Does consideration of larger study areas yield more accurate estimates of air pollution health effects?: An illustration of the bias-variance trade-off in air pollution epidemiology

Marie Pedersen; Valérie Siroux; Isabelle Pin; Marie Aline Charles; Anne Forhan; Agnés Hulin; Julien Galineau; Johanna Lepeule; Lise Giorgis-Allemand; Jordi Sunyer; Isabella Annesi-Maesano; Rémy Slama; ‘EDEN Mother–Child’ Cohort Study Group

doi:10.1016/j.envint.2013.07.005

Does consideration of larger study areas yield more accurate estimates of air pollution health effects? An illustration of the bias-variance trade-off in air pollution epidemiology

Marie Pedersen, Valérie Siroux, Isabelle Pin, Marie Aline Charles, Anne Forhan, Agnés Hulin, Julien Galineau, Johanna Lepeule, Lise Giorgis-Allemand, Jordi Sunyer, Isabella Annesi-Maesano, Rémy Slama, ‘EDEN Mother–Child’ Cohort Study Group

Center of Epidemiology and Screening

12 Citations (Scopus)

Abstract

BACKGROUND: Spatially-resolved air pollution models can be developed in large areas. The resulting increased exposure contrasts and population size offer opportunities to better characterize the effect of atmospheric pollutants on respiratory health. However the heterogeneity of these areas may also enhance the potential for confounding. We aimed to discuss some analytical approaches to handle this trade-off.

METHODS: We modeled NO2 and PM10 concentrations at the home addresses of 1082 pregnant mothers from EDEN cohort living in and around urban areas, using ADMS dispersion model. Simulations were performed to identify the best strategy to limit confounding by unmeasured factors varying with area type. We examined the relation between modeled concentrations and respiratory health in infants using regression models with and without adjustment or interaction terms with area type.

RESULTS: Simulations indicated that adjustment for area limited the bias due to unmeasured confounders varying with area at the costs of a slight decrease in statistical power. In our cohort, rural and urban areas differed for air pollution levels and for many factors associated with respiratory health and exposure. Area tended to modify effect measures of air pollution on respiratory health.

CONCLUSIONS: Increasing the size of the study area also increases the potential for residual confounding. Our simulations suggest that adjusting for type of area is a good option to limit residual confounding due to area-associated factors without restricting the area size. Other statistical approaches developed in the field of spatial epidemiology are an alternative to control for poorly-measured spatially-varying confounders.

Original language	English
Journal	Environment International
Volume	60
Pages (from-to)	23-30
Number of pages	8
ISSN	0160-4120
DOIs	https://doi.org/10.1016/j.envint.2013.07.005
Publication status	Published - Oct 2013

UN SDGs

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

Access to Document

10.1016/j.envint.2013.07.005

Cite this

Pedersen, M., Siroux, V., Pin, I., Charles, M. A., Forhan, A., Hulin, A., Galineau, J., Lepeule, J., Giorgis-Allemand, L., Sunyer, J., Annesi-Maesano, I., Slama, R., & ‘EDEN Mother–Child’ Cohort Study Group (2013). Does consideration of larger study areas yield more accurate estimates of air pollution health effects? An illustration of the bias-variance trade-off in air pollution epidemiology. Environment International, 60, 23-30. https://doi.org/10.1016/j.envint.2013.07.005

Does consideration of larger study areas yield more accurate estimates of air pollution health effects? An illustration of the bias-variance trade-off in air pollution epidemiology. / Pedersen, Marie; Siroux, Valérie; Pin, Isabelle et al.

In: Environment International, Vol. 60, 10.2013, p. 23-30.

Research output: Contribution to journal › Journal article › Research › peer-review

Pedersen, M, Siroux, V, Pin, I, Charles, MA, Forhan, A, Hulin, A, Galineau, J, Lepeule, J, Giorgis-Allemand, L, Sunyer, J, Annesi-Maesano, I, Slama, R & ‘EDEN Mother–Child’ Cohort Study Group 2013, 'Does consideration of larger study areas yield more accurate estimates of air pollution health effects? An illustration of the bias-variance trade-off in air pollution epidemiology', Environment International, vol. 60, pp. 23-30. https://doi.org/10.1016/j.envint.2013.07.005

@article{ce24dd87650a4b7293c9e375fdf9ae30,

title = "Does consideration of larger study areas yield more accurate estimates of air pollution health effects?: An illustration of the bias-variance trade-off in air pollution epidemiology",

abstract = "BACKGROUND: Spatially-resolved air pollution models can be developed in large areas. The resulting increased exposure contrasts and population size offer opportunities to better characterize the effect of atmospheric pollutants on respiratory health. However the heterogeneity of these areas may also enhance the potential for confounding. We aimed to discuss some analytical approaches to handle this trade-off.METHODS: We modeled NO2 and PM10 concentrations at the home addresses of 1082 pregnant mothers from EDEN cohort living in and around urban areas, using ADMS dispersion model. Simulations were performed to identify the best strategy to limit confounding by unmeasured factors varying with area type. We examined the relation between modeled concentrations and respiratory health in infants using regression models with and without adjustment or interaction terms with area type.RESULTS: Simulations indicated that adjustment for area limited the bias due to unmeasured confounders varying with area at the costs of a slight decrease in statistical power. In our cohort, rural and urban areas differed for air pollution levels and for many factors associated with respiratory health and exposure. Area tended to modify effect measures of air pollution on respiratory health.CONCLUSIONS: Increasing the size of the study area also increases the potential for residual confounding. Our simulations suggest that adjusting for type of area is a good option to limit residual confounding due to area-associated factors without restricting the area size. Other statistical approaches developed in the field of spatial epidemiology are an alternative to control for poorly-measured spatially-varying confounders.",

keywords = "Air Pollutants, Air Pollution, Asthma, Cohort Studies, Computer Simulation, Confounding Factors (Epidemiology), Environmental Monitoring, Female, Humans, Infant, Logistic Models, Models, Chemical, Models, Statistical, Nitrogen Dioxide, Particulate Matter, Pregnancy, Prenatal Exposure Delayed Effects, Prevalence, Questionnaires, Risk Factors, Rural Population, Sample Size, Urban Health, Urban Population",

author = "Marie Pedersen and Val{\'e}rie Siroux and Isabelle Pin and Charles, {Marie Aline} and Anne Forhan and Agn{\'e}s Hulin and Julien Galineau and Johanna Lepeule and Lise Giorgis-Allemand and Jordi Sunyer and Isabella Annesi-Maesano and R{\'e}my Slama and {{\textquoteleft}EDEN Mother–Child{\textquoteright} Cohort Study Group}",

year = "2013",

month = oct,

doi = "10.1016/j.envint.2013.07.005",

language = "English",

volume = "60",

pages = "23--30",

journal = "Environment international",

issn = "0160-4120",

publisher = "Pergamon Press",

}

TY - JOUR

T1 - Does consideration of larger study areas yield more accurate estimates of air pollution health effects?

T2 - An illustration of the bias-variance trade-off in air pollution epidemiology

AU - Pedersen, Marie

AU - Siroux, Valérie

AU - Pin, Isabelle

AU - Charles, Marie Aline

AU - Forhan, Anne

AU - Hulin, Agnés

AU - Galineau, Julien

AU - Lepeule, Johanna

AU - Giorgis-Allemand, Lise

AU - Sunyer, Jordi

AU - Annesi-Maesano, Isabella

AU - Slama, Rémy

AU - ‘EDEN Mother–Child’ Cohort Study Group

PY - 2013/10

Y1 - 2013/10

N2 - BACKGROUND: Spatially-resolved air pollution models can be developed in large areas. The resulting increased exposure contrasts and population size offer opportunities to better characterize the effect of atmospheric pollutants on respiratory health. However the heterogeneity of these areas may also enhance the potential for confounding. We aimed to discuss some analytical approaches to handle this trade-off.METHODS: We modeled NO2 and PM10 concentrations at the home addresses of 1082 pregnant mothers from EDEN cohort living in and around urban areas, using ADMS dispersion model. Simulations were performed to identify the best strategy to limit confounding by unmeasured factors varying with area type. We examined the relation between modeled concentrations and respiratory health in infants using regression models with and without adjustment or interaction terms with area type.RESULTS: Simulations indicated that adjustment for area limited the bias due to unmeasured confounders varying with area at the costs of a slight decrease in statistical power. In our cohort, rural and urban areas differed for air pollution levels and for many factors associated with respiratory health and exposure. Area tended to modify effect measures of air pollution on respiratory health.CONCLUSIONS: Increasing the size of the study area also increases the potential for residual confounding. Our simulations suggest that adjusting for type of area is a good option to limit residual confounding due to area-associated factors without restricting the area size. Other statistical approaches developed in the field of spatial epidemiology are an alternative to control for poorly-measured spatially-varying confounders.

AB - BACKGROUND: Spatially-resolved air pollution models can be developed in large areas. The resulting increased exposure contrasts and population size offer opportunities to better characterize the effect of atmospheric pollutants on respiratory health. However the heterogeneity of these areas may also enhance the potential for confounding. We aimed to discuss some analytical approaches to handle this trade-off.METHODS: We modeled NO2 and PM10 concentrations at the home addresses of 1082 pregnant mothers from EDEN cohort living in and around urban areas, using ADMS dispersion model. Simulations were performed to identify the best strategy to limit confounding by unmeasured factors varying with area type. We examined the relation between modeled concentrations and respiratory health in infants using regression models with and without adjustment or interaction terms with area type.RESULTS: Simulations indicated that adjustment for area limited the bias due to unmeasured confounders varying with area at the costs of a slight decrease in statistical power. In our cohort, rural and urban areas differed for air pollution levels and for many factors associated with respiratory health and exposure. Area tended to modify effect measures of air pollution on respiratory health.CONCLUSIONS: Increasing the size of the study area also increases the potential for residual confounding. Our simulations suggest that adjusting for type of area is a good option to limit residual confounding due to area-associated factors without restricting the area size. Other statistical approaches developed in the field of spatial epidemiology are an alternative to control for poorly-measured spatially-varying confounders.

KW - Air Pollutants

KW - Air Pollution

KW - Asthma

KW - Cohort Studies

KW - Computer Simulation

KW - Confounding Factors (Epidemiology)

KW - Environmental Monitoring

KW - Female

KW - Humans

KW - Infant

KW - Logistic Models

KW - Models, Chemical

KW - Models, Statistical

KW - Nitrogen Dioxide

KW - Particulate Matter

KW - Pregnancy

KW - Prenatal Exposure Delayed Effects

KW - Prevalence

KW - Questionnaires

KW - Risk Factors

KW - Rural Population

KW - Sample Size

KW - Urban Health

KW - Urban Population

U2 - 10.1016/j.envint.2013.07.005

DO - 10.1016/j.envint.2013.07.005

M3 - Journal article

C2 - 23994839

SN - 0160-4120

VL - 60

SP - 23

EP - 30

JO - Environment international

JF - Environment international

ER -

Does consideration of larger study areas yield more accurate estimates of air pollution health effects? An illustration of the bias-variance trade-off in air pollution epidemiology

Abstract

UN SDGs

Access to Document

Fingerprint

Cite this