Survival analyses for time to event outcomes in longitudinal studies
Xiangrong Kong, Johns Hopkins University, Baltimore, Maryland, United States
DisclosureBlock: Xiangrong Kong, None
Description
The overarching goal of this presentation is to introduce the fundamental statistical concepts, methods and computational tools in analyzing survival (i.e. time-to-event) data in ophthalmological studies involving observations from both eyes of an individual. The presentation will start by reviewing the key terminologies in describing time-to-event data, such as censoring, survivor function and hazard function, as well as reviewing commonly used methods to summarize and analyze survival data (Kaplan-Meier method, logrank tests, and Cox proportional hazards models).
The presentation will introduce extended Cox modeling for analysis of correlated time-to-event data (i.e. data available from both eyes). Real-world ophthalmological datasets will be used to illustrate the concepts and methods. Computational tools will also be introduced.
Participants are expected to fulfill the following learning objectives:
- To have a better understanding of situations when time-to-even data may arise in ophthalmological studies and what kinds of scientific interests may be addressed by analysis of time-to-event data.
- To be able to understand and interpret correctly key statistical concepts particularly relevant to time-to-event data analysis, such as censoring, truncation, survivor function, and hazard function.
- To have a basic knowledge of the statistical methods that can be used to analyze correlated time-to-event data.
- To have a basic knowledge of the statistical software tools that can be used for survival data analysis.