Author Affiliations: Division of Cardiology, Cedars-Sinai Medical Center, Los Angeles, California.
Chow et al argue that their data are not subject to verification bias because all screened and enrolled patients underwent ICA irrespective of the computed tomography results. This defense falls short on 2 counts.
First, the authors' study comprised 169 patients undergoing coronary angiography among a total of 594 candidates with suspected coronary artery disease. Even if all 169 patients underwent both forms of imaging, the remaining 425 candidates might well exhibit a preferential referral away from diagnostic verification. Without knowing the proportion of positive and negative test responses in all subjects (those referred for verification and those not referred for verification), one cannot determine the presence or degree of verification bias.1
Diamond GA, Kaul S. Using the Appropriate Fishing Net for Computed Tomographic Coronary Angiography in Daily Clinical Practice—Reply. Arch Intern Med. 2011;171(16):1512-1513. doi:10.1001/archinternmed.2011.377