The Behavioural Pediatrics Feeding Assessment Scale in Young Children with Autism Spectrum Disorder: Psychometrics and Associations with Child and Parent Variables

Saturday, May 16, 2015: 3:16 PM
Grand Ballroom A (Grand America Hotel)
S. L. Allen1, I. M. Smith2, E. K. Duku3, T. Vaillancourt4, P. Szatmari5, S. E. Bryson6, E. J. Fombonne7, J. Volden8, C. Waddell9, L. Zwaigenbaum10, W. Roberts11, P. Mirenda12, T. A. Bennett13, M. Elsabbagh14 and S. Georgiades15, (1)Clinical Psychology, Dalhousie University, Halifax, NS, Canada, (2)Dalhousie University / IWK Health Centre, Halifax, NS, Canada, (3)Offord Centre for Child Studies & McMaster University, Hamilton, ON, Canada, (4)University of Ottawa, Ottawa, ON, Canada, (5)University of Toronto, Toronto, ON, Canada, (6)Autism Research Centre, Dalhousie/IWK Health Centre, Halifax, NS, Canada, (7)Oregon Health & Science University, Portland, OR, (8)Communication Sciences and Disorders, University of Alberta, Edmonton, AB, Canada, (9)Faculty of Health Sciences, Simon Fraser University, Vancouver, BC, Canada, (10)University of Alberta, Edmonton, AB, Canada, (11)Pediatrics, University of Toronto, Toronto, ON, Canada, (12)University of British Columbia, Vancouver, BC, Canada, (13)Psychiatry and Behavioural Neurosciences, Offord Centre for Child Studies, McMaster University, Hamilton, ON, Canada, (14)Department of Psychiatry, McGill University, Montreal, QC, Canada, (15)McMaster University, Hamilton, ON, Canada
Background: Feeding problems negatively affect the health and wellbeing of children with ASD and their caregivers (Sharp et al., 2013), and reportedly occur at very high rates in this population (e.g., 46-89%; Ledford & Gast, 2006). Addressing feeding problems in children with ASD should therefore be a priority; however, research and clinical work have been limited by the lack of adequate standardized measures for quantifying feeding problems in children with ASD. 

Objectives: The purpose of this study was to examine the utility of a widely used measure of feeding problems, the Behavioural Pediatric Feeding Assessment Scale (BPFAS; Crist & Napier-Phillips, 2001), when applied to a large, well-characterized sample of preschoolers with ASD. Our primary goal was to identify an appropriate factor structure. A secondary goal was to evaluate the convergent and divergent validity of the BPFAS by examining correlations between feeding problems and child and parent variables of interest including indices of cognitive functioning, adaptive functioning, autism symptoms and severity, behaviour problems, sleep problems, and parenting stress. 

Methods: Participants were 374 preschoolers with ASD (314 boys; mean age = 40.89 months) and their families. Participants were recruited within 4 months of a diagnosis of ASD, through five publicly funded regional pediatric health care facilities. We began by conducting a confirmatory factor analysis (CFA) of the original BPFAS five-factor model (Crist & Napier-Phillips, 2001), each of the five proposed latent variables, and a rival one-factor model. None of the models was adequate; thus we conducted a categorical exploratory factor analysis (CEFA) to determine an alternative factor structure. Finally, we used Spearman’s rank order correlations to examine the validity of the BPFAS.  

Results: The CEFA identified an acceptable three-factor model (root mean square error of approximation = 0.06), accounting for 43.13% of the cumulative variance (see Table 1 for factor loadings). We labeled Factor 1 as Food Acceptance (Cronbach’s α = .71), Factor 2 as Medical/Oral Motor (Cronbach’s α = .71), and Factor 3 as Mealtime Behaviour (Cronbach’s α = .81). Correlational analyses (see Table 2) indicated that feeding problems were positively related to parent-reported autism symptoms, behaviour problems, sleep problems, and parenting stress, but largely unrelated to language and cognitive abilities, or autism severity as indexed by the ADOS severity metric.

Conclusions: These results provide preliminary support for the BPFAS as a measure of feeding problems in preschoolers with ASD. Further research is needed to establish the measurement invariance of the BPFAS across different samples of children with ASD (e.g., children with clinically significant feeding problems, differing age groups, etc.) and across time. In turn, the BPFAS may prove useful as a screening tool for feeding problems in children with ASD, quantifying treatment progress, and advancing research that seeks to identify treatment targets and variables that contribute to the development and maintenance of feeding problems in children with ASD.