The Dutch Central Sensitization Inventory (CSI) : factor analysis, discriminative power, and test-retest reliability
Faculty of Medicine and Health Sciences
New York, N.Y.
The clinical journal of pain. - New York, N.Y.
, p. 624-630
University of Antwerp
Objectives: A standardized assessment of central sensitization can be performed with the Central Sensitization Inventory (CSI), an English questionnaire consisting of 25 items relating to current health symptoms. The aim of this study was to translate the CSI into Dutch, to perform a factor analysis to reveal the underlying structure, examine its discriminative power, and test-retest reliability. Methods: The CSI was first translated into Dutch. A factor analysis was conducted on CSI data of a large group of chronic pain patients (n=368). The ability to discriminate between chronic pain patients (n=188) and pain-free controls (n=49) was determined and the test-retest reliability for chronic pain patients (n=36) and controls (n=45) with a time interval of 3 weeks was evaluated. Results: The exploratory factor analysis resulted in a 4-factor model based on 20 items, representing the domains General disability and physical symptoms (Cronbach α=0.80), Higher central sensitivity(Cronbach α=0.78), Urological and dermatological symptoms(Cronbach α=0.60), and Emotional distress(Cronbach α=0.80). Furthermore, a parsimonious second-order factor model was found, where the factor General central sensitization was underlying the 4 first-order factors. Chronic pain patients scored significantly worse on all 4 factors. The test-retest reliability was excellent values in both chronic pain patients (ICC=0.88) and controls (ICC=0.91). Discussion: The original CSI was translated into Dutch and did not reveal any problems during data acquisition. The domains represented by the 4 factors may be useful in setting up specific patient profiles and treatment targets. To conclude, the Dutch CSI revealed 4 distinguishable domains, showed good internal consistency for the total score and 3 out of 4 domains, good discriminative power, and excellent test-retest reliability.