A statistical downscaling approach for generating high spatial resolution health risk maps

A case study of road noise and ischemic heart disease mortality in Melbourne, Australia

Ivan C. Hanigan, Timothy B. Chaston, Ben Hinze, Martine Dennekamp, Bin Jalaludin, Yohannes Kinfu, Geoffrey G. Morgan

Research output: Contribution to journalArticle

1 Downloads (Pure)

Abstract

Introduction: Road traffic noise increases the risk of mortality from ischemic heart disease (IHD). Because noise is highly localized, high resolution maps of exposures and health outcomes are key to urban planning interventions that are informed by health risks. In Australia, publicly accessible IHD deaths data are only available at the coarse spatial aggregation level of local government area (LGA), in which about 130,000 people reside. Herein, we addressed this limitation of health data using statistical downscaling and generated environmental health risk maps for noise at the meshblock level (MB; ~ 90 people). Methods: We estimated noise exposures at the MB level using a model of road traffic noise in Melbourne, Australia, from 2011. As recommended by the World Health Organization, a non-linear exposure-response function for traffic noise and IHD was used to calculate odds ratios for noise related IHD in all MBs. Noise attributable risks of IHD death were then estimated by statistically downscaling LGA-level IHD rates to the MB level. Results: Noise levels of 80 dB were recorded in some MBs. From the given noise maps, approximately 5% of the population was exposed to traffic noise above the risk threshold of 55 dB. Maps of excess risk at the MB level identified areas in which noise levels and exposed populations are large. Attributable rates of IHD deaths due to noise were generally very low, but some were as high as 5-10 per 100,000, and in extremely noisy and populated MBs represented more than 8% excess risk of IHD death. We presented results as interactive maps of excess risk due to noise at the small neighbourhood scale. Conclusion: Our method accommodates low-resolution health data and could be used to inform urban planning and public health decision making for various environmental health concerns. Estimated noise related IHD deaths were relatively few in Melbourne in 2011, likely because road traffic is one of many noise sources and the current noise model underestimates exposures. Nonetheless, this novel computational framework could be used globally to generate maps of noise related health risks using scant health outcomes data.

Original languageEnglish
Article number20
Pages (from-to)20
Number of pages10
JournalInternational Journal of Health Geographics
Volume18
Issue number1
DOIs
Publication statusPublished - 5 Sep 2019

Fingerprint

Health risks
Myocardial Ischemia
Noise
Mortality
Health
Urban planning
City Planning
Health risk
Ischemic heart disease
Roads
Local Government
Environmental Health
Public health
Agglomeration
Decision making
Urban Health

Cite this

@article{84e3e8632988460388a227376db65c8a,
title = "A statistical downscaling approach for generating high spatial resolution health risk maps: A case study of road noise and ischemic heart disease mortality in Melbourne, Australia",
abstract = "Introduction: Road traffic noise increases the risk of mortality from ischemic heart disease (IHD). Because noise is highly localized, high resolution maps of exposures and health outcomes are key to urban planning interventions that are informed by health risks. In Australia, publicly accessible IHD deaths data are only available at the coarse spatial aggregation level of local government area (LGA), in which about 130,000 people reside. Herein, we addressed this limitation of health data using statistical downscaling and generated environmental health risk maps for noise at the meshblock level (MB; ~ 90 people). Methods: We estimated noise exposures at the MB level using a model of road traffic noise in Melbourne, Australia, from 2011. As recommended by the World Health Organization, a non-linear exposure-response function for traffic noise and IHD was used to calculate odds ratios for noise related IHD in all MBs. Noise attributable risks of IHD death were then estimated by statistically downscaling LGA-level IHD rates to the MB level. Results: Noise levels of 80 dB were recorded in some MBs. From the given noise maps, approximately 5{\%} of the population was exposed to traffic noise above the risk threshold of 55 dB. Maps of excess risk at the MB level identified areas in which noise levels and exposed populations are large. Attributable rates of IHD deaths due to noise were generally very low, but some were as high as 5-10 per 100,000, and in extremely noisy and populated MBs represented more than 8{\%} excess risk of IHD death. We presented results as interactive maps of excess risk due to noise at the small neighbourhood scale. Conclusion: Our method accommodates low-resolution health data and could be used to inform urban planning and public health decision making for various environmental health concerns. Estimated noise related IHD deaths were relatively few in Melbourne in 2011, likely because road traffic is one of many noise sources and the current noise model underestimates exposures. Nonetheless, this novel computational framework could be used globally to generate maps of noise related health risks using scant health outcomes data.",
keywords = "Environmental health risk assessment, Exposure–response function, Mortality, Noise, Statistical downscaling, Exposure-response function",
author = "Hanigan, {Ivan C.} and Chaston, {Timothy B.} and Ben Hinze and Martine Dennekamp and Bin Jalaludin and Yohannes Kinfu and Morgan, {Geoffrey G.}",
year = "2019",
month = "9",
day = "5",
doi = "10.1186/s12942-019-0184-x",
language = "English",
volume = "18",
pages = "20",
journal = "International Journal of Health Geographics",
issn = "1476-072X",
publisher = "BioMed Central",
number = "1",

}

A statistical downscaling approach for generating high spatial resolution health risk maps : A case study of road noise and ischemic heart disease mortality in Melbourne, Australia. / Hanigan, Ivan C.; Chaston, Timothy B.; Hinze, Ben; Dennekamp, Martine; Jalaludin, Bin; Kinfu, Yohannes; Morgan, Geoffrey G.

In: International Journal of Health Geographics, Vol. 18, No. 1, 20, 05.09.2019, p. 20.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A statistical downscaling approach for generating high spatial resolution health risk maps

T2 - A case study of road noise and ischemic heart disease mortality in Melbourne, Australia

AU - Hanigan, Ivan C.

AU - Chaston, Timothy B.

AU - Hinze, Ben

AU - Dennekamp, Martine

AU - Jalaludin, Bin

AU - Kinfu, Yohannes

AU - Morgan, Geoffrey G.

PY - 2019/9/5

Y1 - 2019/9/5

N2 - Introduction: Road traffic noise increases the risk of mortality from ischemic heart disease (IHD). Because noise is highly localized, high resolution maps of exposures and health outcomes are key to urban planning interventions that are informed by health risks. In Australia, publicly accessible IHD deaths data are only available at the coarse spatial aggregation level of local government area (LGA), in which about 130,000 people reside. Herein, we addressed this limitation of health data using statistical downscaling and generated environmental health risk maps for noise at the meshblock level (MB; ~ 90 people). Methods: We estimated noise exposures at the MB level using a model of road traffic noise in Melbourne, Australia, from 2011. As recommended by the World Health Organization, a non-linear exposure-response function for traffic noise and IHD was used to calculate odds ratios for noise related IHD in all MBs. Noise attributable risks of IHD death were then estimated by statistically downscaling LGA-level IHD rates to the MB level. Results: Noise levels of 80 dB were recorded in some MBs. From the given noise maps, approximately 5% of the population was exposed to traffic noise above the risk threshold of 55 dB. Maps of excess risk at the MB level identified areas in which noise levels and exposed populations are large. Attributable rates of IHD deaths due to noise were generally very low, but some were as high as 5-10 per 100,000, and in extremely noisy and populated MBs represented more than 8% excess risk of IHD death. We presented results as interactive maps of excess risk due to noise at the small neighbourhood scale. Conclusion: Our method accommodates low-resolution health data and could be used to inform urban planning and public health decision making for various environmental health concerns. Estimated noise related IHD deaths were relatively few in Melbourne in 2011, likely because road traffic is one of many noise sources and the current noise model underestimates exposures. Nonetheless, this novel computational framework could be used globally to generate maps of noise related health risks using scant health outcomes data.

AB - Introduction: Road traffic noise increases the risk of mortality from ischemic heart disease (IHD). Because noise is highly localized, high resolution maps of exposures and health outcomes are key to urban planning interventions that are informed by health risks. In Australia, publicly accessible IHD deaths data are only available at the coarse spatial aggregation level of local government area (LGA), in which about 130,000 people reside. Herein, we addressed this limitation of health data using statistical downscaling and generated environmental health risk maps for noise at the meshblock level (MB; ~ 90 people). Methods: We estimated noise exposures at the MB level using a model of road traffic noise in Melbourne, Australia, from 2011. As recommended by the World Health Organization, a non-linear exposure-response function for traffic noise and IHD was used to calculate odds ratios for noise related IHD in all MBs. Noise attributable risks of IHD death were then estimated by statistically downscaling LGA-level IHD rates to the MB level. Results: Noise levels of 80 dB were recorded in some MBs. From the given noise maps, approximately 5% of the population was exposed to traffic noise above the risk threshold of 55 dB. Maps of excess risk at the MB level identified areas in which noise levels and exposed populations are large. Attributable rates of IHD deaths due to noise were generally very low, but some were as high as 5-10 per 100,000, and in extremely noisy and populated MBs represented more than 8% excess risk of IHD death. We presented results as interactive maps of excess risk due to noise at the small neighbourhood scale. Conclusion: Our method accommodates low-resolution health data and could be used to inform urban planning and public health decision making for various environmental health concerns. Estimated noise related IHD deaths were relatively few in Melbourne in 2011, likely because road traffic is one of many noise sources and the current noise model underestimates exposures. Nonetheless, this novel computational framework could be used globally to generate maps of noise related health risks using scant health outcomes data.

KW - Environmental health risk assessment

KW - Exposure–response function

KW - Mortality

KW - Noise

KW - Statistical downscaling

KW - Exposure-response function

UR - http://www.scopus.com/inward/record.url?scp=85071760244&partnerID=8YFLogxK

UR - http://www.mendeley.com/research/statistical-downscaling-approach-generating-high-spatial-resolution-health-risk-maps-case-study-road

U2 - 10.1186/s12942-019-0184-x

DO - 10.1186/s12942-019-0184-x

M3 - Article

VL - 18

SP - 20

JO - International Journal of Health Geographics

JF - International Journal of Health Geographics

SN - 1476-072X

IS - 1

M1 - 20

ER -