Psychometric evaluation of the Chinese revised Sensory Integration and Praxis Tests in children with amblyopia

Children with amblyopia commonly present with sensory integration difficulties, however, there is a paucity of validated tools to assess these issues in this clinical population. This study aimed to evaluate the psychometric properties of the Chinese Revised version of the Sensory Integration and Praxis Tests (SIPT-R) for use in children with amblyopia.

We conducted a psychometric validation study using baseline data from the China Amblyopia Behavioral Cohort (CABC). Participants included 629 children aged 4-9 years diagnosed with amblyopia; subtypes included bilateral ametropic (55.7%), anisometropic (38.8%), and strabismic amblyopia (5.6%). Based on severity, 32.1% were classified as mild amblyopia, 59.3% as moderate amblyopia, and 8.6% as severe. Because the learning ability subscale was only applicable to children aged 6 years and older, scale structure differed between age groups. Thus, reliability and validity were evaluated separately for the aged 4-5 and 6-9 years. Internal consistency and construct validity were examined via Cronbach's α, split-half reliability, and confirmatory factor analysis (CFA). Convergent and discriminant validity were assessed via factor loadings, composite reliability (CR), and average variance extracted (AVE).

The SIPT-R demonstrated excellent internal consistency (Cronbach's α = 0.968 - 0.969) and strong subscale reliability (α = 0.906 - 0.959). All split-half coefficients exceeded 0.80. CFA indicated an acceptable model fit for both age groups: for aged 4-5 years, χ ²/df = 2.712, root mean square error of approximation (RMSEA) = 0.079, comparative fit index (CFI) = 0.832, incremental fit index (IFI) = 0.833; for aged 6-9 years, χ ²/df = 2.951, RMSEA = 0.074, CFI = 0.804, IFI = 0.805. Most items demonstrated adequate factor loadings, with CR values ranging from 0.91 to 0.96 and AVEs between 0.41 and 0.68. Discriminant validity was largely satisfactory; however, a high correlation was observed between tactile defensiveness and proprioception (r = 0.86), which exceeded the square root of their respective AVE (0.69), suggesting some overlap between these constructs.

The Chinese revised SIPT-R demonstrates strong reliability and acceptable validity for screening sensory integration function in children with amblyopia. Further research should focus on refining specific subscales and evaluating additional psychometric properties, such as test-retest reliability and criterion validity.