Ph.D. candidate
Vo Van Thi
Department of Pediatrics, Can Tho University of Medicine and Pharmacy; Can Tho Children’s Hospital
Title: Clinical and NLR-Based Prediction of Respiratory Support Requirement in Children Hospitalized with Pneumonia in Southern Vietnam
Name(s): Thi Van Vo, Phuong Minh Nguyen, Minh Dien Thai.
Department: Department of Pediatrics, Can Tho University of Medicine and Pharmacy, Can Tho, Vietnam; Can Tho Children’s Hospital, Can Tho, Vietnam; Faculty of Medicine, Nam Can Tho University, Can Tho, Vietnam
Abstract
Background:
Respiratory support requirement among children hospitalized with pneumonia is an important marker of disease severity and hospital resource needs. However, practical risk stratification tools for routine pediatric hospital settings in Southern Vietnam remain limited. This study aimed to develop and evaluate clinical and laboratory-based prediction models for in-hospital respiratory support requirement among children aged 2–59 months hospitalized with pneumonia.
Methods:
We conducted a retrospective cohort study at a tertiary pediatric hospital in Southern Vietnam from July 2024 to November 2025. Children aged 2–59 months hospitalized with pneumonia were included after predefined exclusions. The outcome was respiratory support requirement during hospitalization, including oxygen therapy, continuous positive airway pressure, or invasive mechanical ventilation, analyzed as a binary endpoint for model development. Candidate predictors included age <12 months, malnutrition, recurrent pneumonia, cyanosis, tachypnea, chest indrawing, and complete blood count-derived inflammatory indices. Univariable logistic regression and two multivariable logistic regression models were developed: Model 1 using clinical predictors only and Model 2 using clinical predictors plus neutrophil-to-lymphocyte ratio (NLR). Model discrimination was assessed using area under the receiver operating characteristic curve, and calibration was evaluated using the Hosmer–Lemeshow test and observed-to-expected ratio.
Results:
Among 1,797 children included in the final analysis, 154 (8.6%) required respiratory support. Of these, 114 received oxygen only, 24 required continuous positive airway pressure without intubation, and 16 required invasive mechanical ventilation. Children requiring respiratory support were younger and had higher proportions of malnutrition, recurrent pneumonia, cyanosis, tachypnea, and chest indrawing. In the primary model, independent predictors of respiratory support were age <12 months (adjusted odds ratio [aOR] 2.57, 95% CI 1.69–3.92), malnutrition (aOR 4.33, 95% CI 2.56–7.33), recurrent pneumonia (aOR 1.82, 95% CI 1.18–2.81), cyanosis (aOR 24.02, 95% CI 7.41–77.87), chest indrawing (aOR 4.19, 95% CI 2.73–6.43), and higher NLR (aOR 1.49 per 1-unit increase, 95% CI 1.38–1.60). Tachypnea was not independently associated after adjustment. Discrimination improved from the clinical-only model (AUC 0.754) to the clinical plus NLR model (AUC 0.840, 95% CI 0.805–0.873). At the optimal probability cut-off of 0.122, the final model achieved 66.2% sensitivity, 86.2% specificity, 31.1% positive predictive value, 96.5% negative predictive value, and 84.5% accuracy. Calibration was acceptable, with Hosmer–Lemeshow p=0.662 and an observed-to-expected ratio of 1.00.
Conclusion:
A simple clinical model strengthened by NLR provided good discrimination and calibration for predicting in-hospital respiratory support requirement among children under five hospitalized with pneumonia in Southern Vietnam. This model may support early triage, prioritization of monitoring intensity, and escalation readiness in resource-constrained pediatric settings. External validation is warranted before wider implementation.
Keywords: pediatric pneumonia; respiratory support; risk prediction; neutrophil-to-lymphocyte ratio; logistic regression; early diagnostics; Vietnam
Biography of the presenting author
Dr. Vo Van Thi, MD, MSc, is a pediatric neurologist and lecturer in Pediatrics at Can Tho University of Medicine and Pharmacy, Vietnam, and a pediatric neurologist at Can Tho Children’s Hospital. He is an admitted PhD candidate at Karolinska Institutet, Sweden, with a research focus on integrated early screening and precision biomarkers for autism in Vietnam and Sweden, and has also received a PhD offer from the University of New South Wales, Australia. He holds a Master’s degree in Pediatrics, a Bachelor’s degree in English with distinction, and professional training in pediatric speech therapy. He completed an international clinical observership in complex autism and neurodevelopment at Monash University and Monash Children’s Hospital, Australia. His academic and clinical work focuses on pediatric neurology, epilepsy, neurodevelopmental disorders, autism, pediatric infectious diseases, and artificial intelligence applications in child health. He is a Fellow of the ICNA Global Burden of Disease Program 2025 and a member of the International Child Neurology Association and the American Epilepsy Society.
Presenting author details
Full name: Vo Van Thi
Contact number: +84 336 780 819
Email: vvthi@ctump.edu.vn
LinkedIn: N/A
ResearchGate: N/A
Google Scholar: N/A
ResearcherID: N/A
ORCID: N/A
Website: N/A
Session name/number: Early diagnostics
Category: Oral presentation