Comparison of two nutritional risk screening tools in hospitalized children with Japanese encephalitis: a causal inference of clinical outcomes and implications for optimized management

Abstract

Abstract Background This study used two nutritional risk screening (NRS) tools to explore the causal relationship between nutritional risk and clinical outcomes (length of hospital stay and cost), as well as clinical results (incidence of sequelae), in hospitalized children with Japanese encephalitis (JE). The goal is to screen for a more suitable nutrition risk tool for JE reveal the underlying mechanisms, accurately quantify the impact, and provide a reliable basis for optimizing clinical management and reducing the burden of the disease in affected children. Methods The classical Screening Tool for Risk of Nutrition in Growth Kids (STRONGkids) and Screening Tool for Assessment of Malnutrition in Pediatrics (STAMP) were utilized to evaluate the nutritional risk of the children. A heatmap analysis was conducted to investigate the correlation between variables influencing the STRONGkids score and STAMP score. Subsequently, a decision tree was employed to identify the main factors influencing the STRONGkids score and STAMP score. Finally, causal inference was employed to calculate the causal effects between the NRS score, clinical outcomes, and clinical results. Results Dysphagia was the most significant factors affecting STRONGKids scores, and the weight and height was the most significant factors affecting STAMP scores. Causal analysis revealed that for every unit increase in the severity of JE type, the STRONGkids score increased by 0.515 units, and 1.339 units for STAMP. Moreover, the presence of dysphagia led to a 1.944-unit increase in the STRONGkids score, and 1.497-unit for STAMP. Additionally, for every unit increase in the STRONGkids score, the length of hospital stay increased by 2.541 days, and hospitalization costs increased by $612.507. Similarly, for every unit increase in the STAMP score, the length of hospital stay increased by 1.571 days, and hospitalization costs increased by $425.595. Conclusions Based on decision tree, causal analysis and the actual situation of SNI, the internal structural setup of the STAMP tool is more suitable for screening pediatric patients with JE, making it a more reasonable choice for this purpose when compared to STRONGkids.

Keywords

• Children
• Japanese encephalitis
• Nutritional risk
• Clinical outcomes
• Causal analysis
• Decision tree (machine learning)