An AIoT-Based Intelligent Neonatal Jaundice Monitoring System with Real-Time Data Validation: ClinicalStatistical Analysis of Efficacy and Safety

Abstract

Neonatal jaundice, a common condition requiring prompt and effective treatment, often faces challenges such as inconsistent therapeutic outcomes and safety risks in traditional phototherapy. This study presents an AIoT-based intelligent neonatal jaundice monitoring system designed to enhance treatment efficacy and safety through real-time data validation and dynamic parameter adjustments. The system integrates artificial intelligence (AI) algorithms, including Convolutional Neural Networks (CNNs) and Multilayer Perceptrons (MLPs), with IoT-enabled sensors for continuous monitoring of physiological parameters and environmental conditions (e.g., light intensity, incubator temperature). Clinical trials involving 200 neonates demonstrated significant reductions in serum bilirubin levels (from 15 mg/dL to 5 mg/dL within 24 hours) and treatment duration (18 hours for double-sided phototherapy vs. 24 hours for single-sided methods). Statistical analysis revealed a 90% cure rate in the system group, compared to 78% in the control group (P=0.021), alongside significantly lower incidences of adverse events. The system’s ability to dynamically optimize light distribution and intensity based on real-time data ensures uniform therapy delivery while prioritizing patient comfort and safety. These results underscore the potential of AIoT-driven systems to revolutionize neonatal jaundice management by combining precision, adaptability, and clinical reliability.