Sleep quality as a potential target in the behavioral network of children and adolescents with multimorbidity: a machine learning and computer‑simulated network intervention study

With the transformation of lifestyles, the phenomenon of multimorbidity among children and adolescents has become increasingly severe. Therefore, identifying key behaviors and formulating precise, prioritized intervention strategies are vital for improving the situation of multimorbidity among children and adolescents. This study aims to explore this key issue from a multi-methodological perspective. A total of 10,182 children and adolescents were included. The random forest model in machine learning was utilized, and mean decrease in Gini were used to measure the importance of behaviors; a behavioral network for children and adolescents with comorbidities was constructed. Furthermore, computer-simulated network intervention was used to identify potential, simulation-identified intervention and prevention targets for behaviors in children and adolescents with multimorbidity. The random forest model indicated that sleep quality was the most important behavioral factor. Network analysis indicated, the Sleep quality–MVPA edge differed significantly between groups and had a higher weight in the multimorbidity group. Computer-simulated intervention results indicated that dietary diversity may be a potential intervention target for mitigating behavioral networks in children and adolescents with multimorbidity; sleep quality may be a potential preventive target. The integrated approach based on machine learning and computer simulation of network interventions offers novel insights into behavioral interventions for children and adolescents with comorbidities. Overall, sleep quality serves as a potential core target for improving multimorbidity among children and adolescents.