Background: Coronary CT angiography (CCTA) has been proven accurate and is incorporated in clinical recommendations for coronary artery disease (CAD) diagnosis workup, but cost-effectiveness data, especially in comparison to other methods such as myocardial single photon emission CT (SPECT) are insufficient.
Objective: To compare the cost-effectiveness of CCTA and myocardial SPECT in a real-world setting.
Methods: We performed a retrospective cohort study on consecutive patients with suspected CAD and a pretest probability between 10% and 90%. Test accuracy was compared by correcting referral bias to coronary angiography depending on noninvasive test results based on the Bayes' theorem and also by incorporating 1-year follow-up results. Cost-effectiveness was analyzed using test accuracy and quality-adjusted life year (QALY). The model using diagnostic accuracy used the number of patients accurately diagnosed among 1000 persons as the effect and contained only expenses for diagnostic testing as the cost. In the model using QALY, a decision tree was developed, and the time horizon was 1 year.
Results: CCTA was performed in 635 patients and SPECT in 997 patients. An accurate diagnosis per 1000 patients was achieved in 725 patients by CCTA vs 661 patients by SPECT. In the model using diagnostic accuracy, CCTA was more effective and less expensive than SPECT ($725.38 for CCTA vs $661.46 for SPECT). In the model using QALY, CCTA was generally more effective in terms of life quality (0.00221 QALY) and cost ($513) than SPECT. However, cost utility varied among subgroups, with SPECT outperforming CCTA in patients with a pretest probability of 30% to 60% (0.01890 QALY; $113).
Conclusion: These results suggest that CCTA may be more cost-effective than myocardial SPECT.
Keywords: Coronary artery disease; Coronary computed tomography angiography; Cost-effectiveness; Myocardial single photon emission computed tomography.
Copyright © 2015 Society of Cardiovascular Computed Tomography. Published by Elsevier Inc. All rights reserved.