Insomnia Severity Index (ISI)

What ISI measures

The Insomnia Severity Index is a 7-item self-report measure of insomnia severity over the past 2 weeks. It captures both nocturnal symptoms (difficulty falling asleep, staying asleep, early waking) and daytime impact (impairment, distress, satisfaction). Each item scores 0–4. Total range: 0–28.

The 2-week window matters: insomnia by definition is a persistent problem. A few bad nights after a stressful event do not constitute insomnia disorder.

Interpretation table (Morin 2011)

The 8–14 band is meaningful: many people in this range have early or atypical insomnia that responds well to behavioural intervention before it becomes chronic. Sleep hygiene plus sleep-restriction principles for 2–4 weeks is reasonable as a self-managed first step.

Why CBT-I, not pills

AASM 2021 clinical practice guidelines recommend Cognitive Behavioural Therapy for Insomnia (CBT-I) as the first-line treatment for chronic insomnia disorder. The evidence is strong:

• Effect sizes of 0.7–1.0 for sleep onset, total sleep time, and ISI score reduction
• Benefits maintained at 6 and 12 months post-treatment
• No rebound, no tolerance, no risk of dependence
• Comparable benefits via face-to-face therapy or digital CBT-I (Sleepio, Somryst)

Sleep medication has a more limited role: short-term use during acute stress, or as adjunct when CBT-I is unavailable. Long-term hypnotic use is associated with tolerance, rebound insomnia on discontinuation, and accumulating evidence of cognitive and falls risk in older adults.

Sleep hygiene basics — necessary but rarely sufficient

If your score is in the subthreshold band (8–14), trying these for 2–4 weeks is reasonable before seeking specialist care:

• Consistent bed and wake times, even on weekends (within ±30 min)
• Bed only for sleep and intimacy — no laptop, phone, or TV in bed
• Dark, cool bedroom (16–19°C is the optimum range in most studies)
• No caffeine after early afternoon — see our caffeine half-life calculator for personalized timing
• Morning daylight within 30 minutes of waking — strongest signal regulating circadian rhythm
• Stop forcing sleep: if you’ve been awake >20 minutes, get up, do something quiet in dim light, return when sleepy

If the score does not improve after 4 weeks of consistent hygiene, or you scored ≥15, hygiene alone is unlikely to be enough — proceed to CBT-I.

What’s not insomnia

A high ISI score may actually reflect:

• Obstructive sleep apnoea — loud snoring, witnessed pauses, daytime sleepiness despite adequate time in bed. Needs polysomnography or home sleep test.
• Restless legs syndrome — uncomfortable urge to move legs, worse in evening.
• Circadian rhythm disorders — sleep is fine when timing is unrestricted; problems arise from work/school schedule mismatch.
• Depression or anxiety — early-morning waking is classic for depression; difficulty falling asleep with racing thoughts is classic for anxiety. Treating the mood disorder often resolves the sleep problem.
• Medication contributors — beta-blockers, decongestants, certain antidepressants, corticosteroids.

A clinical evaluation can disentangle these, often within one consultation.

Related calculators

• Sleep cycle calculator — best bedtimes and wake times by 90-minute cycles.
• Caffeine half-life calculator — personalised cut-off for evening alertness.
• PHQ-9 depression test — depression and insomnia are bidirectionally linked.
• GAD-7 anxiety test — anxiety frequently underlies insomnia.

Limitations

• Recent acute stressors. A bereavement, relationship rupture, or major life event will raise the score in ways that may resolve naturally. Repeat after 4 weeks.
• Shift workers have sleep fragmentation that is environmental, not insomnia per se. Specialist sleep-medicine input is preferable.
• Adolescents — ISI is validated in adults. Modified versions exist for youth.
• Older adults typically score slightly higher than younger adults at any given clinical level; thresholds are still useful but interpret in context.

Privacy

All calculations run in your browser. We never see, log, or store your individual answers. Only an anonymous event (locale and severity band) is sent to a privacy-respecting analytics service.