INSOMNIA SEVERITY INDEX
The Insomnia Severity Index (ISI) is designed to be both a brief screening measure of insomnia and an outcomes measure for use in treatment research (1,2).
Scale content corresponds in part to DSM-IV criteria for insomnia, and measures the subject's current (within the past 2 weeks) perception of symptom severity, distress, and daytime impairment. Items include: the severity of sleep onset and maintenance (middle and early morning awakening) difficulties, satisfaction with current sleep pattern, interference with daily functioning, appearance of impairment attributed to the sleep problem, and the degree of concern caused by insomnia.
Charles M. Morin, PhD. E-mail: .
Primary version is self-administered; clinician-administered, significant other, and French language versions are available.
Number of items in scale.
There are 7 items.
The psychometric properties of ISI were reported from a sample of adults (ages 17–84 years) with insomnia complaints (primary and secondary to medical, psychiatric, or other sleep disorders) (2).
The ISI has also been validated on samples of young adults (mean age 20 years) (3), and older adults (mean age 65 years) with primary insomnia (2).
WHO ICF Components.
Body function, Impairment, Activity limitation, Participation restriction, Environmental factor.
Self-administration; paper and pencil.
Time to administer/complete.
Five minutes or less.
No cost; available from authors. Copy available at the Arthritis Care & Research Web site at http://www.interscience.wiley.com/jpages/0004-3591:1/suppmat/index.html.
Five-point Likert scale (0 = not at all, 4 = extremely).
Range is 0–28.
Interpretation of scores.
Suggested guidelines for interpretation: 0–7 = no clinically significant insomnia; 8–14 = subthreshold insomnia; 15–21 = clinical insomnia of moderate severity; 21–28 = severe clinical insomnia. Guidelines require additional validation. Smith and Trinder (3) found a cutoff score of 14 distinguished subjects with insomnia from normal controls with a sensitivity = 94% and specificity = 94%.
Method of scoring.
Individual items are summed by hand.
Time to score.
Less than one minute.
Training to score.
Training to interpret.
Traditional norms are not available. Mean (SD) for insomnia patients (n = 145) is 19.7 (4.1). Mean (SD) older adults with DSM-IV diagnosis of primary insomnia is 15.4 (4.2). ISI scores for the 5 insomnia subgroups in the original sample were pain conditions (20.2); psychophysiological (19.5); psychiatric (21.0); idiopathic (19.7); alcohol/substance abuse (19.8); and other (19.5) (2).
Cronbach's alpha = 0.74 to 0.78. Individual item correlations to the total score ranged from 0.36 (difficulty falling asleep) to 0.67 (interference with daily functioning), with a mean of 0.54. Internal reliability coefficients of individual items to total score demonstrated increased stability following treatment with means of 0.69 at post-treatment and 0.72 at followup (2).
The ISI has good face validity with the concept of insomnia as defined by DSM-IV. Formal evaluation of content validity was demonstrated via principal component analysis, which yielded 3 components (impact, severity, and satisfaction). These components are consistent with the diagnostic criteria of insomnia and captured 72% of the variance (2).
A cutoff score of 14 demonstrated a sensitivity of 94% and a specificity of 94% in distinguishing individuals diagnosed with primary insomnia from good sleeper controls. Diagnoses were established based on integration of expert clinical interview, polysomnography, and psychometric testing (3).
Bastien and colleagues reported significant correlations of select ISI items and Total Score with relevant polysomnography (PSG) variables before and after treatment (2). The ISI Sleep Onset item was the only significant pretreatment correlation (r = 0.45; P < 0.05, PSG Sleep Latency). At posttreatment, ISI Total Score correlated with PSG Sleep Efficiency (−0.35; P < 0.05). ISI Sleep Onset and middle insomnia items correlated with PSG Sleep Latency and Wake After Sleep Onset (0.39 and 0.45; P < 0.05, respectively).
In regards to sleep diatries, Bastien and colleagues (2) reported significant correlations of select ISI items with corresponding daily sleep diary parameters in 2 separate samples, one of which included correlations between measures both before and after treatment. Pretreatment correlations were weak to moderate (0.32–0.55) and posttreatment correlations were moderate to strong (0.55–0.99). The ISI Total Score correlated weakly to moderately with diary measures of Sleep Efficiency (Pearson's r values ranging between −0.19 and −0.61).
Correlation coefficients between clinician ISI ratings and the patient version for ISI total score ranged between 0.57 and 0.71, P < 0.01. Individual item correlations between the 2 versions ranged between 0.50 (difficulty staying asleep) and 0.69 (problem waking up too early).
Clinician ratings of insomnia severity predicted ISI Total Score. R2 ranged from 0.37 (pretreatment) to 0.61 (post-treatment) P < 0.05) (2).
Sensitivity/responsiveness to change.
Morin and colleagues demonstrated sensitivity to change following pharmacologic and/or behavioral intervention for primary insomnia in older adults (4).
Comments and Critique
The ISI is a face valid index of insomnia severity demonstrating criterion validity and other adequate psychometric properties. It has been validated against both polysomnographic and prospective sleep diary measures and demonstrates convergence with clinical interview criteria. It may be particularly useful for treatment outcome research in which insomnia is a secondary endpoint, and a brief, low subject burden instrument is needed. It may also be useful for diagnosis and treatment planning. The cutoff score may be useful as a guideline for clinicians in evaluating the clinical significance of the insomnia complaint. While not specifically developed for rheumatology patients, scale development included a heterogenous group of patients with insomnia secondary to pain conditions.
Due to its brevity and Likert scale data, this instrument would not be an appropriate stand-alone measure of sleep disturbance. The scale is limited to questions pertaining to insomnia severity/impact and does not assess frequency of symptoms. It also does not include items relevant to other sleep disorders, which may occur more frequently in chronic pain populations (e.g., periodic limb movements, restless legs syndrome, or sleep apnea). More research demonstrating the psychometric properties of the ISI, and establishing cutoff scores and sensitivity to change in rheumatologic populations is needed.
1. (Original) Morin CM. Insomnia: psychological assessment and management. New York: Guilford Press; 1993.
2. Bastien CH, Vallie'res A, Morin CM. Validation of the Insomnia Severity Index as an outcome measure for insomnia research. Sleep Med 2000;2:297–307.
3. Smith S, Trinder J. Detecting insomnia: comparison of four self-report measures of sleep in a young adult population. J Sleep Res 2001;10:229–35.
4. Morin CM, Colecchi C, Stone J, Sood R, Brink D. Behavioral and pharmacological therapies for late-life insomnia: a randomized controlled trial. JAMA 1999;281:991–9.