Endgames Statistical Question

Cohen’s coefficient κ

BMJ 2012; 344 doi: https://doi.org/10.1136/bmj.e1178 (Published 22 February 2012) Cite this as: BMJ 2012;344:e1178
  1. Philip Sedgwick, senior lecturer in medical statistics
  1. 1Centre for Medical and Healthcare Education, St George’s, University of London, Tooting, London, UK
  1. p.sedgwick{at}sgul.ac.uk

Chest radiographs are the best method for diagnosing pneumonia but are often not available in developing countries. Therefore, in 1990 the World Health Organization developed guidelines for diagnosis of non-severe pneumonia that comprised clinical symptoms of fast breathing alone. However, fast breathing can have causes other than pneumonia, and thus children who are given a diagnosis of non-severe pneumonia on the basis of fast breathing alone may receive antibiotics unnecessarily.

Children aged 2 to 59 months with non-severe pneumonia diagnosed on the basis of the WHO guidelines were invited to participate from outpatient departments of six hospitals in Pakistan. In total 2000 children were enrolled, for whom 1848 chest radiographs were available for assessment. Two consultant radiologists used standardised criteria to evaluate the chest radiographs, with no clinical information available to them. The primary outcome was diagnosis of pneumonia (absent or present) from chest radiographs.1

Cohen’s coefficient κ for agreement between the two radiologists in their diagnoses was 0.46. A small number of children were given a diagnosis of bronchiolitis. The researchers concluded that most children with non-severe pneumonia diagnosed on the basis of the current WHO definition had normal chest radiographs.

Which of the following statements, …

View Full Text

Sign in

Log in through your institution

Free trial

Register for a free trial to thebmj.com to receive unlimited access to all content on thebmj.com for 14 days.
Sign up for a free trial

Subscribe