Diagnostic tests for tuberculosis and their evaluation.

Abstract

In addition to clinical and radiological examinations, the diagnosis of tuberculosis may be made by detection of the causative organism or components thereof, by assay of specific immune responses and by detection of various secondary phenomena resulting from the disease process. Much effort has been expended on the development of sensitive, specific and robust tests for tuberculosis but with limited success. Much of the problem stems from the ways in which new tests have been evaluated and compared with others. The sensitivity and specificity of a new test may be easily calculated but are of limited use unless the predictive values of positive and negative results, appropriate to the circumstances under which the test will be used, are also calculated. Tests must not only be evaluated for their efficiency. It is of even greater importance to evaluate the robustness of a test so as to determine whether it will be suitable for providing the basis for a sustained, reliable and clinically useful diagnostic service. For this purpose, international co-operative and comparative studies, together with guidelines for the assessment of the value of new tests are required.