Nonparametric Estimation of Agreement Measure Between Ordinal and Censored Continuous Outcomes

• Activity Number: 81
• Type: Contributed
• Date/Time: Sunday, August 9, 2015 : 4:00 PM to 5:50 PM
• Sponsor: Biometrics Section
• Author(s): Tian Dai* and Ying Guo and Limin Peng and Amita K. Manatunga
• Companies: and Emory University and Emory University and Emory University
• Keywords: Agreement ; Censoring

Abstract

The concept of broad sense agreement (BSA) has recently been proposed for evaluating the capability of interpreting a continuous measurement according to an ordinal scale (Peng et al., JASA, 2011). The current estimator for BSA is only applicable for data with completely observed continuous and ordinal scales. In this work, we consider the problem of evaluating BSA index when the continuous variable is subject to censoring. We propose a simple plug-in estimator that uses nonparametric estimators of survival functions. The proposed BSA estimator is proven to be consistent and asymptotically normal. To achieve better performance in small sample size settings, we further study some useful variants of the proposed estimator by using the idea of multiple imputation. Our simulation results show that the imputation-based estimators have much improved performance in small sample scenarios. We illustrate our methods via an application to a peripheral artery disease data.