Describing the Heterogeneity of Parent-Child Dyads Including Minimally Verbal Children with Autism

Background: Approximately 30% of children with autism will remain minimally verbal (less than 20 spontaneous words) by school entry (Anderson et al., 2007). This population is highly heterogenous in both cognitive and language abilities. Further, parent-child interaction appears highly varied however, dyads with minimally verbal children with ASD (mvcASD) have yet to be described. For dyads with mvcASD, supporting parents’ ability to foster language building blocks including joint engagement may be essential due to the severity of children’s communication delays. Learning about the ways parents engage their child with complex communication needs prior to intervention may inform the content and delivery of intervention. 

Objectives: To determine reliable subgroups of parent-child dyads including mvcASD. 

Methods: Participants. Participants included 59 children with autism and their primary caregivers, enrolled through a multisite intervention trial. Included children were: 1) 5- 8 years old with developmental age of at least 24 months, 2) diagnosed with autism, 3) used less than 20 words, and 4) had at least 2 years of early intervention.

Measures and outcomes. Multiple measures were taken at study entry including standardized measures of cognition (Leiter International Performance Scale: Roid & Miller, 1997) and language (Peabody Picture Vocabulary Test: Dunn & Dunn, 1995; Test of Early Language Development: Hresko et al., 1981). Dyads were also observed during a 10-minute play interaction coded for caregivers’ responsivity and directiveness (Mahoney & Perales, 2003), interaction strategies (Patterson, Goods, & Kasari, in progress), children’s state of engagement from unengaged through joint engagement (Adamson et al., 2009) as well as children’s initiation of joint engagement. 

Results: K-means cluster analysis was conducted, identifying four unique clusters. Assessments of children’s cognition and language and parental responsivity were not significantly different amongst clusters. Across clusters, significant differences were found in children’s total time in each type of engagement: unengaged (p<.03), object engaged (p<.01), and joint engagement (p<.01) as well as parents’ directiveness (p<.01) and appropriate use of basic interaction strategies (p<.01).

Four clusters were identified. Cluster one (n=4) included dyads with the greatest total time in child initiated joint engagement, least time unengaged with parents demonstrating greatest use of basic skills and moderate directiveness compared to other clusters. In contrast, children in cluster 2 (n=8) spent the least time jointly engaged, most time unengaged with parents remaining passive in the interaction. Cluster 3 (n=32) included children who were primarily object engaged, achieving limited time jointly engaged with moderately directive parents. Last, cluster 4 (n=15) described highly directive parents who spent a moderate amount of time jointly engaged with few to no child initiation of joint engagement. 

Conclusions: Clusters were not defined by children’s cognitive and language abilities, rather, children’s state of engagement and their initiation of joint engagement significantly defined the clusters. Further parents’ strategies and style of interaction also defined the clusters. However, parental responsivity was uniformly low across clusters. As such, parents’ interaction style during play and the dyad’s ability to enter into and maintain a joint engaged state may be key considerations for communication interventions targeting this population.