I've had chronic insomnia for about 3 ye
Ask
I've had chronic insomnia for about 3 years. I take 45–90 minutes to fall asleep most nights, wake up 2–3 times, and get maybe 5 hours of actual sleep despite being in bed 8–9 hours. I'm exhausted all day. I've tried sleep hygiene (dark room, no screens, cool temp, consistent bedtime) for over a year with minimal improvement.
My doctor finally referred me to a sleep specialist who recommends Cognitive Behavioral Therapy for Insomnia (CBT-I) instead of sleep medication. She said it's the "gold standard" but I'm confused about what CBT-I actually involves in practice.
Specifically:
1. What are the core components of CBT-I — I've heard "sleep restriction" is involved and it sounds terrifying. Does it mean I'll be sleeping even less before I sleep more?
2. Sleep restriction therapy: how does it actually work mechanically? If I'm currently getting 5 hours, does it restrict me to 5 hours, and then I work up from there?
3. How long does a full CBT-I program take before most people see improvement? I've read 6–8 weeks — is that accurate, and what does progress look like?
4. Are there digital/app-based CBT-I programs (like Sleepio or similar) that have evidence behind them, or do I need to work with a therapist in person?
Answer
I'll pull the CBT-I card and the chronic-insomnia fork card first, then check the current evidence on digital programs.
Yes — CBT-I is the right call. The 2021 AASM clinical practice guideline makes multicomponent CBT-I the strong recommendation for chronic insomnia in adults, and sleep hygiene alone (which you've already given a real year of) is no longer considered sufficient for chronic insomnia disorder. Your specialist isn't steering you wrong.
Your pattern — 45–90 min to fall asleep, 2–3 middle wakes, 5 hours of actual sleep in 8–9 in bed, 3+ years, near-nightly — is a textbook **onset + maintenance** mix, with sleep extension (early bedtime, lying in) likely the perpetuating factor. CBT-I is built to target exactly that.
## The six components
| Component | What it does |
| --- | --- |
| Sleep hygiene education | emoves obvious sabotage — caffeine, alcohol, room, timing. The layer you've already worked. |
| Stimulus control | e-pairs bed/bedroom with sleep. Bed only for sleep and sex; get up after ~20 min awake; one fixed wake time. |
| **Sleep restriction** | Clinician compresses time in bed to match actual sleep, then expands as efficiency improves. |
| Cognitive therapy | Catches catastrophising ("tomorrow's ruined"), clock-watching, sleep-as-performance. |
| elaxation / counter-arousal | Breathing, imagery, calming audio — quiets the hyperarousal side. |
| elapse prevention | After a bad week: keep the wake time, restart get-up-if-awake, no catch-up sleep, no restarting pills. |
Source: Edinger et al., *J Clin Sleep Med* 2021, doi:10.5664/jcsm.8986 (AASM 2021 guideline); NHS insomnia pages.
## About sleep restriction (the scary part)
You're right that it sounds backwards, and yes — the first 1–3 weeks often feel **worse** before they feel better. Mechanically:
1. **Build sleep pressure.** Limiting time in bed lets sleep drive accumulate faster, so falling asleep gets easier and middle wakes shorten.
2. **Break the "bed = frustration" pairing.** Lying awake 3 hours a night reinforces the bad association. A tighter, more successful window reverses it.
3. **The clinician sets the window, anchored to a fixed wake time** — typically *close to, not below*, your current average sleep time. The specialist sets it from a 1–2 week sleep diary and titrates in ~15-min expansions when sleep efficiency hits ~85–90%. I don't compute this for you; the diary drives it.
4. **The first 2–3 weeks have a real daytime-sleepiness peak.** Kyle et al. 2014 found Epworth scores rose at weeks 1, 2, and 3, with impaired vigilance and roughly a third of participants reporting driving was affected. **Sleepiness returns toward baseline by ~3 months** — the early dip is not the method failing.
Source: Kyle et al., *Sleep* 2014, doi:10.5665/sleep.3386; Sleepio suitability guidance; AASM 2021.
**Who must not DIY restriction:** bipolar or psychotic symptoms, untreated sleep apnea, narcolepsy, epilepsy/seizures, pregnancy without clinician oversight, or safety-critical work (driving, machinery, traffic control) during the acute phase. aise any of these with your specialist at session 1.
## Timeline
6–8 weeks is the standard course (4–8 weekly sessions is typical).
| Phase | What it usually looks like |
| --- | --- |
| Week 1 | Diary begins; window set; first nights feel rough; daytime sleepiness rises |
| Weeks 2–3 | Often the hardest — restricted TIB, low sleep, daytime fog. Expected, not failure. |
| Weeks 3–5 | Sleep starts consolidating; falling asleep faster; middle wakes shorten |
| Weeks 6–8 | Window expands in 15-min steps as efficiency improves; most see meaningful gains |
| ~3 months | Gains typically consolidate; long-term outcomes beat pills |
Source: AASM 2021; NHS insomnia pages; dCBT-I trial literature.
## Digital vs in-person
The evidence base for digital CBT-I is strong and growing. Two names with the most data:
- **Sleepio** — multiple CTs; a 2025 trial showed sustained gains at 6 months. Listed by NICE (UK, 2022) as a cost-saving option in primary care; available through some NHS GP patches and in the US through employers/insurers.
- **Somryst** (formerly SHUTi) — FDA-cleared dCBT-I in the US. Pivotal studies showed >40% remission. **Public availability has been unstable since 2023 — verify before relying on it.**
Other doors: **CBT-I Coach** (US VA, free, best paired with a clinician); **somnio** and **HelloBetter Schlafen** (Germany DiGA, prescription, statutory-insurer paid); **Sleepstation** (some NHS GP practices); **internetpsykiatri.se** (Sweden).
Source: NICE HTG624; AASM 2024 dCBT-I platforms review; JMI 2025 Sleepiox trial; DiGA directory; Somryst pivotal studies.
**Which one fits depends on where you are and what your insurer or health system covers.** I don't know your country, so I won't point you at the wrong door:
- **UK** — ask your GP whether Sleepio or Sleepstation is on your local NHS pathway.
- **Germany** — somnio or HelloBetter on prescription (DiGA).
- **US** — CBT-I Coach is free; Sleepio/Somryst access depends on employer/insurer.
- **Australia** — healthdirect routes to a GP/sleep specialist; no public automated programme listed.
- **Most other countries** — no publicly listed self-serve programme. In-person or guided telehealth with a CBT-I-trained clinician is then the route.
That's a commissioning fact, not a quality verdict on the apps.
## What to ask your specialist at session 1
- "Do you run in-person CBT-I, or refer to a guided digital programme?"
- "Will sleep restriction be part of my plan, and how will you titrate it?"
- "Is there a publicly funded or insurer-covered programme available here?"
- "Can I start a sleep diary now so week 1 isn't a blank?" (Most clinicians want 1–2 weeks of diary before session 1.)
## Honest limits
I'm not your clinician. I can't titrate your sleep window, run CBT-I sessions, or decide whether restriction is safe for you. The specialist designs and adjusts the plan from your diary. If you're already falling asleep unintentionally during the day, or driving while drowsy, raise that at session 1 — the early weeks need a safety plan before you start.