EVALUATION OF COMPETING TESTS FOR THE DIAGNOSIS OF PULMONARY-EMBOLISMAND DEEP-VEIN THROMBOSIS .1.

Tests for pulmonary embolism (PE) and its most frequent source, deep v ein thrombosis (DVT), include angiography, ventilation-perfusion nucle ar medicine scans, venous sonography with Doppler, and contrast and ra dionuclide leg venography. Although selective angiography is the defin itive procedure for the diagnosis of PE, the associated risk of death, although small, as well as the morbidity associated with injection of contrast agents, are high enough that alternative, less accurate, but safer diagnostic procedures are performed in an attempt to avoid the higher-risk procedure. Effective cost (EC) of each test represents the dollars spent per unit of diagnostic information and is defined as th e ratio of the expected direct cost (EDC) of the test to its diagnosti c performance (DU). EDC includes the base cost or charge of the test a nd the estimated cost of the morbidity and mortality that can be incur red in performing the test, while DU is determined from the test sensi tivity and specificity. With the lowest EC as the selection criterion for the best test and representative costs, sensitivity, specificity, and morbidity and mortality rates, five different tests for PE or DVT were compared. Doppler sonography yielded the most diagnostic informat ion per dollar spent, as its EC was the lowest, primarily because its base cost was low compared to that of the other tests. Radionuclide le g venography had the second lowest EC. Selection among the remaining t hree tests depended on the prevalence of PE and morbidity and mortalit y costs. This methodology allows for comparison of individual tests in any given clinical setting as well as determination of the maximum co st and minimum performance characteristics for any new test to be comp etitive with already existing ones.