Accuracy of the interpretation of chest radiographs for the diagnosis of paediatric pneumonia

PLoS One. 2014 Aug 22;9(8):e106051. doi: 10.1371/journal.pone.0106051. eCollection 2014.


Introduction: World Health Organization (WHO) radiological classification remains an important entry criterion in epidemiological studies of pneumonia in children. We report inter-observer variability in the interpretation of 169 chest radiographs in children suspected of having pneumonia.

Methods: An 18-month prospective aetiological study of pneumonia was undertaken in Northern England. Chest radiographs were performed on eligible children aged ≤16 years with clinical features of pneumonia. The initial radiology report was compared with a subsequent assessment by a consultant cardiothoracic radiologist. Chest radiographic changes were categorised according to the WHO classification.

Results: There was significant disagreement (22%) between the first and second reports (kappa = 0.70, P<0.001), notably in those aged <5 years (26%, kappa = 0.66, P<0.001). The most frequent sources of disagreement were the reporting of patchy and perihilar changes.

Conclusion: This substantial inter-observer variability highlights the need for experts from different countries to create a consensus to review the radiological definition of pneumonia in children.

Publication types

  • Multicenter Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adolescent
  • Child
  • Child, Preschool
  • England
  • Female
  • Humans
  • Infant
  • Male
  • Observer Variation*
  • Pneumonia / diagnosis
  • Pneumonia / diagnostic imaging*
  • Pneumonia / etiology
  • Prospective Studies
  • Radiography, Thoracic*
  • World Health Organization

Grant support

The authors declare that this study was supported by a grant (No: 0887X1-4479) from Pfizer Vaccines UK. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.