Comparing the performance of air pollution models for nitrogen dioxide and ozone in the context of a multilevel epidemiological analysis

Barbara K. Butland; Evangelia Samoli; Richard W. Atkinson; Benjamin Barratt; Sean D. Beevers; Nutthida Kitwiroon; Konstantina Dimakopoulou; Sophia Rodopoulou; Joel D. Schwartz; Klea Katsouyanni

doi:10.1097/EE9.0000000000000093

Comparing the performance of air pollution models for nitrogen dioxide and ozone in the context of a multilevel epidemiological analysis

Barbara K. Butland^*, Evangelia Samoli, Richard W. Atkinson, Benjamin Barratt, Sean D. Beevers, Nutthida Kitwiroon, Konstantina Dimakopoulou, Sophia Rodopoulou, Joel D. Schwartz, Klea Katsouyanni

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

17 Citations (Scopus)

Abstract

Background: Using modeled air pollutant predictions as exposure variables in epidemiological analyses can produce bias in health effect estimation. We used statistical simulation to estimate these biases and compare different air pollution models for London. Methods: Our simulations were based on a sample of 1,000 small geographical areas within London, United Kingdom. "True" pollutant data (daily mean nitrogen dioxide [NO₂] and ozone [O₃]) were simulated to include spatio-temporal variation and spatial covariance. All-cause mortality and cardiovascular hospital admissions were simulated from "true" pollution data using prespecified effect parameters for short and long-term exposure within a multilevel Poisson model. We compared: land use regression (LUR) models, dispersion models, LUR models including dispersion output as a spline (hybrid1), and generalized additive models combining splines in LUR and dispersion outputs (hybrid2). Validation datasets (model versus fixed-site monitor) were used to define simulation scenarios. Results: For the LUR models, bias estimates ranged from -56% to +7% for short-term exposure and -98% to -68% for long-term exposure and for the dispersion models from -33% to -15% and -52% to +0.5%, respectively. Hybrid1 provided little if any additional benefit, but hybrid2 appeared optimal in terms of bias estimates for short-term (-17% to +11%) and long-term (-28% to +11%) exposure and in preserving coverage probability and statistical power. Conclusions: Although exposure error can produce substantial negative bias (i.e., towards the null), combining outputs from different air pollution modeling approaches may reduce bias in health effect estimation leading to improved impact evaluation of abatement policies.

Original language	English
Article number	e093
Journal	Environmental Epidemiology
Volume	4
Issue number	3
DOIs	https://doi.org/10.1097/EE9.0000000000000093
Publication status	Published - 2020

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1097/EE9.0000000000000093

Cite this

Butland, B. K., Samoli, E., Atkinson, R. W., Barratt, B., Beevers, S. D., Kitwiroon, N., Dimakopoulou, K., Rodopoulou, S., Schwartz, J. D., & Katsouyanni, K. (2020). Comparing the performance of air pollution models for nitrogen dioxide and ozone in the context of a multilevel epidemiological analysis. Environmental Epidemiology, 4(3), Article e093. https://doi.org/10.1097/EE9.0000000000000093

@article{432e2f125f1c41b1a710e7ca744a991e,

title = "Comparing the performance of air pollution models for nitrogen dioxide and ozone in the context of a multilevel epidemiological analysis",

abstract = "Background: Using modeled air pollutant predictions as exposure variables in epidemiological analyses can produce bias in health effect estimation. We used statistical simulation to estimate these biases and compare different air pollution models for London. Methods: Our simulations were based on a sample of 1,000 small geographical areas within London, United Kingdom. {"}True{"} pollutant data (daily mean nitrogen dioxide [NO2] and ozone [O3]) were simulated to include spatio-temporal variation and spatial covariance. All-cause mortality and cardiovascular hospital admissions were simulated from {"}true{"} pollution data using prespecified effect parameters for short and long-term exposure within a multilevel Poisson model. We compared: land use regression (LUR) models, dispersion models, LUR models including dispersion output as a spline (hybrid1), and generalized additive models combining splines in LUR and dispersion outputs (hybrid2). Validation datasets (model versus fixed-site monitor) were used to define simulation scenarios. Results: For the LUR models, bias estimates ranged from -56% to +7% for short-term exposure and -98% to -68% for long-term exposure and for the dispersion models from -33% to -15% and -52% to +0.5%, respectively. Hybrid1 provided little if any additional benefit, but hybrid2 appeared optimal in terms of bias estimates for short-term (-17% to +11%) and long-term (-28% to +11%) exposure and in preserving coverage probability and statistical power. Conclusions: Although exposure error can produce substantial negative bias (i.e., towards the null), combining outputs from different air pollution modeling approaches may reduce bias in health effect estimation leading to improved impact evaluation of abatement policies.",

author = "Butland, {Barbara K.} and Evangelia Samoli and Atkinson, {Richard W.} and Benjamin Barratt and Beevers, {Sean D.} and Nutthida Kitwiroon and Konstantina Dimakopoulou and Sophia Rodopoulou and Schwartz, {Joel D.} and Klea Katsouyanni",

year = "2020",

doi = "10.1097/EE9.0000000000000093",

language = "English",

volume = "4",

journal = "Environmental Epidemiology",

issn = "2474-7882",

publisher = "Wolters Kluwer",

number = "3",

}

Butland, BK, Samoli, E, Atkinson, RW, Barratt, B , Beevers, SD, Kitwiroon, N, Dimakopoulou, K, Rodopoulou, S, Schwartz, JD & Katsouyanni, K 2020, 'Comparing the performance of air pollution models for nitrogen dioxide and ozone in the context of a multilevel epidemiological analysis', Environmental Epidemiology, vol. 4, no. 3, e093. https://doi.org/10.1097/EE9.0000000000000093

TY - JOUR

T1 - Comparing the performance of air pollution models for nitrogen dioxide and ozone in the context of a multilevel epidemiological analysis

AU - Butland, Barbara K.

AU - Samoli, Evangelia

AU - Atkinson, Richard W.

AU - Barratt, Benjamin

AU - Beevers, Sean D.

AU - Kitwiroon, Nutthida

AU - Dimakopoulou, Konstantina

AU - Rodopoulou, Sophia

AU - Schwartz, Joel D.

AU - Katsouyanni, Klea

PY - 2020

Y1 - 2020

N2 - Background: Using modeled air pollutant predictions as exposure variables in epidemiological analyses can produce bias in health effect estimation. We used statistical simulation to estimate these biases and compare different air pollution models for London. Methods: Our simulations were based on a sample of 1,000 small geographical areas within London, United Kingdom. "True" pollutant data (daily mean nitrogen dioxide [NO2] and ozone [O3]) were simulated to include spatio-temporal variation and spatial covariance. All-cause mortality and cardiovascular hospital admissions were simulated from "true" pollution data using prespecified effect parameters for short and long-term exposure within a multilevel Poisson model. We compared: land use regression (LUR) models, dispersion models, LUR models including dispersion output as a spline (hybrid1), and generalized additive models combining splines in LUR and dispersion outputs (hybrid2). Validation datasets (model versus fixed-site monitor) were used to define simulation scenarios. Results: For the LUR models, bias estimates ranged from -56% to +7% for short-term exposure and -98% to -68% for long-term exposure and for the dispersion models from -33% to -15% and -52% to +0.5%, respectively. Hybrid1 provided little if any additional benefit, but hybrid2 appeared optimal in terms of bias estimates for short-term (-17% to +11%) and long-term (-28% to +11%) exposure and in preserving coverage probability and statistical power. Conclusions: Although exposure error can produce substantial negative bias (i.e., towards the null), combining outputs from different air pollution modeling approaches may reduce bias in health effect estimation leading to improved impact evaluation of abatement policies.

AB - Background: Using modeled air pollutant predictions as exposure variables in epidemiological analyses can produce bias in health effect estimation. We used statistical simulation to estimate these biases and compare different air pollution models for London. Methods: Our simulations were based on a sample of 1,000 small geographical areas within London, United Kingdom. "True" pollutant data (daily mean nitrogen dioxide [NO2] and ozone [O3]) were simulated to include spatio-temporal variation and spatial covariance. All-cause mortality and cardiovascular hospital admissions were simulated from "true" pollution data using prespecified effect parameters for short and long-term exposure within a multilevel Poisson model. We compared: land use regression (LUR) models, dispersion models, LUR models including dispersion output as a spline (hybrid1), and generalized additive models combining splines in LUR and dispersion outputs (hybrid2). Validation datasets (model versus fixed-site monitor) were used to define simulation scenarios. Results: For the LUR models, bias estimates ranged from -56% to +7% for short-term exposure and -98% to -68% for long-term exposure and for the dispersion models from -33% to -15% and -52% to +0.5%, respectively. Hybrid1 provided little if any additional benefit, but hybrid2 appeared optimal in terms of bias estimates for short-term (-17% to +11%) and long-term (-28% to +11%) exposure and in preserving coverage probability and statistical power. Conclusions: Although exposure error can produce substantial negative bias (i.e., towards the null), combining outputs from different air pollution modeling approaches may reduce bias in health effect estimation leading to improved impact evaluation of abatement policies.

UR - http://www.scopus.com/inward/record.url?scp=85095846664&partnerID=8YFLogxK

U2 - 10.1097/EE9.0000000000000093

DO - 10.1097/EE9.0000000000000093

M3 - Article

AN - SCOPUS:85095846664

SN - 2474-7882

VL - 4

JO - Environmental Epidemiology

JF - Environmental Epidemiology

IS - 3

M1 - e093

ER -

Comparing the performance of air pollution models for nitrogen dioxide and ozone in the context of a multilevel epidemiological analysis

Abstract

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this