Threshold regression to accommodate a censored covariate

Jing Qian, Sy Han Chiou, Jacqueline E. Maye, Folefac Atem, Keith A. Johnson, Rebecca Betensky

Research output: Contribution to journalArticle

Abstract

In several common study designs, regression modeling is complicated by the presence of censored covariates. Examples of such covariates include maternal age of onset of dementia that may be right censored in an Alzheimer's amyloid imaging study of healthy subjects, metabolite measurements that are subject to limit of detection censoring in a case-control study of cardiovascular disease, and progressive biomarkers whose baseline values are of interest, but are measured post-baseline in longitudinal neuropsychological studies of Alzheimer's disease. We propose threshold regression approaches for linear regression models with a covariate that is subject to random censoring. Threshold regression methods allow for immediate testing of the significance of the effect of a censored covariate. In addition, they provide for unbiased estimation of the regression coefficient of the censored covariate. We derive the asymptotic properties of the resulting estimators under mild regularity conditions. Simulations demonstrate that the proposed estimators have good finite-sample performance, and often offer improved efficiency over existing methods. We also derive a principled method for selection of the threshold. We illustrate the approach in application to an Alzheimer's disease study that investigated brain amyloid levels in older individuals, as measured through positron emission tomography scans, as a function of maternal age of dementia onset, with adjustment for other covariates. We have developed an R package, censCov, for implementation of our method, available at CRAN.

Original languageEnglish (US)
Pages (from-to)1261-1270
Number of pages10
JournalBiometrics
Volume74
Issue number4
DOIs
StatePublished - Dec 1 2018

Fingerprint

Covariates
dementia
Regression
amyloid
Alzheimer disease
Amyloid
Maternal Age
Age of Onset
Dementia
positron-emission tomography
Linear Models
Alzheimer Disease
Positron emission tomography
Alzheimer's Disease
Biomarkers
selection methods
longitudinal studies
Metabolites
case-control studies
Linear regression

Keywords

  • Alzheimer's disease
  • Bias correction
  • Censored predictor
  • Cox proportional hazards model
  • Kaplan–Meier estimator
  • Limit of detection

ASJC Scopus subject areas

  • Statistics and Probability
  • Biochemistry, Genetics and Molecular Biology(all)
  • Immunology and Microbiology(all)
  • Agricultural and Biological Sciences(all)
  • Applied Mathematics

Cite this

Qian, J., Chiou, S. H., Maye, J. E., Atem, F., Johnson, K. A., & Betensky, R. (2018). Threshold regression to accommodate a censored covariate. Biometrics, 74(4), 1261-1270. https://doi.org/10.1111/biom.12922

Threshold regression to accommodate a censored covariate. / Qian, Jing; Chiou, Sy Han; Maye, Jacqueline E.; Atem, Folefac; Johnson, Keith A.; Betensky, Rebecca.

In: Biometrics, Vol. 74, No. 4, 01.12.2018, p. 1261-1270.

Research output: Contribution to journalArticle

Qian, J, Chiou, SH, Maye, JE, Atem, F, Johnson, KA & Betensky, R 2018, 'Threshold regression to accommodate a censored covariate', Biometrics, vol. 74, no. 4, pp. 1261-1270. https://doi.org/10.1111/biom.12922
Qian J, Chiou SH, Maye JE, Atem F, Johnson KA, Betensky R. Threshold regression to accommodate a censored covariate. Biometrics. 2018 Dec 1;74(4):1261-1270. https://doi.org/10.1111/biom.12922
Qian, Jing ; Chiou, Sy Han ; Maye, Jacqueline E. ; Atem, Folefac ; Johnson, Keith A. ; Betensky, Rebecca. / Threshold regression to accommodate a censored covariate. In: Biometrics. 2018 ; Vol. 74, No. 4. pp. 1261-1270.
@article{8b0b4eaad8e8466d86fb13472eac4b36,
title = "Threshold regression to accommodate a censored covariate",
abstract = "In several common study designs, regression modeling is complicated by the presence of censored covariates. Examples of such covariates include maternal age of onset of dementia that may be right censored in an Alzheimer's amyloid imaging study of healthy subjects, metabolite measurements that are subject to limit of detection censoring in a case-control study of cardiovascular disease, and progressive biomarkers whose baseline values are of interest, but are measured post-baseline in longitudinal neuropsychological studies of Alzheimer's disease. We propose threshold regression approaches for linear regression models with a covariate that is subject to random censoring. Threshold regression methods allow for immediate testing of the significance of the effect of a censored covariate. In addition, they provide for unbiased estimation of the regression coefficient of the censored covariate. We derive the asymptotic properties of the resulting estimators under mild regularity conditions. Simulations demonstrate that the proposed estimators have good finite-sample performance, and often offer improved efficiency over existing methods. We also derive a principled method for selection of the threshold. We illustrate the approach in application to an Alzheimer's disease study that investigated brain amyloid levels in older individuals, as measured through positron emission tomography scans, as a function of maternal age of dementia onset, with adjustment for other covariates. We have developed an R package, censCov, for implementation of our method, available at CRAN.",
keywords = "Alzheimer's disease, Bias correction, Censored predictor, Cox proportional hazards model, Kaplan–Meier estimator, Limit of detection",
author = "Jing Qian and Chiou, {Sy Han} and Maye, {Jacqueline E.} and Folefac Atem and Johnson, {Keith A.} and Rebecca Betensky",
year = "2018",
month = "12",
day = "1",
doi = "10.1111/biom.12922",
language = "English (US)",
volume = "74",
pages = "1261--1270",
journal = "Biometrics",
issn = "0006-341X",
publisher = "Wiley-Blackwell",
number = "4",

}

TY - JOUR

T1 - Threshold regression to accommodate a censored covariate

AU - Qian, Jing

AU - Chiou, Sy Han

AU - Maye, Jacqueline E.

AU - Atem, Folefac

AU - Johnson, Keith A.

AU - Betensky, Rebecca

PY - 2018/12/1

Y1 - 2018/12/1

N2 - In several common study designs, regression modeling is complicated by the presence of censored covariates. Examples of such covariates include maternal age of onset of dementia that may be right censored in an Alzheimer's amyloid imaging study of healthy subjects, metabolite measurements that are subject to limit of detection censoring in a case-control study of cardiovascular disease, and progressive biomarkers whose baseline values are of interest, but are measured post-baseline in longitudinal neuropsychological studies of Alzheimer's disease. We propose threshold regression approaches for linear regression models with a covariate that is subject to random censoring. Threshold regression methods allow for immediate testing of the significance of the effect of a censored covariate. In addition, they provide for unbiased estimation of the regression coefficient of the censored covariate. We derive the asymptotic properties of the resulting estimators under mild regularity conditions. Simulations demonstrate that the proposed estimators have good finite-sample performance, and often offer improved efficiency over existing methods. We also derive a principled method for selection of the threshold. We illustrate the approach in application to an Alzheimer's disease study that investigated brain amyloid levels in older individuals, as measured through positron emission tomography scans, as a function of maternal age of dementia onset, with adjustment for other covariates. We have developed an R package, censCov, for implementation of our method, available at CRAN.

AB - In several common study designs, regression modeling is complicated by the presence of censored covariates. Examples of such covariates include maternal age of onset of dementia that may be right censored in an Alzheimer's amyloid imaging study of healthy subjects, metabolite measurements that are subject to limit of detection censoring in a case-control study of cardiovascular disease, and progressive biomarkers whose baseline values are of interest, but are measured post-baseline in longitudinal neuropsychological studies of Alzheimer's disease. We propose threshold regression approaches for linear regression models with a covariate that is subject to random censoring. Threshold regression methods allow for immediate testing of the significance of the effect of a censored covariate. In addition, they provide for unbiased estimation of the regression coefficient of the censored covariate. We derive the asymptotic properties of the resulting estimators under mild regularity conditions. Simulations demonstrate that the proposed estimators have good finite-sample performance, and often offer improved efficiency over existing methods. We also derive a principled method for selection of the threshold. We illustrate the approach in application to an Alzheimer's disease study that investigated brain amyloid levels in older individuals, as measured through positron emission tomography scans, as a function of maternal age of dementia onset, with adjustment for other covariates. We have developed an R package, censCov, for implementation of our method, available at CRAN.

KW - Alzheimer's disease

KW - Bias correction

KW - Censored predictor

KW - Cox proportional hazards model

KW - Kaplan–Meier estimator

KW - Limit of detection

UR - http://www.scopus.com/inward/record.url?scp=85060986750&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85060986750&partnerID=8YFLogxK

U2 - 10.1111/biom.12922

DO - 10.1111/biom.12922

M3 - Article

C2 - 29933515

AN - SCOPUS:85060986750

VL - 74

SP - 1261

EP - 1270

JO - Biometrics

JF - Biometrics

SN - 0006-341X

IS - 4

ER -