Air pollution and mortality: Estimating regional and national dose-response relationships

F. Dominici, M. Daniels, Scott Zeger, J. M. Samet

Research output: Contribution to journalArticle

Abstract

We analyzed a national data base of air pollution and mortality for the 88 largest U.S. cities for the period 1987-1994, to estimate relative rates of mortality associated with airborne particulate matter smaller than 10 microns (PM10) and the form of the relationship between PM10 concentration and mortality. To estimate city-specific relative rates of mortality associated with PM10, we built log-linear models that included nonparametric adjustments for weather variables and longer term trends. To estimate PM10 mortality dose-response curves, we modeled the logarithm of the expected value of daily mortality as a function of PM10 using natural cubic splines with unknown numbers and locations of knots. We also developed spatial models to investigate the heterogeneity of relative mortality rates and of the shapes of PM10 mortality dose-response curves across cities and geographical regions. To determine whether variability in effect estimates can be explained by city-specific factors, we explored the dependence of relative mortality rates on mean pollution levels, demographic variables, reliability of the pollution data, and specific constituents of particulate matter. We implemented estimation with simulation-based methods, including data augmentation to impute the missing data of the city-specific covariates and the reversible jump Markov chain Monte Carlo (RJMCMC) to sample from the posterior distribution of the parameters in the hierarchical spline model, We found that previous-day PM10 concentrations were positively associated with total mortality in most the locations, with a .5% increment for a 10 μg/m3 increase in PM10. The effect was strongest in the Northeast region, where the increase in the death rate was twice as high as the average for the other cities. Overall, we found that the pooled concentration-response relationship for the nation was linear.

Original languageEnglish (US)
Pages (from-to)100-111
Number of pages12
JournalJournal of the American Statistical Association
Volume97
Issue number457
DOIs
StatePublished - Mar 2002

Fingerprint

Dose-response
Air Pollution
Mortality
Dose-response Curve
Particulate Matter
Mortality Rate
Pollution
Estimate
Reversible Jump Markov Chain Monte Carlo
Data Augmentation
Log-linear Models
Relationships
Air pollution
Cubic Spline
Spatial Model
Posterior distribution
Missing Data
Expected Value
Logarithm
Weather

Keywords

  • Air pollution
  • Data augmentation
  • Generalized additive model
  • Hierarchical model
  • Natural cubic spline
  • Particulate matter
  • Relative rate

ASJC Scopus subject areas

  • Mathematics(all)
  • Statistics and Probability

Cite this

Air pollution and mortality : Estimating regional and national dose-response relationships. / Dominici, F.; Daniels, M.; Zeger, Scott; Samet, J. M.

In: Journal of the American Statistical Association, Vol. 97, No. 457, 03.2002, p. 100-111.

Research output: Contribution to journalArticle

@article{853588e0965f45dd9d78b0b228d15911,
title = "Air pollution and mortality: Estimating regional and national dose-response relationships",
abstract = "We analyzed a national data base of air pollution and mortality for the 88 largest U.S. cities for the period 1987-1994, to estimate relative rates of mortality associated with airborne particulate matter smaller than 10 microns (PM10) and the form of the relationship between PM10 concentration and mortality. To estimate city-specific relative rates of mortality associated with PM10, we built log-linear models that included nonparametric adjustments for weather variables and longer term trends. To estimate PM10 mortality dose-response curves, we modeled the logarithm of the expected value of daily mortality as a function of PM10 using natural cubic splines with unknown numbers and locations of knots. We also developed spatial models to investigate the heterogeneity of relative mortality rates and of the shapes of PM10 mortality dose-response curves across cities and geographical regions. To determine whether variability in effect estimates can be explained by city-specific factors, we explored the dependence of relative mortality rates on mean pollution levels, demographic variables, reliability of the pollution data, and specific constituents of particulate matter. We implemented estimation with simulation-based methods, including data augmentation to impute the missing data of the city-specific covariates and the reversible jump Markov chain Monte Carlo (RJMCMC) to sample from the posterior distribution of the parameters in the hierarchical spline model, We found that previous-day PM10 concentrations were positively associated with total mortality in most the locations, with a .5{\%} increment for a 10 μg/m3 increase in PM10. The effect was strongest in the Northeast region, where the increase in the death rate was twice as high as the average for the other cities. Overall, we found that the pooled concentration-response relationship for the nation was linear.",
keywords = "Air pollution, Data augmentation, Generalized additive model, Hierarchical model, Natural cubic spline, Particulate matter, Relative rate",
author = "F. Dominici and M. Daniels and Scott Zeger and Samet, {J. M.}",
year = "2002",
month = "3",
doi = "10.1198/016214502753479266",
language = "English (US)",
volume = "97",
pages = "100--111",
journal = "Journal of the American Statistical Association",
issn = "0162-1459",
publisher = "Taylor and Francis Ltd.",
number = "457",

}

TY - JOUR

T1 - Air pollution and mortality

T2 - Estimating regional and national dose-response relationships

AU - Dominici, F.

AU - Daniels, M.

AU - Zeger, Scott

AU - Samet, J. M.

PY - 2002/3

Y1 - 2002/3

N2 - We analyzed a national data base of air pollution and mortality for the 88 largest U.S. cities for the period 1987-1994, to estimate relative rates of mortality associated with airborne particulate matter smaller than 10 microns (PM10) and the form of the relationship between PM10 concentration and mortality. To estimate city-specific relative rates of mortality associated with PM10, we built log-linear models that included nonparametric adjustments for weather variables and longer term trends. To estimate PM10 mortality dose-response curves, we modeled the logarithm of the expected value of daily mortality as a function of PM10 using natural cubic splines with unknown numbers and locations of knots. We also developed spatial models to investigate the heterogeneity of relative mortality rates and of the shapes of PM10 mortality dose-response curves across cities and geographical regions. To determine whether variability in effect estimates can be explained by city-specific factors, we explored the dependence of relative mortality rates on mean pollution levels, demographic variables, reliability of the pollution data, and specific constituents of particulate matter. We implemented estimation with simulation-based methods, including data augmentation to impute the missing data of the city-specific covariates and the reversible jump Markov chain Monte Carlo (RJMCMC) to sample from the posterior distribution of the parameters in the hierarchical spline model, We found that previous-day PM10 concentrations were positively associated with total mortality in most the locations, with a .5% increment for a 10 μg/m3 increase in PM10. The effect was strongest in the Northeast region, where the increase in the death rate was twice as high as the average for the other cities. Overall, we found that the pooled concentration-response relationship for the nation was linear.

AB - We analyzed a national data base of air pollution and mortality for the 88 largest U.S. cities for the period 1987-1994, to estimate relative rates of mortality associated with airborne particulate matter smaller than 10 microns (PM10) and the form of the relationship between PM10 concentration and mortality. To estimate city-specific relative rates of mortality associated with PM10, we built log-linear models that included nonparametric adjustments for weather variables and longer term trends. To estimate PM10 mortality dose-response curves, we modeled the logarithm of the expected value of daily mortality as a function of PM10 using natural cubic splines with unknown numbers and locations of knots. We also developed spatial models to investigate the heterogeneity of relative mortality rates and of the shapes of PM10 mortality dose-response curves across cities and geographical regions. To determine whether variability in effect estimates can be explained by city-specific factors, we explored the dependence of relative mortality rates on mean pollution levels, demographic variables, reliability of the pollution data, and specific constituents of particulate matter. We implemented estimation with simulation-based methods, including data augmentation to impute the missing data of the city-specific covariates and the reversible jump Markov chain Monte Carlo (RJMCMC) to sample from the posterior distribution of the parameters in the hierarchical spline model, We found that previous-day PM10 concentrations were positively associated with total mortality in most the locations, with a .5% increment for a 10 μg/m3 increase in PM10. The effect was strongest in the Northeast region, where the increase in the death rate was twice as high as the average for the other cities. Overall, we found that the pooled concentration-response relationship for the nation was linear.

KW - Air pollution

KW - Data augmentation

KW - Generalized additive model

KW - Hierarchical model

KW - Natural cubic spline

KW - Particulate matter

KW - Relative rate

UR - http://www.scopus.com/inward/record.url?scp=0036489049&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036489049&partnerID=8YFLogxK

U2 - 10.1198/016214502753479266

DO - 10.1198/016214502753479266

M3 - Article

AN - SCOPUS:0036489049

VL - 97

SP - 100

EP - 111

JO - Journal of the American Statistical Association

JF - Journal of the American Statistical Association

SN - 0162-1459

IS - 457

ER -