Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 76 - Contributed Poster Presentations: Section on Statistics in Epidemiology
Type: Contributed
Date/Time: Monday, August 3, 2020 : 10:00 AM to 2:00 PM
Sponsor: Section on Statistics in Epidemiology
Abstract #312973
Title: Restricted Cubic Spline to Model Relationship Between Time to Treatment with Survival for Oral Cavity Cancer
Author(s): SONGZHU ZHAO* and Lai Wei and steve Kang and Rygalski Chandler
Companies: OSU and OSU and OSU and OSU
Keywords: Restricted cubic spline ; survival; time to treatment initiation; log-hazard ; oral cavity cancer; National Cancer Data Base

Numerous study found increasing time to treatment initiation (TTI) associated with worsening survival. The most common approach was to apply cox proportional hazard models, which assume the linear association between TTI with log-hazard. However continuous covariates can affect the hazard through more complicated nonlinear functional forms. The purpose of this study was to investigate non-linear effect of TTI on survival using restricted cubic spline. Using the National Cancer Data Base (NCDB), Cox proportional hazard models with or without restricted cubic splines were used to determine the association between TTI on survival. In unadjusted cubic spline plot, log-hazard increased with TTI before flattening when time over 90 days. The adjusted cubic spline plot showed a fairly constant log-hazard over time. This confirmed the results from cox models that longer TTI was associated with worsening survival on univariate analysis but not significant in multivariate analysis. Restricted cubic spline was a flexible approach to produce smooth curve which is easy to interpret the relationship between time to treatment initiation and log-hazard.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2020 program