←Back
527 – Diagnostic Tests: Student Papers and Correlated Data
Methods for Diagnostic Performance Comparison with Correlated Data
Sponsor: Section on Medical Devices and Diagnostics
Keywords: diagnostic, correlated, Z-test, Bootstrap, Bayesian, Power
Xuan Ye
U.S. Food and Drug Administration, Silver Spring, Maryland
In new diagnostic device application, such as 510(k), the new measurement method or diagnostic device is often compared to a predicate device or a comparator. We need to demonstrate that the diagnostic performance of the subject device is substantially equivalent to the predicate's. In clinical trials, both methods are often applied to the same set of patients producing correlated dataset. In this talk, we compare various parametric and non-parametric methods to address the issue and discuss the analysis results through simulation studies.