Active Engagement in Toddlers Referred for Possible Autism Spectrum Disorder

Background: The National Research Council (2001) recommended 25 hours of active engagement (AE) per week of early intervention for children with autism spectrum disorders (ASD). However, there has been a marked absence of research addressing AE for children with ASD. Additionally, there is pressing need for better treatment outcome measures in ASD in light of the priority of early intervention.

Objectives: The purpose of this investigation from the prospective, longitudinal study of the FIRST WORDS Project was to quantify AE in toddlers referred for possible ASD using a new measure of AE.

Methods: Video observations were recorded in the homes of 35 children (17-35 months, m=22.5) recruited from the FIRST WORDS Project with red flags of ASD on the CSBS Behavior Sample. Twenty seven received a diagnosis of ASD. Observations were parent-child interactions during 3-5 activities for 40 minutes. Six dimensions of child AE were coded (regulation, productivity, social orientation, eye gaze to face, response to bid for interaction, and communicative act) using 30 second intervals to determine the proportion of AE time. In addition, six dimensions of parent strategies that support AE were measured (promoting participation, routine features, opportunities for communication, balance of turns, synchrony/modeling, and expectations). Measures of nonverbal and verbal development were obtained from the Mullen Scales of Early Learning (17-38 months, m=25.2). Adaptive behavior was obtained from the Vineland Adaptive Behavior Scales (17-37 months, m=25.7). Measures of autism symptoms were obtained using the ADOS (18-36 months, m=26.5).

Results: Young children referred for possible ASD displayed varying levels of AE with regulation occurring at a high proportion of intervals and social orientation and response occurring at low proportions. Significant correlations were observed between regulation and productivity, social orientation and eye gaze, and eye gaze and communicative act. Time samples of varying lengths were not significantly different from one another and were significantly correlated. Forty minute cumulative time samples of response, eye gaze, and communicative act showed significant correlations with outcome measures and added unique variance to predictive models. Although no consistent pattern emerged suggesting that any 10 minute time sample predicted better than the others, 40 minute cumulative samples were the best consistent predictors. Finally, significant correlations were observed between child AE measures and parent support measures.

Conclusions: Measures of AE suggest that regulation and productivity occurred at high proportions in this natural home environment. Social orientation and response to bid occurred at very low proportions suggesting that these may be sensitive measures to detect growth over time. Forty minute intervals were the best consistent predictors of outcome measures suggesting that observations of AE should be a least 40 minutes in length. These findings have important implications for operationalizing AE as a treatment outcome measure of the density of the learning environment for children with ASD.