Interpreting score differences in the Insomnia Severity Index: using health-related outcomes to define the minimally important difference

Curr Med Res Opin. 2009 Oct;25(10):2487-94. doi: 10.1185/03007990903167415.


Objective: To estimate the minimally important difference (MID) for the Insomnia Severity Index (ISI) by examining the association of score differences of the ISI with health-related outcomes including health-related quality of life, productivity, and fatigue.

Methods: Data came from a randomized, placebo-controlled clinical trial evaluating the long-term efficacy of eszopiclone for primary insomnia (N = 828). Logistic regression models were used to characterize the relationship between ISI change scores (from baseline to 6 months post-treatment) and outcomes/anchors from the SF-36 Health Survey, Work Limitations Questionnaire (WLQ), and Fatigue Severity Scale (FSS). Odds ratios were derived from the regression coefficients to calculate the probability of a given outcome being associated with different ISI change scores. Convergence between anchor- and distribution-based estimates was assessed.

Results: Higher ISI scores (indicating more severe insomnia) were significantly associated with higher probabilities of negative outcome in all models. Individuals with a 6-point score reduction in ISI scores (which corresponded to 1(1/2) standard deviations) were 48% less likely to report 'feeling worn out' (SF-36) at 6 months, 46% less likely to be 'unable to think clearly' (WLQ), and 52% less likely to report 'feeling fatigued' (FSS). Similar results were found across a broad spectrum of all selected anchors.

Conclusions: Based on results of the study, a 6-point reduction is recommended to represent a clinically meaningful improvement in individuals with primary insomnia. Research on generalizability of the recommended MID in this study to other patient populations and other type of treatment interventions is needed.

Publication types

  • Randomized Controlled Trial
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Fatigue
  • Female
  • Humans
  • Logistic Models
  • Male
  • Middle Aged
  • Placebos
  • Severity of Illness Index*
  • Sleep Initiation and Maintenance Disorders / physiopathology*
  • Surveys and Questionnaires


  • Placebos