Predictive Value of Social Communicative Abilities in Toddlers Screening Positive for ASD towards Outcome At Age 7-8y

Background: Early detection of autism spectrum disorder (ASD) is important for clinical (e.g., to facilitate early intervention) as well as theoretical (e.g., to improve our knowledge about the developmental pathway of the disorder) reasons. Therefore, routine screening for ASD of all children has been recommended (American Academy of Pediatrics, 2007). In recent years, many screening instruments for ASD in toddlers were developed but they often generate a high false positive rate (e.g., Kleinman et al., 2008). These false positive screen children may have other developmental difficulties that also need thorough assessment and early intervention (e.g., Dietz et al., 2006). However, only a few studies prospectively followed the development of positive screen children into childhood (e.g., Charman et al., 2005) and these studies usually did not include false positive screens.

Objectives: The first aim of this study was to determine if language, intelligence, and theory of mind (ToM) differ in children with and without a clinical diagnosis of ASD at the age of 7-8y, recruited from a group of children with a positive screen for ASD in toddlerhood. Secondly, this prospective study wanted to examine if social communicative abilities in toddlers screening positive for ASD are predictive for outcome at age 7-8y. 

Methods: Seventy-nine children were identified as at-risk for ASD based on a positive screen for ASD and were seen at the university lab between ages 2-4 (Dereu et al., 2012). At the age of 3, the early social-communicative skills imitation, joint attention, and pretend play were assessed. At age 7-8y, all children of the at-risk group were asked to participate in a follow-up study to measure their language, intelligence, and ToM. Preliminary results based on one third of the sample are summarized here. At the meeting, complete data will be presented.

Results: Significant differences were found between children who developed ASD and the group who did not for all variables measured at the age of 7-8y. Children with ASD had poorer language skills, a lower IQ, and scored worse on ToM tasks than children without ASD. Hierarchical linear regression analyses revealed that social communicative abilities at age 3y explained unique variance in outcome measures at age 7-8y on top of the variance explained by diagnostic status (total variances explained were between 61% and 91%). Procedural imitation and initiated behavioural requests explained respectively 13-15% and 33-53% additional variance in language, intelligence, and ToM at 7-8y. In addition, initiated joint attention explained respectively 41% and 36% additional variance in language and intelligence. Finally, pretend play only predicted ToM at age 7-8y (16% additional variance explained).

Conclusions: Social communicative skills measured at 3y can predict outcome at age 7-8y. Especially initiated joint attention and behavioural requests explained additional variance in language, intelligence, and ToM on top of variance explained by diagnosis (true versus false positives). So, thorough assessment of social communicative abilities may also be important in false positive screens and, if impaired, they should be targets of early intervention in at-risk groups.