Evaluation for Coronary Artery Disease and Medicare Spending

To the Editor: Dr Shreibati and colleagues proposed several important explanations for the differences in expenditures between Medicare patients evaluated for coronary heart disease (CHD) with coronary computed tomography angiography (CCTA) or conventional stress testing.¹ However, they place too little emphasis on the potential for an unmeasured difference in sensitivity between the 2 approaches. This difference is likely to be substantially more than what one might infer from the literature because of a failure to adjust for verification bias (sometimes called workup or referral bias) in most studies that have evaluated stress testing.² Because patients with positive stress test results are more likely to undergo follow-up cardiac catheterization, the sensitivity and specificity derived from a population selected for angiographic confirmation are overestimated and underestimated, respectively.^3- 4 Studies of CCTA may be affected by the same bias, but to a lesser degree because the test is frequently performed as an adjunct study and its results may not drive the decision to pursue cardiac catheterization to the same extent.