Individual Differences in Developmental Gains across One Year of Early Intervention for Pre-Schoolers with Autism.

Background: Many children with autism benefit from intensive early intervention. However, individual responses vary greatly and little is known about the profile of skills associated with more favourable outcomes. Age at intake and amount of intervention received are commonly identified as contributing to positive outcomes, as are symptom severity and developmental level. However, these factors have rarely been examined together, and among child characteristics, total scores are often used to measure skills, which may mask information about specific behavioural predictors.

Objectives: We aimed to identify specific behaviours associated with developmental gains across one year of intervention, and further examine the role of intervention dose and child age at intake, in a large sample. In this cohort, while we can control for the influence of intervention type, as all children received the same program - participant characteristics and intervention dose naturally vary, enabling us to examine the association of this variation on outcomes.

Methods: Participants were a cohort of 110 children (M age = 2.7 years at intake; range 1 to 4.8 years) who received approximately 1-year of Group-Early Start Denver Model (Vivanti et al., 2017) between 2015 and 2017. The Autism Diagnostic Observation Schedule (ADOS-2) was conducted at intake, and single items representing key foundational skills were examined as potential predictors (e.g., Pointing, Response to Joint Attention, etc.) of outcome. The primary outcome of interest was change in Mullen Scales of Early Learning (MSEL) Verbal and Non-Verbal age equivalent (V/NVAE) scores; administered at intake (T1) and exit (T2). Earlier cohorts have shown improvement on the MSEL through this program (Vivanti et al., 2014). Intervention dose was estimated as days/week*months enrolled.

Results: Large variability was evident in developmental gains on the MSEL (see Figure 1). Controlling for T1 AE scores, correlations were evident between younger age at intake and higher T2 NVAE (r=21, p<.05), and between higher intervention dose and T2 VAE (r=.24, p<.05). Hierarchical regressions examining the unique role of potential predictors, controlling for T1 AE scores, revealed higher T2 NVAE was predicted by ADOS Pointing and intervention dose and (less) ADOS Overactivity (F(1,83)=54.53, p<.001, R²=.72). ADOS Response to Joint Attention was associated with higher T2 VAE (F(1,88)=220.82 p<.001, R²=.83). Whilst correlated with outcomes, age at intake was not a significant unique predictor of T2 V/NVAE, nor were other ADOS social-communication skills (e.g., Play, Eye Contact and Language) when entered alongside the aforementioned significant unique predictors.

Conclusions: Identifying specific skills/behaviours associated with more favourable developmental outcomes can help inform future individualised treatment decisions. Developmental age may play a greater role in predicting gains than chronological age. Nevertheless, these data reinforce the importance of early intervention to ensure learning supports are in place to minimise the extent of the gap between chronological and developmental age over time. While examination of associated factors in a large cohort receiving a common intervention informs our understanding of predictors of outcome, replication of findings within a comparative study is required to disentangle genuine predictors of intervention outcome from broad indicators of good prognosis.