Revolutionizing pediatric obesity intervention strategies: From traditional growth reference tools to AI-enabled pediatric obesity clinical decision support systems

Abstract

Background: Childhood obesity is a growing public health concern that can seriously impact children's physical development and long-term health if left unaddressed. Effective monitoring and early intervention are crucial for managing this risk. Objective: This study aims to evaluate the performance and clinical relevance of the Pediatric Obesity and Weight Management (PedOWM) tool, a semantic rule-based Clinical Decision Support System (CDSS) designed to assess and manage childhood and adolescent obesity using national anthropometric data. Method: PedOWM was tested using retrospective auxological data from 100 Turkish children, predominantly composed of overweight and obese individuals, but also including cases from other BMI categories such as underweight and normal weight. The system analyzed anthropometric measurements and generated treatment recommendations based on its semantic rules. Its results(e.g., BMI, Z-scores, percentiles) were compared with three widely used pediatric growth reference tools: & Ccedil;EDD-& Ccedil;oz & uuml;m (Turkey), BCM (Baylor College of Medicine, USA), and CDC (Centers for Disease Control and Prevention, USA). The evaluation included visual plot comparisons, expert assessments by a pediatric endocrinologist, statistical analyses, and a performance comparison between PedOWM's recommendations and those of the clinical expert. Results: PedOWM's results were largely consistent with those produced by the three tools, particularly & Ccedil;EDD & Ccedil;oz & uuml;m, supporting its reliability and validity in clinical evaluations. Statistically significant discrepancies observed with CDC and BCM were primarily attributed to differences in reference populations and the absence of data for children under two years of age. Furthermore, PedOWM's treatment recommendations showed strong concordance with expert clinical decisions, achieving performance metrics of 99.5 % accuracy, 97.1 % precision, 97.5 % recall, and 97.6 % F1-score. Conclusion: Increasing societal awareness of the risks associated with childhood obesity can drive proactive measures, leading to timely interventions that significantly enhance health outcomes for children. The results obtained demonstrate promising findings regarding the applicability, effectiveness, and efficiency of PedOWM, which facilitates collaboration among healthcare professionals, parents, patients, and dietitians.