Dependent censoring in piecewise exponential survival models

Staplin, Natalie; Kimber, A. C.; Collett, David; Roderick, PJ

doi:10.1177/0962280214544018

Cited by 29 publications

(17 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Further work, following perhaps on that of Staplin et al . (30), is needed to identify and address potential bias due to dependent censoring in analyses of risk based on Poisson regression of person-year data.…”

Section: Discussionmentioning

confidence: 99%

Effect of Heterogeneity in Background Incidence on Inference about the Solid-Cancer Radiation Dose Response in Atomic Bomb Survivors

Cologne¹,

Sugiyama

French³

et al. 2019

Radiation Research

View full text Add to dashboard Cite

A recent analysis of solid cancer incidence in the Life Span Study of atomic bomb survivors (Hiroshima and Nagasaki, Japan) found evidence of a nonlinear, upwardly curving radiation dose response among males but not among females. Further analysis of this new and unexpected finding was necessary. We used two approaches to investigate this finding. In one approach, we excluded individual cancer sites or groups of sites from all solid cancers. In the other approach, we used joint analysis to allow for heterogeneity in background-rate parameters across groups of cancers with dissimilar trends in background rates. Exclusion of a few sites led to the disappearance of curvature among males in the remaining collection of solid cancers; some of these influential sites have unique features in their background age-specific incidence that are not captured by a background-rate model fit to all solid cancers combined. Exclusion of a few sites also led to an appearance of curvature among females. Misspecification of background rates can cause bias in inference about the shape of the dose response, so heterogeneity of background rates might explain at least part of the all solid cancer dose-response difference in curvature between males and females. We conclude that analysis based on all solid cancers as a single outcome is not the optimal method to assess radiation risk for solid cancer in the Life Span Study; joint analysis with suitable choices of cancer groups might be preferable by allowing for background-rate heterogeneity across sites while providing greater power to assess radiation risk than analyses of individual sites.

show abstract

Section: Discussionmentioning

confidence: 99%

Effect of Heterogeneity in Background Incidence on Inference about the Solid-Cancer Radiation Dose Response in Atomic Bomb Survivors

Cologne¹,

Sugiyama

French³

et al. 2019

Radiation Research

View full text Add to dashboard Cite

show abstract

“…Independent censoring is known to hold if T is stochastically independent of C [3] but the assumption is generally not testable based on the observed data [21, 22]. However, for a study with staggered entry and administrative end of follow-up, independence of the administrative part of the censoring is equivalent to the absence of calendar time trends which is in fact testable.…”

Section: Methodsmentioning

confidence: 99%

“…Intuitively, it may be thought of as the requirement that subjects who remain at risk are representative for the sample without censoring with respect to their disease experience at any given time. In a general time-to-event setting it is however not possible, based on the observed data, to determine whether a censoring mechanism is independent or not [21, 22]. …”

Section: Introductionmentioning

confidence: 99%

Estimating a population cumulative incidence under calendar time trends

Hansen

Overgaard

Andersen

et al. 2017

BMC Med Res Methodol

View full text Add to dashboard Cite

BackgroundThe risk of a disease or psychiatric disorder is frequently measured by the age-specific cumulative incidence. Cumulative incidence estimates are often derived in cohort studies with individuals recruited over calendar time and with the end of follow-up governed by a specific date. It is common practice to apply the Kaplan–Meier or Aalen–Johansen estimator to the total sample and report either the estimated cumulative incidence curve or just a single point on the curve as a description of the disease risk.MethodsWe argue that, whenever the disease or disorder of interest is influenced by calendar time trends, the total sample Kaplan–Meier and Aalen–Johansen estimators do not provide useful estimates of the general risk in the target population. We present some alternatives to this type of analysis.ResultsWe show how a proportional hazards model may be used to extrapolate disease risk estimates if proportionality is a reasonable assumption. If not reasonable, we instead advocate that a more useful description of the disease risk lies in the age-specific cumulative incidence curves across strata given by time of entry or perhaps just the end of follow-up estimates across all strata. Finally, we argue that a weighted average of these end of follow-up estimates may be a useful summary measure of the disease risk within the study period.ConclusionsTime trends in a disease risk will render total sample estimators less useful in observational studies with staggered entry and administrative censoring. An analysis based on proportional hazards or a stratified analysis may be better alternatives.Electronic supplementary materialThe online version of this article (doi:10.1186/s12874-016-0280-6) contains supplementary material, which is available to authorized users.

show abstract

“…Emura and Chen 21 ; Staplin et al. 22 ; Willems et al. 23 ) but has not caught up in the tree-based methods literature.…”

Section: Introductionmentioning

confidence: 99%

Survival forests for data with dependent censoring

Moradian

Larocque

Bellavance

2017

Stat Methods Med Res

View full text Add to dashboard Cite

Tree-based methods are very powerful and popular tools for analysing survival data with right-censoring. The existing methods assume that the true time-to-event and the censoring times are independent given the covariates. We propose different ways to build survival forests when dependent censoring is suspected, by using an appropriate estimator of the survival function when aggregating the individual trees and/or by modifying the splitting rule. The appropriate estimator used in this paper is the copula-graphic estimator. We also propose a new method for building survival forests, called p-forest, that may be used not only when dependent censoring is suspected, but also as a new survival forest method in general. The results from a simulation study indicate that these modifications improve greatly the estimation of the survival function in situations of dependent censoring. A real data example illustrates how the proposed methods can be used to perform a sensitivity analysis.

show abstract

Dependent censoring in piecewise exponential survival models

Cited by 29 publications

References 18 publications

Effect of Heterogeneity in Background Incidence on Inference about the Solid-Cancer Radiation Dose Response in Atomic Bomb Survivors

Effect of Heterogeneity in Background Incidence on Inference about the Solid-Cancer Radiation Dose Response in Atomic Bomb Survivors

Estimating a population cumulative incidence under calendar time trends

Survival forests for data with dependent censoring

Contact Info

Product

Resources

About