Analysis of Mixed Symptomatic and Asymptomatic Data Using a Current Status Design

• Activity Number: 574
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Victor Giancarlo Sal Y Rosas*+ and James Hughes
• Companies: Pontificia Universidad Católica del Perú and University of Washington
• Keywords: Current status data ; Outcome misclassification ; Sensitivity ; Specificity ; symptomatic data

Abstract

In this article, we consider the statistical analysis of a current status data design when the laboratory test that measures the outcome of interest is subject to error and when some visits are informative due to symptomatic events. Our proposed methodology establishes a general framework using the EM algorithm to estimate the cumulative distribution function of the disease of interest (one sample problem). In addition, we proposed a statistic based on the mean difference of the survival functions of two groups for the two-sample hypothesis testing problem. Finally, we extend the EM algorithm to implement a proportional hazard model that takes into account the misclassification rates and the symptomatic data. Our methods are motivated and demonstrated by data collected from an infectious disease study in Seattle, WA