Abstract W P285: Novel Imaging Technology to Select Patients with a Recent Transient Ischemic Attack or Minor Ischemic Stroke for Carotid Endarterectomy: The Relationship between Test Performance and Cost-Effectiveness

Abstract

Background and Purpose: Non-invasive molecular imaging tests are being developed to improve the ability to predict future strokes in patients with a recent transient ischemic attack (TIA) or minor ischemic stroke. Greater predictive ability can improve patient management, e.g., by identifying which patients will benefit more from a carotid endarterectomy than from medication alone. We estimated the minimum performance (i.e., sensitivity and specificity) that a new test must have in order for it to be cost-effective versus currently available strategies. Methods: We compared the cost-effectiveness of using a new imaging test (as a confirmatory test after an initial duplex ultrasonography) with a guidelines-based strategy and three strategies found in daily practice. Cost-effectiveness modelling was used to estimate the long-term costs and health outcomes of each strategy. A willingness-to-pay threshold of є30,000 per QALY gained was used to evaluate cost-effectiveness. We examined the results in two hypothetical populations (60-year-old and 80-year-old men) and varied the sensitivity and specificity to estimate the minimum test performance needed in order for the new strategy to be cost-effective versus the alternatives. Results: A perfect confirmatory test (100% sensitivity and specificity) at a cost of є390 is cost-effective for 60-year-old men versus all comparators. A test that is 100% sensitive must be at least 71% specific to be cost-effective versus the guidelines. A test that is 100% specific must be at least 52% sensitive to be cost-effective. Assuming 90% sensitivity, a test must have a specificity of at least 77% to be cost-effective. The minimum required performance for 80-year-old men was higher; e.g., a test that is 100% sensitive must be 88% (vs. 71%) specific to be cost-effective. Conclusions: A new strategy that improves risk prediction in patients with a recent TIA or minor ischemic stroke may help to reduce the risk of recurrent stroke and thereby improve health outcomes and cost-effectiveness. However, the minimum required performance may not be achievable in all patient subgroups.