Home > Research > Publications & Outputs > Modelling time to event with observations made ...
View graph of relations

Modelling time to event with observations made at arbitrary times

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Published

Standard

Modelling time to event with observations made at arbitrary times. / Sperrin, Matthew; Buchan, Iain.
In: Statistics in Medicine, Vol. 32, No. 1, 15.01.2013, p. 99-109.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Harvard

Sperrin, M & Buchan, I 2013, 'Modelling time to event with observations made at arbitrary times', Statistics in Medicine, vol. 32, no. 1, pp. 99-109. https://doi.org/10.1002/sim.5509

APA

Sperrin, M., & Buchan, I. (2013). Modelling time to event with observations made at arbitrary times. Statistics in Medicine, 32(1), 99-109. https://doi.org/10.1002/sim.5509

Vancouver

Sperrin M, Buchan I. Modelling time to event with observations made at arbitrary times. Statistics in Medicine. 2013 Jan 15;32(1):99-109. doi: 10.1002/sim.5509

Author

Sperrin, Matthew ; Buchan, Iain. / Modelling time to event with observations made at arbitrary times. In: Statistics in Medicine. 2013 ; Vol. 32, No. 1. pp. 99-109.

Bibtex

@article{bd03b53a90484544a0fdd8d8c2ab11fb,
title = "Modelling time to event with observations made at arbitrary times",
abstract = "We introduce new methods of analysing time to event data via extended versions of the proportional hazards and accelerated failure time (AFT) models. In many time to event studies, the time of first observation is arbitrary, in the sense that no risk modifying event occurs. This is particularly common in epidemiological studies. We show formally that, in these situations, it is not sensible to take the first observation as the time origin, either in AFT or proportional hazards type models. Instead, we advocate using age of the subject as the time scale. We account for the fact that baseline observations may be made at different ages in different patients via a two stage procedure. First, we marginally regress any potentially age-varying covariates against age, retaining the residuals. These residuals are then used as covariates in the fitting of either an AFT model or a proportional hazards model. We call the procedures residual accelerated failure time (RAFT) regression and residual proportional hazards (RPH) regression respectively. We compare standard AFT with RAFT, and demonstrate superior predictive ability of RAFT in real examples. In epidemiology, this has real implications in terms of risk communication to both patients and policy makers.",
keywords = "accelerated failure time, proportional hazards, survival analysis, time origin, time-varying covariates",
author = "Matthew Sperrin and Iain Buchan",
year = "2013",
month = jan,
day = "15",
doi = "10.1002/sim.5509",
language = "English",
volume = "32",
pages = "99--109",
journal = "Statistics in Medicine",
issn = "1097-0258",
publisher = "John Wiley and Sons Ltd",
number = "1",

}

RIS

TY - JOUR

T1 - Modelling time to event with observations made at arbitrary times

AU - Sperrin, Matthew

AU - Buchan, Iain

PY - 2013/1/15

Y1 - 2013/1/15

N2 - We introduce new methods of analysing time to event data via extended versions of the proportional hazards and accelerated failure time (AFT) models. In many time to event studies, the time of first observation is arbitrary, in the sense that no risk modifying event occurs. This is particularly common in epidemiological studies. We show formally that, in these situations, it is not sensible to take the first observation as the time origin, either in AFT or proportional hazards type models. Instead, we advocate using age of the subject as the time scale. We account for the fact that baseline observations may be made at different ages in different patients via a two stage procedure. First, we marginally regress any potentially age-varying covariates against age, retaining the residuals. These residuals are then used as covariates in the fitting of either an AFT model or a proportional hazards model. We call the procedures residual accelerated failure time (RAFT) regression and residual proportional hazards (RPH) regression respectively. We compare standard AFT with RAFT, and demonstrate superior predictive ability of RAFT in real examples. In epidemiology, this has real implications in terms of risk communication to both patients and policy makers.

AB - We introduce new methods of analysing time to event data via extended versions of the proportional hazards and accelerated failure time (AFT) models. In many time to event studies, the time of first observation is arbitrary, in the sense that no risk modifying event occurs. This is particularly common in epidemiological studies. We show formally that, in these situations, it is not sensible to take the first observation as the time origin, either in AFT or proportional hazards type models. Instead, we advocate using age of the subject as the time scale. We account for the fact that baseline observations may be made at different ages in different patients via a two stage procedure. First, we marginally regress any potentially age-varying covariates against age, retaining the residuals. These residuals are then used as covariates in the fitting of either an AFT model or a proportional hazards model. We call the procedures residual accelerated failure time (RAFT) regression and residual proportional hazards (RPH) regression respectively. We compare standard AFT with RAFT, and demonstrate superior predictive ability of RAFT in real examples. In epidemiology, this has real implications in terms of risk communication to both patients and policy makers.

KW - accelerated failure time

KW - proportional hazards

KW - survival analysis

KW - time origin

KW - time-varying covariates

U2 - 10.1002/sim.5509

DO - 10.1002/sim.5509

M3 - Journal article

VL - 32

SP - 99

EP - 109

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 1097-0258

IS - 1

ER -