Background Patients with recurrent syncope frequently undergo extensive investigations that consume significant health care resources. Recent advances in long-term monitoring techniques have enhanced diagnostic yield in patients with infrequent symptoms. There is little information on the relative cost-effective profile of the investigative tools used in patients with syncope. Methods Two methods to determine health care costs in patients with syncope were used. In the first, health care resource utilization was determined in 24 patients with recurrent unexplained syncope and negative investigations who underwent insertion of the implantable loop recorder (ILR) during a pilot study of the feasibility of the device. The costs of investigations before, during, and after ILR implantation in each patient were calculated on the basis of median charges for an index investigation and a regression analysis of 1018 US Medicare hospital claims for syncope from 1993. Charges were converted to costs using a cost-to-charge ratio of 0.64. The second method was based on estimated costs per diagnosis and published diagnostic yields of 6 commonly applied tests in patients with syncope. A cohort simulation using theoretic models of 100 patients undergoing investigation for syncope was created to compare the diagnostic yield and cost per diagnosis of various diagnostic cascades. Results In the pilot study, the cost of investigation of syncope in the 2 years before ILR insertion was $7584 per patient. After the ILR was inserted, a diagnosis was obtained in 21 of 24 patients (diagnostic yield 88%). The cost of therapy was $2452, followed by a reduction in cost of care to $596 over 30 ± 10 months of follow-up. In the second method, the diagnostic yield of individual tests ranged from 3% for echocardiography to 88% for the ILR. The cost per diagnosis obtained ranged from $529 for the external loop recorder to $73,260 for electrophysiologic testing in patients without structural heart disease. An approach to syncope similar to that of the ILR pilot study resulted in a cost per diagnosis of $3193 and a diagnostic yield of 98%. Performance of echocardiography in half of the patients and electrophysiologic testing only in the presence of structural heart disease reduced the cost to $2494 and retained a diagnostic yield of 98%. Conclusions The cost of investigation of syncope is high. The ILR may reduce health care resource utilization by providing a diagnosis permitting definitive therapy. The cost per diagnosis profile of current diagnostic tests commonly used in patients with syncope is highly variable. A cost-effective approach to diagnosing this disorder can retain a high diagnostic yield with a reduction in resource utilization compared with a conventional approach. (Am Heart J 1999;137:870-7.)