The Role of Early Intervention on Communication Development in Children with Autism Spectrum Disorder: A Systematic Review of Randomized Controlled Trials

Abstract

Introduction: Impairment of social communication is a defining feature of autism spectrum disorder (ASD), and the pre-school years constitute a period of maximal neurodevelopmental plasticity. Although early intervention is universally recommended, the magnitude and generalizability of its effect on communication development remain contested. This systematic review was undertaken to quantify the effect of early intervention on communication development in children with ASD using exclusively randomized evidence. Methods: The study strictly adhered to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) 2020 guidelines. Eligible studies were randomized controlled trials enrolling children with ASD (mean age ≤72 months) comparing early intervention against control conditions and reporting at least one communication outcome. Risk of bias was assessed using Cochrane RoB 2, and certainty of evidence was graded using GRADE. A structured narrative synthesis with effect-direction analysis was performed due to substantial heterogeneity. Results: Twenty randomized controlled trials comprising 1,713 children were included. Consistent and statistically significant benefit was demonstrated for proximal dyadic communication outcomes: caregiver synchronous responsiveness improved significantly in 4 of 6 trials (effect sizes 0.33 to 1.61), and child-initiated communication improved significantly in 6 of 8 trials (effect sizes 0.20 to 0.99). Initiation of joint attention improved in 4 of 5 trials, and joint engagement in all 4 trials measuring it. Expressive language improved in only 6 of 13 trials, receptive language in 4 of 8, and blinded autism symptom severity in 5 of 10, with the two largest contemporary pragmatic trials entirely null on their primary blinded endpoints (effect sizes 0.04 and −0.06). Effects attenuated monotonically as measurement moved from proximal dyadic outcomes to distal standardized instruments. Discussion: The evidence supports a coherent, mechanistically explicable but hierarchically attenuating effect: intervention reliably changes targeted caregiver behaviour and child communicative initiation, with effect thereafter attenuating progressively as measurement moves away from the intervention context toward standardized, decontextualized instruments. Parent-mediated interventions delivered by non-specialist health workers reproduced proximal effects obtained by specialist therapists. Conclusion: Early intervention produces robust, replicated improvement in proximal communication architecture—caregiver responsiveness, child communicative initiation, joint attention, and joint engagement—with moderate certainty. Effects on standardized language and blinded symptom severity are smaller and less consistent. Parent-mediated, naturalistic developmental behavioural interventions should be offered as first-line management, with outcome monitoring anchored on dyadic communication measures.