The Swedish version of the Insomnia Severity Index: Factor structure analysis and psychometric properties in chronic pain patients.

Objective Insomnia is the most commonly diagnosed comorbidity disorder among patients with chronic pain. This circumstance requests brief and valid instruments for screening insomnia in epidemiological studies. The main object of this study was to assess the psychometric properties and factor structure of the Swedish version of the Insomnia Severity Index (ISI). The ISI is a short instrument designed to measure clinical insomnia and one of the most common used scales both in clinical and research practice. However there is no study in Sweden that guarantees neither its factor structure nor its feasibility in chronic pain patients. We further examined the measurement invariance property of the ISI across the two sexes. Methods The ISI was administered to 836 (269 men and 567 women) chronic pain patients from the Swedish Quality Registry for Pain Rehabilitation. This study used demographic data, the Hospital Anxiety and Depression Scale (HADS), the Mental Summary Component (MSC) of the Health Survey (SF-36) and the item 7 from Multidimensional Pain Inventory (MPI). The sample was divided into two random halves: exploratory factor analysis (EFA) was performed in the first sample (N1 = 334, 40%) and confirmatory factor analysis (CFA) in the second half of the sample (N2 = 502, 60%). The measurement and structural invariance of the proposed structure (4-item version) between the two sexes as well as reliability and validity indexes were further assessed. Results Exploratory factor analysis using the principal axis factoring method generated one global factor structure for the ISI, explaining 63.1% of the total variance. The one factor solution was stable between the two sexes. Principal component analysis was also applied and indicated almost identical results. The structure was further assessed by CFA, resulting in an adequate fit only after omitting three items. The difference on structural and measurement invariance in the loadings by participants' sex was not significant (Δχ2 = 10.6; df = 3; p = .69 and Δχ2 = 2.86; df = 3; p = 41 respectively). The shorter version four-item Insomnia Severity Index (ISI-4) was analysed further. The Chronbach's alpha for the global ISI-4 score was 0.88. The construct validity of the ISI-4 was also supported by the, Hospital Anxiety and Depression Scale, the Mental Summary Component of quality of life and quality of sleep data. Pain intensity was significantly associated with the ISI-4 score (beta = .29, p < 001) whereas no significant correlation between four-item Insomnia Severity Index score and age was observed (p > 05). Conclusions and implications Although short, the four-item Insomnia Severity Index (ISI-4) version seemed to effectively assess insomnia in chronic pain patients. An important clinical implication is that the four-item Swedish Insomnia Severity Index can be used in chronic pain cohorts when screening for insomnia problems. Its measurement and structural invariance property across the two sexes shows that the ISI-4 is a valid measure of the insomnia across groups of chronic patients. Our results also suggest its utility both in pain clinical practice and research purposes.