An intuitive Bayesian spatial model for disease mapping that accounts for scaling

Andrea Riebler, Sigrunn H. Sørbye, Daniel Simpson, Håvard Rue

Research output: Contribution to journalArticle

35 Citations (Scopus)

Abstract

In recent years, disease mapping studies have become a routine application within geographical epidemiology and are typically analysed within a Bayesian hierarchical model formulation. A variety of model formulations for the latent level have been proposed but all come with inherent issues. In the classical BYM model, the spatially structured component cannot be seen independently from the unstructured component. This makes prior definitions for the hyperparameters of the two random effects challenging. There are alternative model formulations that address this confounding, however, the issue on how to choose interpretable hyperpriors is still unsolved. Here, we discuss a recently proposed parameterisation of the BYM model that leads to improved parameter control as the hyperparameters can be seen independently from each other. Furthermore, the need for a scaled spatial component is addressed, which facilitates assignment of interpretable hyperpriors and make these transferable between spatial applications with different graph structures. We provide implementation details for the new model formulation which preserve sparsity properties, and we investigate systematically the model performance and compare it to existing parameterisations. Through a simulation study, we show that the new model performs well, both showing good learning abilities and good shrinkage behaviour. In terms of model choice criteria, the proposed model performs at least equally well as existing parameterisations, but only the new formulation offers parameters that are interpretable and hyperpriors that have a clear meaning.
Original languageEnglish
Pages (from-to)1145-1165
JournalStatistical Methods in Medical Research
Volume25
Issue number4
DOIs
Publication statusPublished - 1 Aug 2016

Fingerprint

Disease Mapping
Aptitude
Spatial Model
Bayesian Model
Intuitive
Epidemiology
Learning
Scaling
Formulation
Parameterization
Hyperparameters
Model
Model Choice
Bayesian Hierarchical Model
Confounding
Performance Model
Shrinkage
Random Effects
Sparsity
Control Parameter

Keywords

  • stat.ME

Cite this

An intuitive Bayesian spatial model for disease mapping that accounts for scaling. / Riebler, Andrea; Sørbye, Sigrunn H.; Simpson, Daniel; Rue, Håvard.

In: Statistical Methods in Medical Research, Vol. 25, No. 4, 01.08.2016, p. 1145-1165.

Research output: Contribution to journalArticle

Riebler, Andrea ; Sørbye, Sigrunn H. ; Simpson, Daniel ; Rue, Håvard. / An intuitive Bayesian spatial model for disease mapping that accounts for scaling. In: Statistical Methods in Medical Research. 2016 ; Vol. 25, No. 4. pp. 1145-1165.
@article{99f82654555c44f1a99a07e51a5fcfe4,
title = "An intuitive Bayesian spatial model for disease mapping that accounts for scaling",
abstract = "In recent years, disease mapping studies have become a routine application within geographical epidemiology and are typically analysed within a Bayesian hierarchical model formulation. A variety of model formulations for the latent level have been proposed but all come with inherent issues. In the classical BYM model, the spatially structured component cannot be seen independently from the unstructured component. This makes prior definitions for the hyperparameters of the two random effects challenging. There are alternative model formulations that address this confounding, however, the issue on how to choose interpretable hyperpriors is still unsolved. Here, we discuss a recently proposed parameterisation of the BYM model that leads to improved parameter control as the hyperparameters can be seen independently from each other. Furthermore, the need for a scaled spatial component is addressed, which facilitates assignment of interpretable hyperpriors and make these transferable between spatial applications with different graph structures. We provide implementation details for the new model formulation which preserve sparsity properties, and we investigate systematically the model performance and compare it to existing parameterisations. Through a simulation study, we show that the new model performs well, both showing good learning abilities and good shrinkage behaviour. In terms of model choice criteria, the proposed model performs at least equally well as existing parameterisations, but only the new formulation offers parameters that are interpretable and hyperpriors that have a clear meaning.",
keywords = "stat.ME",
author = "Andrea Riebler and S{\o}rbye, {Sigrunn H.} and Daniel Simpson and H{\aa}vard Rue",
year = "2016",
month = "8",
day = "1",
doi = "10.1177/0962280216660421",
language = "English",
volume = "25",
pages = "1145--1165",
journal = "Statistical Methods in Medical Research",
issn = "0962-2802",
publisher = "Sage Publications",
number = "4",

}

TY - JOUR

T1 - An intuitive Bayesian spatial model for disease mapping that accounts for scaling

AU - Riebler, Andrea

AU - Sørbye, Sigrunn H.

AU - Simpson, Daniel

AU - Rue, Håvard

PY - 2016/8/1

Y1 - 2016/8/1

N2 - In recent years, disease mapping studies have become a routine application within geographical epidemiology and are typically analysed within a Bayesian hierarchical model formulation. A variety of model formulations for the latent level have been proposed but all come with inherent issues. In the classical BYM model, the spatially structured component cannot be seen independently from the unstructured component. This makes prior definitions for the hyperparameters of the two random effects challenging. There are alternative model formulations that address this confounding, however, the issue on how to choose interpretable hyperpriors is still unsolved. Here, we discuss a recently proposed parameterisation of the BYM model that leads to improved parameter control as the hyperparameters can be seen independently from each other. Furthermore, the need for a scaled spatial component is addressed, which facilitates assignment of interpretable hyperpriors and make these transferable between spatial applications with different graph structures. We provide implementation details for the new model formulation which preserve sparsity properties, and we investigate systematically the model performance and compare it to existing parameterisations. Through a simulation study, we show that the new model performs well, both showing good learning abilities and good shrinkage behaviour. In terms of model choice criteria, the proposed model performs at least equally well as existing parameterisations, but only the new formulation offers parameters that are interpretable and hyperpriors that have a clear meaning.

AB - In recent years, disease mapping studies have become a routine application within geographical epidemiology and are typically analysed within a Bayesian hierarchical model formulation. A variety of model formulations for the latent level have been proposed but all come with inherent issues. In the classical BYM model, the spatially structured component cannot be seen independently from the unstructured component. This makes prior definitions for the hyperparameters of the two random effects challenging. There are alternative model formulations that address this confounding, however, the issue on how to choose interpretable hyperpriors is still unsolved. Here, we discuss a recently proposed parameterisation of the BYM model that leads to improved parameter control as the hyperparameters can be seen independently from each other. Furthermore, the need for a scaled spatial component is addressed, which facilitates assignment of interpretable hyperpriors and make these transferable between spatial applications with different graph structures. We provide implementation details for the new model formulation which preserve sparsity properties, and we investigate systematically the model performance and compare it to existing parameterisations. Through a simulation study, we show that the new model performs well, both showing good learning abilities and good shrinkage behaviour. In terms of model choice criteria, the proposed model performs at least equally well as existing parameterisations, but only the new formulation offers parameters that are interpretable and hyperpriors that have a clear meaning.

KW - stat.ME

UR - http://dx.doi.org/10.1177/0962280216660421

U2 - 10.1177/0962280216660421

DO - 10.1177/0962280216660421

M3 - Article

VL - 25

SP - 1145

EP - 1165

JO - Statistical Methods in Medical Research

JF - Statistical Methods in Medical Research

SN - 0962-2802

IS - 4

ER -