Meta-Analysis on Technology-Based Versus Non-Technology Based Social Communication Interventions for Children with ASD

Background:

Educators and clinicians now have access to a growing number of technologies that they can incorporate into their interventions with children with ASD. Captivating sound and visual effects, as well as consistency in target demonstrations are amongst the reasons that support the use of technology in intervention (Moore & Calvert, 2000; Bernard-Opitz, Sriram, Nakhoda-Sapuan, 2001). Meta-analysis studies that explored the benefits of technology-based social-communication interventions yielded mixed findings. One meta-analysis (Grynszpan, Weiss, Perez-Diaz, & Gal, 2014) demonstrated a positive effect with the use of technology in therapy, whereas another meta-analysis (Wang & Parrila, 2011) found no evidence to support the use of technology. Existing meta-analyses either concentrated on studies that compared children receiving technology-based therapy to a waitlist/typically developing group, or on single-subject intervention studies where a comparison group was lacking. A meta-analysis conducted on studies that included children with ASD in both a technology-based intervention group and a non-technology intervention control group would improve clinical decision-making regarding therapy delivery method.

Objectives:

To compare the effectiveness of technology-based versus non-technology based social communication interventions for children with ASD using a meta-analysis approach.

Methods:

A systematic literature search was performed using keywords including “autism,” “ASD,” “technology,” “computer,” “robot,” “video,” and “virtual.” A total of 5386 studies were identified. 24 studies met the inclusion criteria: a) included at least one participant with diagnosis of ASD; b) employed randomized control trial or alternating treatment designs; c) contain at least one dependent measure of social communication (verbalizations, affect, imitation, joint attention, turn-taking, requesting, pointing, or social interaction). 12 studies did not report Ms or SDs and thus were excluded from further analysis. A total of 12 studies containing 89 participants with ASD were included in the analysis. Across all studies, we extracted 37 dependent measures that were related to social communication. For each dependent measure, we calculated the standardised mean difference (SMD) between the technology-based intervention and non-technology based intervention groups. Then, a Z-test determined whether the SMDs across studies were reliably different from zero.

Results:

Overall, interventions conducted through technology did not result in better social communication outcomes compared to those conducted without technology (Cohen’s d = 0.14; standard error =0.68, p=0.17).

Conclusions:

Based on studies with the most stringent methodological designs, our meta-analysis found that children with ASD did not benefit more from technology-based intervention compared to non-technology based intervention. Studies (e.g. Bernard-Opitz et al., 2001) have suggested that technology-based interventions can result in more long-term improvements of skills in children with ASD. Future studies on this topic should consider both immediate intervention outcomes and the generalizability and sustainability of newly learnt skills.