Date Presented 03/26/20
The 2011 Survey of Pathways to Diagnosis and Services includes 15 sensory items that could inform tailored therapeutic interventions established by OTs. However, these items have not been validated as a measure. Based on psychometric analyses, this study concluded that all 15 items were adequately explained by the general sensory factor and one of four sensory factors: (1) sensory seeking; (2) sensory avoidant; (3) disorganized/overreactive; and (4) repetitive behaviors.
Primary Author and Speaker: Mi Jung Lee
Additional Authors and Speakers: Ickpyo Hong
Contributing Authors: Karen Ratcliff, Claudia Hilton
PURPOSE: The 2011 Survey of Pathways to Diagnosis and Services (Pathways) was conducted on the nationally representative sample of children with autism spectrum disorder, intellectual disability, and developmental delay to examine the developmental problems and caregiving concerns for various symptoms, diagnostic status, and clinical treatments/interventions [1]. Information derived from this national population survey supports researchers and policymakers to identify special health care needs to establish tailored therapeutic interventions. However, these items have not been validated and tested to be used by occupational therapists in clinical settings as a measure. Thus, the purpose of this study was to examine the structural model of the 15 sensory items.
DESIGN: Retrospective data analysis
METHOD: The study cohort includes a total of 6,090 children aged 6 to 17 years who ever had one of the following: autism spectrum disorder, intellectual disability, or developmental delay (4,032 phone interviews and 2,988 mailed surveys). We excluded children who did not complete the Strengths & Difficulties Questionnaire, Children’s Social Behavior Questionnaire, and those who ever had developmental disorders. The 15 sensory items with a 3-point rating scale (1 = does not apply, 2 = sometimes or somewhat applies, and 3 = clearly or often applies) were extracted from Pathways. Exploratory factor analysis (EFA) was conducted to identify the factor structure for the 15 items. Then, different structural models (unidimensional models, unidimensional models with correlations, a hierarchical multidimensional model, and a bifactor model) were tested using multidimensional item response theory to determine the best-fitted measurement model for the items. The model fit criteria for MIRT were; 1) root mean square error of approximation (RMSEA < .08), 2) comparative fit index (CFI > .95), and 3) Tucker-Lewis Index (TLI > .95) [2].
RESULTS: A total of 1,968 children was selected for this study. Exploratory factor analysis (EFA) identified four factors in the measure (eigenvalue 1 = 6.0, eigenvalue 2 = 1.6, eigenvalue 3 = 1.2, and eigenvalue 4 = 1.0). Based on the EFA results and item descriptions, content experts labeled the four factors as: 1) sensory seeking; 2) sensory avoidant; 3) low registration; and 4) repetitive behaviors. Then, MIRT further confirmed that a bi-factor model (RMSEA = .03, CFI = .992, TLI = .988) was the best fit.
CONCLUSION: This study revealed that a bifactor model fits the 15 sensory items from Pathways, suggesting that all 15 items were adequately explained by the general sensory factor and each corresponding sensory factor. This study verified our notion that children’s sensory outcomes are simultaneously influenced by their general sensory and one specific sensory factor. Accordingly, applying a bifactor model is recommended for estimating sensory outcomes for children with autism spectrum disorder, intellectual disability, or developmental delay. Future studies are encouraged to evaluate associations between estimated sensory outcomes with children’s various symptoms, diagnostic status, and clinical treatments/interventions.
IMPACT STATEMENT: By using the most appropriate structural model for 15 sensory items from nationally well-represented samples, estimating sensory seeking, sensory avoidant, low registration, and repetitive behavior scores is feasible to use for occupational therapists in tailoring therapeutic interventions.
References
https://www.cdc.gov/nchs/slaits/spds.htm
Reeve, B. B., Hays, R. D., Bjorner, J. B., Cook, K. F., Crane, P. K., Teresi, J. A., . . . Cella, D. (2007). Psychometric evaluation and calibration of health-related quality of life item banks: plans for the Patient-Reported Outcomes Measurement Information System (PROMIS). Medical Care, 45(5 Suppl 1), S22-31. doi:10.1097/01.mlr.0000250483.85507.04
