Evaluation of diagnostic procedures

In this article by J André Knottnerus and colleagues (23 February, pp 477-80), one of the values in the box showing the derivation of measures of discrimination (p 478) was wrong. The last sentence in the second section of text in this box should read “The prior probability of disease is the prevalence of D in the population to be tested: 200/1000=0.20”, with D being the x ray result (fracture or no fracture).