Related Experiment Videos
[Time-dependent covariates in the Cox proportional hazards model. Theory and practice]
1INSERM-INED U569, Service d'Epidémiologie, Hôpital de Bicêtre, 82, rue du Général-Leclerc, 94276 Le Kremlin-Bicêtre. desquilbet@vjf.inserm.fr
Revue D'Epidemiologie Et De Sante Publique
|May 13, 2005
Summary
This study explains how to handle time-dependent covariates in survival analysis using the Cox proportional hazards model. It details methods for unique and multiple exposure changes, with SAS programs provided.
Area of Science:
- Statistics
- Epidemiology
Context:
- Survival analysis requires accounting for exposures that change over time.
- Time-dependent covariates are crucial in the Cox proportional hazards model.
Purpose:
- To present theoretical and practical methods for incorporating time-dependent covariates in survival analysis.
- To illustrate these methods using a small French HIV cohort study.
Summary:
- The study defines two types of time-dependent covariates: unique and multiple changes.
- It demonstrates how to implement these in the Cox model, addressing challenges like missing data.
- SAS programs are included for practical application.
Impact:
- Improves the accuracy of survival analyses when exposures are dynamic.
- Provides a practical guide for researchers using time-dependent covariates.
- Facilitates better understanding of exposure effects in longitudinal studies.