Textbook of Respiratory Medicine

Abstract

Wegener's granulomatosis Positive (F) ANCA test Negative (F) Negative (D) Diagnostic specificity P(DIF) = 18/21 86% Diagnostic sensitivity P(IDF) = 214/219 98% Positive (D) 18 SI," s 2 1 ~~~~9 Nosological Nosological sensitivity specificity P(FID) = 18/23 P(F/D) = 214/217 78% 99%Nosological and diagnostic sensitivity and specificity of the test for antibodies to neutrophil cytoplasmic antigens (ANCA) used to assess the likelihood of disease according to the result of the test so long as the probabilities of disease in the diagnostic setting and in the patients studied for the paper are similar.On the other hand, nosological values of sensitivity and specificity are of limited direct relevance to the practising physician, who would need to convert these probabilities to the clinically relevant probabilities using Bayes's formula.This conversion is to a large extent dependent on the prior probability of disease in the clinical population studied and this is rarely known.There is a lot to be said for collecting data in a realistic clinical setting, as the authors have done, and quoting diagnostic specificity and sensitivity, rather than evaluating a new test in a group ofpatients who are known to have the disease in question and comparing this with the results in a group of normal people.In these circumstances the nosological value would be more appropriate.Taking the bright, coarsely granular pattern as positive, the authors found that the result of the test was positive in 18 patients with Wegener's granulomatosis and in three who did not have the disease.The result was negative in five patients who had active Wegener's granulomatosis and in 214 other patients, in most of whom, from the limited details given, Wegener's granulomatosis might have been initially suspec- ted but was not subsequently confirmed.On the basis ofthese figures (table) the nosological sensitivity is 78%, the nosological specificity 99%, the diagnostic sensitivity 98%, and the diagnostic specificity 86%.It would seem that the authors have quoted the nosological sensitivity with the diagnostic specificity.I would suggest that it would be more appropriate for them to have quoted the diagnostic sen- sitivity of98%-which, as an added bonus, looks even better than the quoted sensitivity of 78%.