Does the Presence of Symbolic Play Matter In Toddlers with ASD?

Background: With the onset of new AAP guidelines for early screening, an earlier diagnosis of autism spectrum disorder (ASD) has become a reality.  Toddler screening instruments rely on play as a “critical item” for evaluating risk status and employ it as the medium for assessment. Yet, the presence of symbolic play is rarely utilized in the final algorithm for diagnostic decision-making. Does the presence of symbolic play in toddlers with social and communication findings consistent with ASD have the potential for a prognostic role in terms of clinical outcome? 

Objectives: To examine symbolic play behavior in young children who received a diagnosis of ASD by age 24 months (mo) and to assess the relationship between symbolic play at the time of original diagnosis and autism severity and cognitive functioning at follow-up.  

Methods: Retrospective chart review of 74 children presenting by age 24 mo to a University Affiliated early intervention program from 2003 to 2009 who received an ASD diagnosis based on multidisciplinary evaluation and who had follow-up at least 1 year later. Data from initial evaluation included: age, demographics, cognition, autistic features based on the DSM-IV criteria, Childhood Autism Rating Scale (CARS) and, in some, the Autism Diagnostic Observation Schedule (ADOS). Initial play level was determined based on clinical observations and compared to the criterion-referenced Westby Play Scale (1992). Children were classified as having symbolic play (SP) if they demonstrated short pretend schemas enacting familiar everyday activities using self or a doll as agent and non-symbolic (No-SP) if they lacked pretend play. ASD diagnosis at follow-up was based on CARS, DSM IV, ADOS, and educational classification. Data of children with SP and No-SP were compared. Statistical analysis included chi-square, t-test and nonparametric testing.

Results: At presentation, mean age of the sample was 19.9 mo; 68% male; 64% had a cognitive standard score > 70; 46% exhibited SP. At follow-up, mean age was 54.1 mo. 87% of the sample continued to meet ASD criteria.   At follow-up, SP children were more likely to show improvement in autism severity: with a greater likelihood of moving from ASD/Autistic Disorder (AD) to PDD.nos (47% vs. 20%, p=0.013) and of having a CARS score <30 (53% vs. 18%, p=0.009) than the No-SP group.   At follow-up, No-SP children had a higher total CARS score (mean CARS 34.9 ±7.3 vs. 29.9± 5.3, p=0.001) and were more likely to meet criteria for AD (68% vs. 17%, p<0.001). They were also more likely to have lower cognition (SS <70) both at presentation (62% vs. 30%, p=0.006) and follow-up (87% vs. 50%, p=0.03) than SP children. 

Conclusions: Symbolic play was commonly seen in a cohort of children presenting early with social communicative deficits and given an ASD diagnosis. Children demonstrating symbolic play at presentation showed greater likelihood of improved clinical outcome with decreased autism severity and higher cognition at follow-up than those who did not. Symbolic play in the presence of social communicative deficits represents a potentially meaningful prognostic indicator of positive outcome for children receiving an early ASD diagnosis.