Letters to the editors

Abstract

Extract

Madam:- Since a number of animal diseases such as brucellosis and tuberculosis in cattle, and hydatids in sheep are now reaching a low prevalence (P: percentage of diseased animals in a population), it is perhaps appropriate to remind readers that prevalence has a marked influence on the accuracy with which a diagnostic test can identify diseased animals. Whereas an ideal test will give positive results with all diseased animals and negative results with all disease-free animals, the ideal is rarely attainable and most tests return some false positive and false negative results. Many published diagnostic tests(2) ⁽⁴⁾ (5) have a sensitivity (Se: percentage of diseased animals with a positive result) and a specificity (Sp: percentage of disease-free animals with a negative result) of between 90 and 100%, and perform well when the prevalence of a particular disease is high. But, when it is low, as is the case when a disease is nearly eradicated, or in a group of animals selected for export, the same tests have a much lower accuracy in the detection of diseased animals. ⁽³⁾ It is not the tests that have changed, but the conditions under which they are used.