The Insomnia Severity Index might be the most effective screening tool at identifying insomnia among outpatients with mental disorders, according to a study published in Sleep Medicine.
The cross-sectional study compared six self-administered sleep measures – the Pittsburgh Sleep Quality Index, Insomnia Severity Index (ISI), Epworth Sleepiness Scale, Flinders Fatigue Scale, Functional Outcomes of Sleep Questionnaire, and Dysfunctional Beliefs and Attitudes about Sleep scale – in 400 psychiatric outpatients.
Of those, the Insomnia Severity Index was the most accurate way to discriminate between cases of insomnia and noncases according to both the DSM-5 and ICD-10 criteria. In fact, the Insomnia Severity Index was the only scale that was able to discriminate both with good accuracy.
The area under the curve for the ISI was 0.88 for the ICD definition, and 0.82 for the DSM-5 criteria. Researchers found that the best sensitivity and specificity for the ISI was achieved using cutoff scores of less than or equal to 14 for ICD-10 insomnia and less than or equal to 11 for DSM-5 insomnia.
A cutoff of 14 or above for the ISI yielded a sensitivity of 81.3%, specificity of 80.9%, positive predictive value of 66.7%, and negative predictive value of 90.2%.
The Pittsburgh Sleep Quality Index was found to have good accuracy in discriminating between cases and noncases using the ICD-10 criteria, but only had fair accuracy for the DSM-5 criteria. However, it was slightly better than the ISI at detecting insomnia cases, according to the DSM-5 criteria, in people with either bipolar affective or anxiety disorders.
The Flinders Fatigue Scale, Functional Outcomes of Sleep Questionnaire, and Dysfunctional Beliefs and Attitudes about Sleep scale all showed fair accuracy for the ICD-10 criteria but low accuracy for the DSM-5 criteria, while the Epworth Sleepiness Scale had low accuracy for the ICD-10 criteria and was nondiscriminatory for the DSM-5 criteria.