Decision Rule Falls Short in Identifying Children at Low Risk for Pneumonia

In a prospective cohort study of 2574 patients (age, <21 years; median age, 2.3 years) who presented to a tertiary pediatric emergency department and who underwent chest radiography for suspected pneumonia, investigators sought to develop a clinical decision rule for chest radiography.

Overall, 422 patients (16%) had radiographic pneumonia (radiograph suggestive of pneumonia); a subset of 199 patients (8% of the total) had definite pneumonia (radiograph diagnostic for pneumonia). In multivariate analysis, significant predictors of both radiographic pneumonia and definite pneumonia (with odds ratios ranging from 1.37–3.69) were oxygen saturation (O₂ sat) ≤92% at triage, chest pain, focal rales, and O₂ sat between 93% and 96% at triage. Wheezing on exam was a significant negative predictor in the entire cohort but not in patients younger than 5 years.

Recursive partitioning stratified patients into risk groups as follows: 20 patients with O₂ sat ≤92%, no wheeze, and focal rales had the highest risk for radiographic pneumonia (pneumonia rate, 70.0%); 38 patients with O₂ sat >92%, no history of fever, no focal decreased breath sounds, and no focal rales were at lowest risk (pneumonia rate, 7.6%). Among patients younger than 5 years, O₂ sat ≤92% at triage was the best predictor of radiographic pneumonia (pneumonia rate 39% vs. 15% for those with O₂ sat >92%); investigators were unable to characterize low-risk characteristics in patients younger than 5.