs nothing.

Bad news: week one is worse than what you have now.

First, make sure it's actually insomnia

Insomnia means you can't fall asleep, can't stay asleep, or wake too early, and it wrecks your next day. The key word is opportunity. You had seven hours available and you got four.

If your two-year-old is up at 2:40 every morning and again at 5, you don't have insomnia. You have a toddler. That's sleep deprivation, and the fix is logistics and a spouse who takes alternate nights, not a therapy protocol. Running sleep restriction on a parent who's already short on sleep is a bad idea.

Also worth ruling out before you start: snoring loud enough to be heard through a door, a spouse who's seen you stop breathing, waking with a dry mouth and a headache, blood pressure that won't come down. That cluster points at sleep apnea, and no behavioral program fixes an airway that closes. Crawling, aching legs at night that improve when you move them point somewhere else again. Both of those are doctor conversations and possibly a sleep study.

And if you're currently taking a prescription sleep medication, don't stop it on your own because an article told you the behavioral route works better. Some of them, particularly the older benzodiazepine class, need a supervised taper. Ask the person who prescribed it.

Week one is a notebook

For seven nights, write down four things when you wake up. What time you got into bed. Roughly how long it took to fall asleep. How long you were awake in the middle of the night, in total, your best guess. What time you got out of bed for good.

No stopwatch. No smartwatch score. The wearable's "deep sleep" number is going to make you anxious about sleep, which is the exact mechanism you're trying to dismantle. Estimates are fine. That's what the protocol was built on.

Then you do the math. Add up your estimated sleep for the week, divide by seven, and you've got your average total sleep time. Divide that by your average time in bed and you've got sleep efficiency as a percentage.

A worked example. In bed at 11:00, out at 6:30, so 7.5 hours of opportunity. Forty minutes to fall asleep most nights, plus an hour staring at the ceiling around 3. Call it five hours of sleep. Five divided by 7.5 is 67 percent.

Healthy sleepers run somewhere north of 85. You're spending two and a half hours a night lying in the dark, awake, teaching your brain that the bed is a place where you think about work email.

The part that works, and hurts

Sleep restriction. You cut your time in bed down to roughly the sleep you're actually getting, plus about thirty minutes.

In the example above, that's five and a half hours. Pick your wake time first, because it's the anchor and it doesn't move, not on Saturday, not after a rough night. Say 6:00 a.m. Then your earliest bedtime is 12:30 a.m.

Most protocols don't go below about five to five and a half hours of time in bed, whatever your diary says. Don't get clever and prescribe yourself four.

What happens next is that you get sleepy. Genuinely, heavily sleepy, in a way you may not have felt in years. Sleep pressure is a real physiological thing and you've been bleeding it off by lying down for eight hours to catch five. Compress the window and the pressure builds, and the pressure is what knocks you out in nine minutes instead of fifty.

Days three through six are the tax. You'll feel foggy, short-tempered, and convinced this is the stupidest thing you've ever tried. Don't drive long distances that week. If your job involves a ladder, a forklift, a scalpel or a squad car, talk to a clinician before you start rather than winging it. People with bipolar disorder or a seizure disorder shouldn't run sleep restriction without supervision at all, because short sleep can trigger episodes.

Then you earn time back

At the end of each week, redo the arithmetic. If your sleep efficiency came in above 85 or 90 percent, move your bedtime fifteen minutes earlier. That's it. Fifteen.

If it's below 80, take fifteen away. If you're between, hold steady another week.

You keep titrating until you land on the amount of sleep your body actually wants, at high efficiency, with a bed that means sleep again. For most people that process runs four to six weeks. Some land at 7 hours 15 minutes and are annoyed it isn't eight. Eight is a marketing number. What you want is a night you sleep through and a day you can think in.

The bed is for sleeping

The second rule is stimulus control, and it's where most people quietly cheat.

Bed is for sleep. Not for scrolling, not for the game, not for a laptop, not for the long-form argument with your wife about her sister. If you're in bed and awake and irritated about it, get up. The usual guidance is roughly twenty minutes, but don't look at the clock to measure it, because clock-watching is its own form of arousal. When it starts to feel bad, that's your twenty minutes.

Go somewhere else. Keep the lights low. Read something on paper that you don't care much about. Fold laundry. Come back when you're sleepy, not when you're bored. If you're up again forty minutes later, do it again.

This feels absurd the first three nights and then it starts to work, because you're breaking a conditioned association. Your brain currently thinks the bedroom is where the worrying happens. It learned that from you, over hundreds of nights, and it can unlearn it in a couple of weeks.

The thinking part

Lying awake at 3 a.m. doing damage projections for the following day is the engine of chronic insomnia. "I'll be useless in the meeting." "I'll snap at the kids again."

Test it against your own record. You've had bad nights before and gone on to do a normal day's work. Not a great day. A normal one. Humans are irritatingly functional on mediocre sleep, which is the one piece of good news the insomnia brain refuses to accept.

Practical version: keep a pad on the kitchen counter, not the nightstand. Anything that shows up in your head at 2 a.m. gets written there when you get out of bed. Tomorrow's problem, tomorrow's paper.

And caffeine has a half-life of around five to six hours, so the 3 p.m. cold brew is still meaningfully in you at 9. Move the cutoff to noon during the protocol. Alcohol is worse than it feels. It gets you down fast and then fragments the back half of the night, which is exactly the half you're trying to repair.

If you'd rather not run it alone

There are structured digital CBT-I programs, including Somryst, which is FDA-cleared as a prescription treatment for chronic insomnia. There are also behavioral sleep medicine providers who do this for a living, and the Society of Behavioral Sleep Medicine keeps a directory. Six sessions is typical. A lot of insurance covers it once somebody names it.

Ask for it by name. "I'd like a referral for CBT-I" gets a different response than "I'm not sleeping well," which gets you a sample packet and a follow-up in three months.

None of this is a diagnosis, and a stubborn sleep problem deserves an actual clinician looking at you.

The strangest part is what you notice around week five. Not that you sleep more. That you stop thinking about it.