In patients suspected of having coronary artery disease (CAD), noninvasive testing has been playing an increasing role in selecting patients who would require coronary angiography for either the "definitive" diagnosis of CAD or as a prelude to planning myocardial revascularization. A mathematic model is presented that defines cost-effective utility of nuclear cardiology testing for diagnosis of CAD and selection of appropriate candidates for coronary angiography, according to quantitative methods of decision analysis. Clinical utility or effectiveness was defined in terms of percent correct diagnosis of CAD. Cost was defined as dollars of medical expenditure. Six competing strategies were compared in subsets of patients with different pretest likelihoods of CAD, based on age, sex, and symptoms. Nuclear cardiology testing was the most cost-effective initial modality of choice in patients with an intermediate pretest likelihood of CAD. In patients with a low pretest likelihood of CAD, nuclear cardiology testing was cost-effective in the subgroup of patients who had abnormal exercise treadmill electrocardiograms. In patients with a high pretest likelihood of CAD, direct referral to coronary angiography was the most cost-effective strategy for diagnosis of CAD. Coronary angiography, however, is performed most often as a prelude to myocardial revascularization. Because these invasive procedures are indicated only in patients who are at high risk with medical therapy, nuclear cardiology procedures, by virtue of incremental prognostic information, identify appropriate candidates for more invasive procedures, aimed at improving survival. Strategies for cost-effective prognostication of CAD depend on not only the patient's pretest likelihood of CAD but also the status of the rest electrocardiogram. In patients with a normal rest electrocardiogram, a low pretest likelihood of CAD indicates a low risk for cardiac events with medical therapy. Therefore coronary angiography is not indicated in these patients. Patients with an intermediate likelihood of CAD should first undergo exercise electrocardiographic testing; a negative response would indicate a low risk for cardiac events and a positive response would indicate the need for nuclear cardiology testing for further cost-effective risk stratification. In patients with a high pretest likelihood of CAD, the combined exercise electrocardiographic and nuclear cardiac testing is the most cost-effective strategy; a negative or a positive nuclear test result would imply low or high risk, respectively. The latter patients would then be candidates for coronary angiography. In all patients with an abnormal rest electrocardiogram, the most cost-effective strategy is uniform referral to nuclear cardiac testing (which is performed in conjunction with exercise electrocardiography), regardless of the pretest likelihood of CAD; a negative or a positive nuclear test result would indicate low or high risk for coronary events, respectively. The latter group would be proper candidates for referral to coronary angiography.