Psychometric properties of the Insomnia Severity Index (ISI) in Mexican adults

Abstract

Background: Insomnia is the sleep disorder with the highest incidence worldwide. It is estimated that this condition increases the risk of developing psychiatric, neurological, and cardiovascular problems. Due to this, it is important to have brief, reliable and valid psychometric instruments that allow health personnel their timely detection in first level health centers. Objectives: Analyze the psychometric properties of the Insomnia Severity Index (ISI) in its version adapted to Spanish in a sample of Mexican adults. Methods: The sample consisted of 310 adults, 223 (71.9%) women and 87 (28.1%) men. The comparison of four ISI measurement models of one, two and three factors and a bifactor model was carried out, its internal consistency was analyzed, an analysis of invariance by sex and correlation analysis with the Athens and Epworth scales. Results: The ISI bifactor model with a general factor (G) and a specific factor for insomnia impact (I) showed the best fit indices (χ2= 29.48, gl = 11, SRMR= 0.03, CFI= 0.98, TLI= 0.96, RMSEA= 0.07), and demonstrated configural, metric, and scalar invariance by sex. Adequate reliability was demonstrated by Omega coefficients (General: ωG= 0.86, Impact: ωI= 0.86) showed adequate reliability; the scale showed very strong correlations with the Athens scale (rAGoverall= 0.84; rAImpact=0.75) and weak to moderate correlations with the Epworth scale (rEGeneral= 0.39; rEImpact=0.44). Conclusions: The ISI bifactor version in Spanish presents adequate psychometric properties for the measurement of insomnia and, as it is a brief tool, it can be used at different levels of health care.