Antiarrhythmic therapy for prevention of sustained ventricular tachycardia is commonly guided by programmed electrical stimulation, and the persistent ability to initiate ventricular tachycardia during drug therapy often leads to the use of nonpharmacologic therapies such as surgery. Prior studies suggest that programmed stimulation has a high sensitivity but a lower specificity for predicting recurrences of ventricular tachycardia during drug therapy. We constructed a model to evaluate the impact of various programmed stimulation specificities and sensitivities on total mortality when patients with inducible ventricular tachycardia during drug therapy proceed to arrhythmia surgery. Assumptions for this model included an 11.5% surgical mortality, a 13% risk of arrhythmia recurrence for surgery survivors, and that 33% of ventricular tachycardia recurrences are fatal. In this model the relative impact of test sensitivity and specificity on total population mortality depended on the pretest probability that drug therapy would be effective. When this probability was high, variations in test sensitivity had relatively little impact on mortality, but a low specificity increased mortality by increasing the number of false positive patients treated surgically who then suffered the surgical mortality. When the probability of effective drug therapy was low, varying test specificity had relatively little impact on mortality, but a low sensitivity increased mortality by increasing the number of patients with a false negative test who suffered fatal arrhythmia recurrences. If the specificity of programmed stimulation is low and the probability that drug therapy will be effective is high, the total population mortality can exceed the mortality expected from drug therapy alone.(ABSTRACT TRUNCATED AT 250 WORDS)