The Psychometric Properties of the Trunk Impairment Scale in Children with Cerebral Palsy

Abstract

The Trunk Impairment Scale (TIS) measures static and dynamic seated trunk control in children with cerebral palsy (CP) who have postural control problems. Studies have investigated the reliability and validity of the TIS. However, the fitness and difficulty of the scale items have not been investigated. This study used Rasch analysis to test the construct validity of TIS for children with CP. TIS data were collected from 60 children with CP and analyzed for person and item fit, item difficulty, rating scale suitability, and separation reliability. Principal component analyses of residuals revealed that TIS had unidimensionality. Five misfit items (static sitting balance (SSB) items 2 and 3, dynamic sitting balance (DSB) items 4 and 5, and coordination (COO) item 3) were identified. DSB8 is the most difficult item, followed by DSB3 and COO4. On the other hand, the SSB3 item was found to be a relatively easy item. The rating scales demonstrated that out of the three subscales, SSB, DSB, and COO, only the SSB subscale did not meet the appropriate criteria. We demonstrated that statistical item analysis with the Rasch model could provide valuable information related to psychometric properties.

Keywords: Rasch analysis; cerebral palsy; children; psychometric; trunk impairment scale.

Figure 1

Study population.

Figure 2

Map of personal proficiencies and item difficulties for the 17 items of the Trunk Impairment Scale. Note. More difficult items are placed on the top of the diagram. M, mean of child or item distribution; S, 1 standard deviation from the mean; T, 2 standard deviations. Each ‘#’ is two persons and each ‘∙’ is one person.