A functional approach to study the relationship between dynamic covariates and survival outcomes: an application to a randomized clinical trial on osteosarcoma

Author(s):
Spreafico, M. (0000-0002-7773-9976); Ieva, F.; Fiocco, M.
Editor(s):
Pollice, A.; Salvati, N.; Schirripa Spagnolo, F.
Published
2020
Publication type
Article in monograph or in proceedings

Abstract

In clinical research, associating dynamic time-varying covariates (e.g. biomarkers or drug assumption) with an event-time outcome represents a challenging task that could be tackled exploiting Functional Data Analysis (FDA). In particular, FDA techniques can be used to represent dynamic time-varying covariates in terms of functions, which can be plugged into a Cox-type regression model to investigate the effect on survival outcomes. Data from MRC BO06/EORTC 80931 randomised controlled trial for treatment of osteosarcoma were analysed. Time-varying covariates related to alkaline phosphatase levels and chemotherapy dose during treatment were considered.

Functional Data AnalysisSurvival ModelsOsteosarcoma
Date issued
2020
Language
English
Publisher
Pearson Italia
Is part of
Book of Short Papers SIS 2020 (2020), p. 727-732
ISBN source
9788891910776
hdl
1887/3673400