Longitudinal Change in the Use of Services in Autism Spectrum Disorder: Understanding the Role of Child Characteristics, Family Demographics, and Parent Cognitions

Background: Despite emerging consensus about many features of effective intervention programs (NRC, 2001; Odom et al., 2010), little information is currently available about the nature and intensity of services children with ASD typically receive in their local communities. 

Objectives: The aim of this study was to examine child characteristics (i.e., gender, age, nonverbal cognitive and language abilities, ASD symptoms), family demographics (i.e., family composition, ethnicity/race, income, parental age, parental education, socioeconomic status), and parent cognitions (i.e., parenting stress, parenting sense of competence, parental concepts of development), that may affect access to early intervention, special education, and related services.  

Methods: The sample included 70 children with ASD (chronological age: M = 57 months; SD = 12) who had limited expressive language abilities (expressive language age: M = 16 months; SD = 9). At baseline, children participated in comprehensive assessments to confirm ASD diagnosis (i.e., ADOS-G, ADI-R) and evaluate development (i.e., Mullen Scales of Early Learning). All parents were enrolled in a short education program, providing each parent with basic information and resources on advocating for a young child with ASD (Parent Advocacy Coaching, PAC). Over a period of four months, parents participated in four educational sessions, held in the families’ home (one session per month, 90 minutes per session). Longitudinal change in children’s intervention programs in the community was evaluated over a period of about 27 months, starting 12 months prior to enrollment in PAC. Data for this study were collected in the context of a clinical trial, evaluating a parent-mediated intervention supporting parent-child communication (Siller et al., 2012). 

Results: Data were analyzed by fitting a series of multilevel models for longitudinal data (SAS Proc Mixed; Singer et al., 2003). Results revealed large individual differences in the intensity of children’s individual and school based services. Despite this variability, only two child characteristics (age, gender) emerged as independent predictors. In contrast, the intensity of children’s intervention program was independently predicted by a broad range of demographic characteristics, including parental education, child ethnicity and race, and family composition. Finally, even after child characteristics and family demographics were statistically controlled, results revealed associations between specific parental cognitions (parenting efficacy, understanding of child development) and the subsequent rate of change in the intensity of children’s intervention programs. 

Conclusions: It is important to emphasize that this is not a study of the outcome of a given form of treatment (i.e., PAC). For such a study, one would need an appropriate control group of families who were not receiving the treatment in question. This being said, by ensuring that a minimum level of advocacy-support is available to all research participants, this research design is ideally positioned to identify factors that predict which families will be able to utilize available supports.