Development and Construct Validity of a Comprehensive Feeding Assessment for Children with Autism Spectrum Disorders

Background: Feeding challenges are estimated to occur in up to 89% of children with autism spectrum disorders (ASD). Restrictive feeding patterns and preferences may lead to nutrient deficient diets, limited participation in family mealtimes, challenging behaviors, and increased caregiver stress during mealtime. Currently, limited assessment tools exist to assess feeding challenges in children with ASD and ones that do are limited in scope and lack validity and reliability evaluation.

Objectives: The Feeding Assessment for Children with Autism was developed to 1) provide a more complete understanding of feeding disorders of children with ASD and 2) confirm validity of the assessment for use in research and clinical settings. The purpose of this study was to determine the construct validity of the Feeding Assessment for Children with Autism and describe association with child and parent characteristics.

Methods: Exploratory methodological research design was used to determine construct validity of the assessment tool. A large national sample (N=406) of caregivers of children (2 to 12 years in age) and reported feeding challenges was recruited through the Interactive Autism Network. Exclusion criteria included primary hearing, visual, or physical impairments, seizure activity in the last 12 months and other conditions known to co-occur with ASD.

The Feeding Assessment for Children with Autism Spectrum Disorders is a 51-question caregiver-report questionnaire developed to characterize feeding challenges in children with ASD. Questionnaire items are believed to be related to four underlying constructs: sensory (n=13), behavior (n=17), health (n=11), and oral motor (n=10) skills. Confirmatory factor analysis, internal consistency, model fit indices, factor loadings, and between-factor correlation matrix were conducted to determine construct validity.

Results: Primary analysis confirmed a four-factor model was appropriate for the questionnaire items with high internal consistency (0.87). Confirmatory model fit indices indicated moderate fit of questionnaire items into the four constructs. Factor loadings for all four theoretical factors were adequate and statistically significant (p < .001) with all loadings >.22. All between factor correlations were statistically significant (p<.02) with small-medium positive Pearson r correlations. The behavior factor had medium positive correlations with both oral-motor (.45) and sensory factors (.30).

Conclusions: Findings of the study provide preliminary confirmation of the construct validity of the Feeding Assessment for Children with Autism Spectrum Disorders, providing researchers and clinicians with a comprehensive and psychometrically sound assessment tool to address children’s unique feeding challenges. Relatively small correlation magnitudes indicate questionnaire items successfully targeted the four unique feeding challenges. The highest between factor correlations were seen between behavior and all other factors, in particular oral motor and sensory. This finding lends support to the hypothesis that mealtime behaviors observed in children with ASD are in response to other challenges these children face including sensory responsiveness, underlying health concerns, and delayed oral motor skills. This tool will allow for more detailed characterization of feeding challenges with this population, and will help optimize individualized and targeted interventions for children with ASD.