Threshold analyses were performed, with both mortality and marginal cost-effectiveness as primary outcome measures, for those variables where no reliable data were available in the literature, or where values may have varied with the local situation. The aim of the threshold analysis was to determine the value of those uncertain variables at which the optimal strategy becomes equally as cost-effective, or as effective overall, as another strategy. Variables varied in the threshold analysis were:
the pretest probability of PE and coexisting DVT (deep vein thrombosis);
the probability of mortality due to untreated PE and DVT;
ANALYSIS>> and the accuracies of spiral CT angiography for detecting PE, of the D-dimer blood assay for detecting fibrinogen degradation products, and of US for detecting DVT.