Academic journal article Bulletin of the World Health Organization

Standardized Interpretation of Paediatric Chest Radiographs for the Diagnosis of Pneumonia in Epidemiological Studies

Academic journal article Bulletin of the World Health Organization

Standardized Interpretation of Paediatric Chest Radiographs for the Diagnosis of Pneumonia in Epidemiological Studies

Article excerpt


Acute lower respiratory tract infection, primarily pneumonia, is the leading cause of death in childhood in developing countries, resulting in an estimated 1.9 million deaths annually (1). However, studies to determine the true burden of pneumonia and the proportion that is preventable by vaccination have been hampered by the lack of an adequate definition of pneumonia. While radiological findings are commonly accepted as the "gold standard" for defining pneumonia, there are no validated definitions for X-ray interpretation. Moreover, inter- and intra-observer variability in the interpretation of chest radiographs is a well recognized problem (2) and has been studied in the diagnosis of tuberculosis (3, 4), pneumoconiosis (5), lung cancer (6) and adult pneumonia (7, 8). For childhood pneumonia, apart from a few studies (9-12), this problem has not been adequately addressed. More specifically, there are no studies reporting multi-observer reviews of paediatric radiographs from developing countries where the quality of X-rays may be less adequate than that in developed countries.

With the availability of effective vaccines against the two leading bacterial pathogens causing childhood pneumonia in developing countries, namely Haemophilus influenzae type b (Hib) and pneumococcal conjugate vaccines, the need for standardized methods to collect data on the pneumonia disease burden and the proportion of the burden preventable by vaccines has become critical.

WHO's Department of Immunization, Vaccines and Biologicals, established a working group to standardize the categorization of radiological pneumonia, for the purpose of establishing burden estimates of likely bacterial pneumonia and estimating vaccine impact (13). This paper describes the process undertaken by the working group to achieve this objective and reports the results of a study of inter- and intra-observer variability in interpreting chest radiographs using the standardized methods. This process was meant only to standardize the interpretation of chest radiographs and not to address the question whether certain radiological patterns represent biologically or pathologically defined pneumonia.


The process of standardization consisted of three stages. These included (1) development and modification of nomenclature for visual descriptors of the characteristics of chest radiographic image; (2) learning and calibration of radiographic image interpretation and refinement of the definitions; and (3) formal measurement of inter- and intra-observer variability in interpretation.

Study participants

The participants in this process were from nine study sites that were proposing to evaluate the impact of Hib or pneumococcal conjugate vaccine. Of these, seven sites participated in the calibration phase, whereas all nine participated in the exercise to measure inter- and intra-observer variation. At each study site there were at least two X-ray readers (one radiologist and one clinician); two sites had three readers each and two sites had one common reader, who was a radiologist.

End-points and definitions

The definitions and end-points were developed through a series of workshops that included review era large number of chest radiographic images, and were further refined at a calibration workshop. In formulating the definitions and end-points, the group took several factors into consideration. Previous studies had shown that "although there is reasonable agreement on the presence of alveolar consolidation, there is considerable disagreement on other findings (8, 11, 14). Also, the presence of significant alveolar consolidation is considered by most authorities to be the most specific radiographic predictor of bacterial pneumonia. Therefore, this was chosen as the primary end-point of interest. Of necessity, the definitions used were framed to be more specific (albeit less sensitive) for likely bacterial pneumonia than those used for clinical purposes; this was considered appropriate for epidemiological studies. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.