Home / Resources / Psychological Measures / ISI
Insomnia Severity Index:
Scoring and Interpretation
The Insomnia Severity Index (ISI) is a seven-item self-report measure of the perceived severity of insomnia symptoms and their daytime consequences, scored from 0 to 28 (Bastien et al., 2001). Before any band is applied, the 0 to 28 sum must be right for Insomnia Severity Index scoring, and the four severity bands and the screening cut-off of 10 must then be kept apart; the post continues with code for SPSS and R, the 2011 psychometric indicators and a methods sentence you can reuse.
What the ISI Measures
The ISI captures how severe a person judges their insomnia to be, including difficulty falling or staying asleep and the degree of dissatisfaction, interference with daily life, visibility to others and worry that follows. Bastien, Vallières and Morin (2001) validated it as a brief outcome measure for insomnia research.
Sleep-medicine trials, behavioral sleep interventions and population surveys use it, usually where severity and distress matter more than the full structure of sleep, which is the territory of the broader PSQI. Later work examined how well it detects insomnia cases and how sensitive it is to treatment change (Morin et al., 2011). Self-report severity is not a diagnosis, and the descriptive bands and the screening cut-off are two different things that are easily confused.
Structure of the ISI
- Items: seven self-report items, each scored from 0 to 4.
- Subscales: none. The standard output is one summed total.
- Reverse-scored items: none in the standard patient version.
- Total: one score from 0 to 28, with higher values indicating greater perceived insomnia severity.
- Reference period: the standard form refers to the last two weeks, and versions with a past-month frame exist, so document which frame was used.
- Forms: patient, clinician and significant-other versions are distinct and should not be pooled.
The instrument itself is available from the American Thoracic Society sleep-measures directory and the original validation publications.
How to Score the ISI
- Check that isi1 to isi7 hold valid, already-keyed responses from 0 to 4 for the version you administered.
- Sum all seven items to form isi_total, which ranges from 0 to 28.
- Decide on a missing-item rule before analysis. The cited psychometric studies set no developer-sanctioned prorating threshold, so the code below uses complete-item scoring. State any other policy in the methods section.
* Illustrative syntax: variables must already be valid item scores (0 to 4).
COMPUTE isi_total = SUM.7(isi1, isi2, isi3, isi4, isi5, isi6, isi7).
EXECUTE.# data: one row per respondent; items isi1 to isi7 coded 0 to 4.
cols <- paste0("isi", 1:7)
data$isi_total <- rowSums(data[cols], na.rm = FALSE)Interpreting ISI Scores
| Score band | Interpretation |
|---|---|
| 0 to 7 | No clinically significant insomnia |
| 8 to 14 | Subthreshold insomnia |
| 15 to 21 | Clinical insomnia, moderate severity |
| 22 to 28 | Clinical insomnia, severe |
Source: Morin et al. (2011); National Sleep Research Resource ISI score documentation.
Higher totals indicate more insomnia symptoms and greater associated distress. Morin et al. (2011) found that a score of 10 or more best identified self-reported insomnia in a community sample (sensitivity 86.1%, specificity 87.7%), whereas 11 was the optimal value in a clinical sample. A screening cut-off therefore differs from the four descriptive severity bands, and the two should not be conflated.
The ISI is a patient-reported index. It does not diagnose an insomnia disorder or separate all physiological causes of poor sleep, so read results with the sleep history and the exact questionnaire version in view.
Psychometric Properties
| Property | Published finding | Source |
|---|---|---|
| Internal consistency | α = .90 in a community sample and α = .91 in a clinical sample | Morin et al. (2011) |
| Convergent validity | Associations with fatigue, sleep quality, anxiety, depression and quality of life | Morin et al. (2011) |
| Screening accuracy | At 10 or more in the community sample: sensitivity 86.1% and specificity 87.7% | Morin et al. (2011) |
| Responsiveness | A reduction of about 8.4 points was associated with independently rated moderate improvement in the clinical sample | Morin et al. (2011) |
| Prior validation | Established use as a clinical screening and treatment-outcome measure | Bastien et al. (2001) |
These are source-specific results and cannot be assumed to hold as reliability or accuracy estimates in a new sample.
Using the ISI in a Study
Analyze the total as a continuous patient-reported outcome. In a randomized comparison, a baseline-adjusted model such as ANCOVA estimates the group difference at follow-up with a confidence interval, and with several assessments Linear Mixed Models (Multilevel Modeling) use all available data under stated missingness assumptions. The severity labels can supplement the primary estimate but should not replace it. Where a trial protocol defines insomnia caseness, report the rule that was prespecified.
If the study sets or evaluates a screening threshold against a reference diagnosis, ROC Curve Analysis shows how sensitivity and specificity trade off across cut-offs, and the optimal value will differ by setting. The .90 and .91 values belong to the 2011 samples, so compute alpha for your own data with Reliability Analysis (Cronbach’s Alpha). The 8.4-point change reflects one clinical anchor and sample, not a universal minimal important difference, so name the anchor and setting if you adopt a change threshold.
Related field: Clinical Trials
Reporting the ISI
Insomnia severity was assessed with the seven-item Insomnia Severity Index (ISI; Bastien et al., 2001). Items were coded from 0 to 4 and summed (range = 0 to 28), with higher scores indicating more severe perceived insomnia. The ISI was analyzed as a [continuous/severity-band] variable, and reliability in the present sample was [α = XX].
References
American Thoracic Society. (n.d.). Insomnia Severity Index (ISI). https://site.thoracic.org/assemblies/srn/sleep-related-questionnaires/isi
Bastien, C. H., Vallières, A., & Morin, C. M. (2001). Validation of the Insomnia Severity Index as an outcome measure for insomnia research. Sleep Medicine, 2(4), 297-307. https://doi.org/10.1016/S1389-9457(00)00065-4
Morin, C. M., Belleville, G., Bélanger, L., & Ivers, H. (2011). The Insomnia Severity Index: Psychometric indicators to detect insomnia cases and evaluate treatment response. Sleep, 34(5), 601-608. https://doi.org/10.1093/sleep/34.5.601
Frequently Asked Questions
What is the highest possible ISI score?
The maximum is 28, because each of the seven items contributes between 0 and 4 points.
Does an ISI score of 10 mean moderate insomnia?
No. A score of 10 sits in the conventional subthreshold band of 8 to 14, even though 10 or more was the best screening cut-off in one community sample. The band and the cut-off answer different questions.
Is an eight-point drop in the ISI clinically important?
Morin et al. linked a reduction of about 8.4 points to moderate improvement in one clinical sample. Whether that transfers to your population is a judgment to document, not an assumption.
Are the ISI and the PSQI interchangeable?
No. The ISI focuses on perceived insomnia severity and its consequences, while the PSQI combines seven components of subjective sleep quality, so their scores and ranges are not comparable.
NEXT STEP
Need ISI Totals
Scored and Checked?
OnlineSPSS is a statistical consultancy for researchers. Share your design and dataset, and a statistically trained Research Coordinator reviews it. The scope and one fixed project fee are confirmed before work begins.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014