Models Accurately Predict Severity of Pediatric Community-Acquired Pneumonia

New models accurately predicted severity in children with community-acquired pneumonia (CAP), demonstrating potential value for clinical decision-making.1

“While only a small percentage of children with pneumonia will have severe outcomes, it’s crucial to identify these patients early so clinicians can act swiftly and aggressively to prevent further deterioration in these children,” lead investigator Todd Florin, MD, MSCE, Associate Division Head for Academic Affairs & Research for the Division of Pediatric Emergency Medicine at Ann & Robert H. Lurie Children’s Hospital of Chicago and Associate Professor of Pediatrics at Northwestern University Feinberg School of Medicine, said in a statement.2 “It is also important to know if the illness will likely be mild, in order to avoid potentially unnecessary tests or treatments or unnecessary hospital stays.”

Florin and colleagues conducted a prospective cohort study of 73 emergency departments in 14 countries including children aged 3 months to <14 years with clinical diagnoses of CAP from February 2019 to June 2021. Exclusion criteria included recent hospitalization and chronic complex conditions. The study defined CAP severity by treatment course and outcomes within 7 days of the ED visit: mild CAP as treated in the outpatient setting or less-than-24-hour hospitalization with no use of oxygen or intravenous fluid, moderate CAP as hospitalization for under 24 hours with oxygen or fluids, or hospitalization of at least 24 hours regardless of interventions but without an outcome qualifying as severe CAP, and severe CAP as needing chest drainage, intensive care unit admission of over 24 hours, positive-pressure ventilation, septic shock, vasoactive infusions, extracorporeal membrane oxygenation, or death.1

“Emergency departments around the world see thousands of children with pneumonia every day, but until now, we haven’t had a reliable way to predict who’s truly at risk of getting sicker,” investigator Nathan Kuppermann, MD, MPH, Executive Vice President, Chief Academic Officer and Director of the Children's National Research Institute in Washington, DC, added.2 “This model gives clinicians a practical tool, rooted in data, to guide that decision and ultimately improve care and outcomes.”

The study included 2222 children, around held of which were female (n = 1103; 49·7%), with a median age of 3 years (IQR 1–5). CAP severities included mild (n = 1290; 58·1%), moderate (n = 812; 36·5%) and severe (n = 120; 5·4%). Primary analyses included 1901 patients with mild (n = 1011; 53.2%), moderate (n = 772; 40.6%), or severe (n = 118; 6.2%) CAP.1

The investigators found that congestion or rhinorrhea was negatively associated with moderate or severe CAP (adjusted odds ratio [aOR], 0·59 [95% CI, 0·46–0·76]), while abdominal pain (aOR, 1·52 [95% CI, 1·17–1·97]), refusal to drink (aOR, 1·57 [95% CI, 1·24–2·00]), antibiotics before ED visit (aOR, 1·64 [95% CI, 1·29–2·10]), chest retractions (aOR, 2·86 [95% CI, 2·24–3·65]), respiratory rate above the 95th percentile for age (aOR, 1·63 [95% CI, 1·29–2·06]), heart rate above the 95th percentile for age (aOR, 1·64 [95% CI, 1·27–2·12]), and hypoxaemia (oxygen saturation 90–92%: aOR, 3·24 [95% CI, 2·46–4·27]; <90%: aOR, 13·39 [95% CI, 8·64–20·73]) were positively associated.1

Florin and colleagues found that the predictive model accurately discriminated between mild CAP and moderate or severe CAP (c-statistic, 0·82 [95% CI, 0·80–0·84]). The model performed similarly in those with radiographic CAP; these participants were also found to have additional predictors of decreased breath sounds and multifocal (c-statistic, 0·82 [95% CI, 0·80–0·85]).1

“Our pediatric pneumonia predictive models show good-to-excellent accuracy,” Florin said.2 “They appear to perform better than clinician judgment alone in predicting illness severity, according to previous research from Lurie Children’s. Once externally validated, our models will provide evidence-based information for clinicians to consider when evaluating pneumonia in children.”

REFERENCES

1. Florin TA, Tancredi DJ, Ambroggio L, et al. Predicting paediatric pneumonia severity in the emergency department: a multinational prospective cohort study of the Pediatric Emergency Research Network. Lancet Child Adolesc Health. 2025; 9(6): 383-392. DOI: 10.1016/S2352-4642(25)00094-X

2. Models Predict Severity of Pneumonia in Kids to Help Guide Treatment. News release. Lurie Childrens. May 15, 2025. https://www.luriechildrens.org/en/news-stories/models-predict-severity-of-pneumonia-in-kids-to-help-guide-treatment/