Measuring Agreement in Method Comparison Studies with Repeated Measurements

• Activity Number: 596
• Type: Topic Contributed
• Date/Time: Wednesday, August 12, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Medical Devices and Diagnostics
• Author(s): Zhiheng Xu* and Meijuan Li
• Companies: FDA and FDA
• Keywords: Diagnostic monitoring device ; correlation ; measurement error ; repeated measurement ; agreement ; predicate

Abstract

For diagnostic monitoring devices such as continuous blood pressure monitor or glucose monitor, each subject could have repeated measurements recorded on both test device and predicate device over a period of time. To assess the agreement between two devices, the within-subject correlation has to be taken into consideration since replicate observations within the same subject could be correlated. In addition, both new device and the predicate have repeated measurements over the time and both covariate (the predicate) and response variable (test device) are subject to measurement errors. Imprecision in the predicate device i.e. the covariate could result in a biased point estimate for the model parameters if it is not properly addressed. In this talk, we will propose and discuss a statistical methodology on how to assess the agreement in method comparison studies with repeated measurements when the correlation structure and measurement errors are both considered.