Incorporating the Presence of Error-Prone Diagnostic Tests and Self-Reported Outcomes in Time-to-Event Models

• Activity Number: 452
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Xiangdong Gu*+ and Raji Balasubramanian
• Companies: and University of Massachusetts, Amherst
• Keywords: self-report ; tim-to-event outcomes

Abstract

We consider studies in which data from sequentially administered, error-prone, laboratory-based diagnostic tests or self-reported questionnaires are collected to determine the occurrence of a silent event. The problem is motivated by the Women's Health Initiative, in which information on diabetes status among the approximately 160,000 women is obtained from annually collected self-reported data. For this setting, we implement a likelihood-based approach and incorporate a proportional hazards model to simultaneously evaluate the association of relevant covariates with outcome, while adjusting for the inherent error in self-reported outcomes. As diabetes status at baseline is also ascertained through self-reports, we also accommodate for the error-prone disease status ascertainment at baseline. We apply this approach to data collected in the Women's Health Initiative to evaluate the effects of elevated depressive symptoms and anti-depressant use on the risk of diabetes, and compare the results obtained using traditional Cox proportional hazards models.