A Bayesian hierarchical model for prediction of latent health states from multiple data sources with application to active surveillance of prostate cancer

Rebecca Yates Coley, Aaron J. Fisher, Mufaddal Mamawala, Herbert Ballentine Carter, Kenneth J. Pienta, Scott L. Zeger

Research output: Research - peer-reviewArticle

Abstract

In this article, we present a Bayesian hierarchical model for predicting a latent health state from longitudinal clinical measurements. Model development is motivated by the need to integrate multiple sources of data to improve clinical decisions about whether to remove or irradiate a patient's prostate cancer. Existing modeling approaches are extended to accommodate measurement error in cancer state determinations based on biopsied tissue, clinical measurements possibly not missing at random, and informative partial observation of the true state. The proposed model enables estimation of whether an individual's underlying prostate cancer is aggressive, requiring surgery and/or radiation, or indolent, permitting continued surveillance. These individualized predictions can then be communicated to clinicians and patients to inform decision-making. We demonstrate the model with data from a cohort of low-risk prostate cancer patients at Johns Hopkins University and assess predictive accuracy among a subset for whom true cancer state is observed. Simulation studies confirm model performance and explore the impact of adjusting for informative missingness on true state predictions. R code is provided in an online supplement and at http://github.com/rycoley/prediction-prostate-surveillance.

LanguageEnglish (US)
Pages625-634
Number of pages10
JournalBiometrics
Volume73
Issue number2
DOIs
StatePublished - Jun 1 2017

Fingerprint

Bayesian Hierarchical Model
Prostate Cancer
Surveillance
Health
Prediction
Model
Information Storage and Retrieval
Prostatic Neoplasms
prostatic neoplasms
prediction
monitoring
Cancer
Neoplasms
Partial Observation
Missing at Random
Performance Model
Measurement Error
Surgery
Decision Making
Integrate

Keywords

  • Latent class analysis
  • Missing data
  • Precision medicine
  • Prostate cancer prognosis
  • Risk classification

ASJC Scopus subject areas

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

Cite this

A Bayesian hierarchical model for prediction of latent health states from multiple data sources with application to active surveillance of prostate cancer. / Coley, Rebecca Yates; Fisher, Aaron J.; Mamawala, Mufaddal; Carter, Herbert Ballentine; Pienta, Kenneth J.; Zeger, Scott L.

In: Biometrics, Vol. 73, No. 2, 01.06.2017, p. 625-634.

Research output: Research - peer-reviewArticle

@article{5888f235033f4853947e0382bf00ef54,
title = "A Bayesian hierarchical model for prediction of latent health states from multiple data sources with application to active surveillance of prostate cancer",
abstract = "In this article, we present a Bayesian hierarchical model for predicting a latent health state from longitudinal clinical measurements. Model development is motivated by the need to integrate multiple sources of data to improve clinical decisions about whether to remove or irradiate a patient's prostate cancer. Existing modeling approaches are extended to accommodate measurement error in cancer state determinations based on biopsied tissue, clinical measurements possibly not missing at random, and informative partial observation of the true state. The proposed model enables estimation of whether an individual's underlying prostate cancer is aggressive, requiring surgery and/or radiation, or indolent, permitting continued surveillance. These individualized predictions can then be communicated to clinicians and patients to inform decision-making. We demonstrate the model with data from a cohort of low-risk prostate cancer patients at Johns Hopkins University and assess predictive accuracy among a subset for whom true cancer state is observed. Simulation studies confirm model performance and explore the impact of adjusting for informative missingness on true state predictions. R code is provided in an online supplement and at http://github.com/rycoley/prediction-prostate-surveillance.",
keywords = "Latent class analysis, Missing data, Precision medicine, Prostate cancer prognosis, Risk classification",
author = "Coley, {Rebecca Yates} and Fisher, {Aaron J.} and Mufaddal Mamawala and Carter, {Herbert Ballentine} and Pienta, {Kenneth J.} and Zeger, {Scott L.}",
year = "2017",
month = "6",
doi = "10.1111/biom.12577",
volume = "73",
pages = "625--634",
journal = "Biometrics",
issn = "0006-341X",
publisher = "Wiley-Blackwell",
number = "2",

}

TY - JOUR

T1 - A Bayesian hierarchical model for prediction of latent health states from multiple data sources with application to active surveillance of prostate cancer

AU - Coley,Rebecca Yates

AU - Fisher,Aaron J.

AU - Mamawala,Mufaddal

AU - Carter,Herbert Ballentine

AU - Pienta,Kenneth J.

AU - Zeger,Scott L.

PY - 2017/6/1

Y1 - 2017/6/1

N2 - In this article, we present a Bayesian hierarchical model for predicting a latent health state from longitudinal clinical measurements. Model development is motivated by the need to integrate multiple sources of data to improve clinical decisions about whether to remove or irradiate a patient's prostate cancer. Existing modeling approaches are extended to accommodate measurement error in cancer state determinations based on biopsied tissue, clinical measurements possibly not missing at random, and informative partial observation of the true state. The proposed model enables estimation of whether an individual's underlying prostate cancer is aggressive, requiring surgery and/or radiation, or indolent, permitting continued surveillance. These individualized predictions can then be communicated to clinicians and patients to inform decision-making. We demonstrate the model with data from a cohort of low-risk prostate cancer patients at Johns Hopkins University and assess predictive accuracy among a subset for whom true cancer state is observed. Simulation studies confirm model performance and explore the impact of adjusting for informative missingness on true state predictions. R code is provided in an online supplement and at http://github.com/rycoley/prediction-prostate-surveillance.

AB - In this article, we present a Bayesian hierarchical model for predicting a latent health state from longitudinal clinical measurements. Model development is motivated by the need to integrate multiple sources of data to improve clinical decisions about whether to remove or irradiate a patient's prostate cancer. Existing modeling approaches are extended to accommodate measurement error in cancer state determinations based on biopsied tissue, clinical measurements possibly not missing at random, and informative partial observation of the true state. The proposed model enables estimation of whether an individual's underlying prostate cancer is aggressive, requiring surgery and/or radiation, or indolent, permitting continued surveillance. These individualized predictions can then be communicated to clinicians and patients to inform decision-making. We demonstrate the model with data from a cohort of low-risk prostate cancer patients at Johns Hopkins University and assess predictive accuracy among a subset for whom true cancer state is observed. Simulation studies confirm model performance and explore the impact of adjusting for informative missingness on true state predictions. R code is provided in an online supplement and at http://github.com/rycoley/prediction-prostate-surveillance.

KW - Latent class analysis

KW - Missing data

KW - Precision medicine

KW - Prostate cancer prognosis

KW - Risk classification

UR - http://www.scopus.com/inward/record.url?scp=85020776734&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85020776734&partnerID=8YFLogxK

U2 - 10.1111/biom.12577

DO - 10.1111/biom.12577

M3 - Article

VL - 73

SP - 625

EP - 634

JO - Biometrics

T2 - Biometrics

JF - Biometrics

SN - 0006-341X

IS - 2

ER -