Statistical Analysis of Clinical Interventions for Emotional Disorders in Children with Autism: A Longitudinal Study
Download as PDF
DOI: 10.25236/icetmr.2025.019
Corresponding Author
Shike Dong
Abstract
This study aims to systematically evaluate the clinical efficacy of behavioral interventions, pharmacological interventions, and combined interventions by longitudinally tracking changes in emotional disorders among children with autism. It also explores individual differences and key influencing factors. Compared to cross-sectional studies, longitudinal research better reveals causal relationships, captures individual dynamic changes, and controls for potential confounding factors. Based on the Autism Intervention Database, the study plans to enroll approximately 400 children (accounting for a 20% attrition rate) with follow-up at baseline, 6 months, 12 months, and 24 months. Demographic characteristics, intervention measures, emotional symptom scales, and covariate data will be collected. Linear mixed-effects models will analyze overall symptom change trends, while growth mixture models will identify potential subgroups of emotional trajectories. Missing data will be imputed using multilevel chained equations, with sensitivity analyses validating robustness. The combined intervention is projected to yield significantly greater improvement than either monotherapy over 24 months, with statistical significance emerging at 12 months (p<0.05). The medication group showed rapid improvement in the first 12 months but subsequently plateaued, while the behavioral group demonstrated slower yet more sustained improvement. Trajectory analysis projected three subgroups: fast responders (approximately 30%), slow responders (approximately 50%), and non-responders (approximately 20%). These differences were closely associated with factors such as age, baseline cognitive level, and ADHD comorbidity. This study employs longitudinal statistical modeling to reveal the dynamic processes and heterogeneous characteristics of emotional disorder interventions in children with autism, providing evidence-based support for developing individualized clinical intervention strategies. Future research may expand sample sources, refine variable collection, and explore more complex statistical methods to enhance interpretability and clinical translation value.
Keywords
Children with autism; Emotional disorders; Clinical intervention; Longitudinal study; Linear mixed-effects model; Growth mixture models; Statistical analysis; Individual differences