[Application of Spline-Based Cox Regression on Analyzing Data from Follow-Up Studies].

Ying Dong,Jin-ming Yu,Da-yi Hu
DOI: https://doi.org/10.3760/cma.j.issn.0254-6450.2012.09.021
2012-01-01
Abstract:With R, this study involved the application of the spline-based Cox regression to analyze data related to follow-up studies when the two basic assumptions of Cox proportional hazards regression were not satisfactory. Results showed that most of the continuous covariates contributed nonlinearly to mortality risk while the effects of three covariates were time-dependent. After considering multiple covariates in spline-based Cox regression, when the ankle brachial index (ABI) decreased by 0.1, the hazard ratio (HR) for all-cause death was 1.071. The spline-based Cox regression method could be applied to analyze the data related to follow-up studies when the assumptions of Cox proportional hazards regression were violated.
What problem does this paper attempt to address?