Related Experiment Video
Updated: Jun 12, 2026
Quantified Assessment of Infant's Gross Motor Abilities Using a Multisensor Wearable
Published on: May 17, 2024
Development and Validation of Non-Invasive Machine-Learning Screening Models for Pediatric Malnutrition in
Petra Klanjšek1, Petra Povalej Bržan2,3, Nataša Marčun Varda2,4
1Faculty of Health Sciences, University of Maribor, Žitna ulica 15, 2000 Maribor, Slovenia.
Insights
New machine learning models effectively screen for child malnutrition using non-invasive data. Simpler models are best for bedside use, aiding early detection of pediatric malnutrition risk.
Area of Science:
- Pediatric Medicine
- Biostatistics
- Computational Biology
Background:
- Child malnutrition is a significant global health issue with severe consequences for growth and development.
- Current screening tools for pediatric malnutrition lack consistent accuracy and practicality.
- Early detection of malnutrition risk is crucial for timely intervention and improved patient outcomes.
Purpose of the Study:
- To develop and validate novel bedside pediatric malnutrition screening models.
- To utilize machine learning and evolutionary computation for complex pattern recognition in non-invasive clinical data.
- To create practical tools for routine ward use in identifying children at risk of malnutrition.
Main Methods:
- A cross-sectional study involving 180 hospitalized children aged 1 month to 18 years.
- Development of screening models using decision trees, random forests, XGBoost, lasso regression, artificial neural networks (ANN), ANFIS, and genetic programming (GP).
- Performance evaluation against the Subjective Global Nutritional Assessment (SGNA) and physician assessment using sensitivity, specificity, AUC, and Cohen's κ.
Main Results:
- Machine learning and intelligent evolutionary models (GP, ANN, ANFIS) demonstrated high diagnostic accuracy (AUC = 0.92-1.00) and agreement (κ = 0.81-1.00) with subjective risk assessment.
- The GP model showed high accuracy but also higher complexity; simpler models like decision trees offered better interpretability and feasibility.
- Validation was limited to a small independent sample without external validation, potentially affecting generalizability.
Conclusions:
- Developed models are non-invasive, cost-effective, and show promise for early malnutrition risk detection upon hospital admission.
- Complex models may function as digital assessment tools, while simpler models are more suitable for rapid bedside screening.
- Further large-scale, multicenter studies are necessary to confirm the generalizability and clinical applicability of these screening models.
Abstract:
Background/Objectives: Child malnutrition is a global health challenge linked to poor growth, impaired development, weakened immunity, and adverse outcomes. Early risk detection is essential, but current screening tools differ in accuracy and feasibility. This study aimed to develop and validate new bedside pediatric malnutrition screening models based on machine learning and evolutionary computation methods that can capture complex patterns in non-invasive clinical indicators while remaining practical for routine ward use. Methods: We conducted a cross-sectional study including 180 hospitalized children (1 month-18 years) recruited consecutively from six pediatric wards. The required sample size (minimum 138 participants) was calculated a priori using national prevalence estimates of pediatric undernutrition (4-9.5%) to ensure adequate precision at a 95% confidence level. Data collection included a questionnaire, anthropometry, subjective malnutrition risk assessment, and the Subjective Global Nutritional Assessment (SGNA) tool. Screening models were developed using decision trees, random forests, XGBoost, lasso regression, artificial neural networks, ANFIS, and genetic programming. Their performance was evaluated against the SGNA tool and physician-based subjective malnutrition risk assessment using sensitivity, specificity, AUC, and Cohen's κ. Results: Machine learning and intelligent evolutionary models (GP, ANN, and ANFIS) showed the best performance in this sample, with substantial to high agreement (κ = 0.81-1.00) and high diagnostic accuracy (AUC = 0.92-1.00) with the subjective malnutrition risk assessment. The GP model demonstrated the highest apparent accuracy in this dataset, but also higher complexity, whereas simpler models such as decision trees showed lower accuracy but greater interpretability and feasibility for routine clinical use. However, validation was performed on a relatively small independent sample, and no external validation was conducted, which may limit the generalizability of the findings. Conclusions: While complex models may serve as digital assessment instruments, simpler models are rapid and more suitable for bedside screening. All developed models are non-invasive and cost-effective and show potential for supportive approaches for early detection of malnutrition risk at hospital admission. However, given the limited validation sample and the absence of external validation, these findings should be interpreted with caution, and further large-scale, multicenter studies are required to confirm generalizability and clinical applicability.