Dataset Information

Multipollutant measurement error in air pollution epidemiology studies arising from predicting exposures with penalized regression splines.

ABSTRACT: Air pollution epidemiology studies are trending towards a multi-pollutant approach. In these studies, exposures at subject locations are unobserved and must be predicted using observed exposures at misaligned monitoring locations. This induces measurement error, which can bias the estimated health effects and affect standard error estimates. We characterize this measurement error and develop an analytic bias correction when using penalized regression splines to predict exposure. Our simulations show bias from multi-pollutant measurement error can be severe, and in opposite directions or simultaneously positive or negative. Our analytic bias correction combined with a non-parametric bootstrap yields accurate coverage of 95% confidence intervals. We apply our methodology to analyze the association of systolic blood pressure with PM2.5 and NO2 in the NIEHS Sister Study. We find that NO2 confounds the association of systolic blood pressure with PM2.5 and vice versa. Elevated systolic blood pressure was significantly associated with increased PM2.5 and decreased NO2. Correcting for measurement error bias strengthened these associations and widened 95% confidence intervals.

SUBMITTER: Bergen S

PROVIDER: S-EPMC5076926 | biostudies-literature | 2016 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Multipollutant measurement error in air pollution epidemiology studies arising from predicting exposures with penalized regression splines.

Bergen Silas S Sheppard Lianne L Kaufman Joel D JD Szpiro Adam A AA

Journal of the Royal Statistical Society. Series C, Applied statistics 20160301 5

Air pollution epidemiology studies are trending towards a multi-pollutant approach. In these studies, exposures at subject locations are unobserved and must be predicted using observed exposures at misaligned monitoring locations. This induces measurement error, which can bias the estimated health effects and affect standard error estimates. We characterize this measurement error and develop an analytic bias correction when using penalized regression splines to predict exposure. Our simulations ...[more]

PMID: 27789915

Similar Datasets

Project description:BackgroundTwo distinctly different types of measurement error are Berkson and classical. Impacts of measurement error in epidemiologic studies of ambient air pollution are expected to depend on error type. We characterize measurement error due to instrument imprecision and spatial variability as multiplicative (i.e. additive on the log scale) and model it over a range of error types to assess impacts on risk ratio estimates both on a per measurement unit basis and on a per interquartile range (IQR) basis in a time-series study in Atlanta.MethodsDaily measures of twelve ambient air pollutants were analyzed: NO2, NOx, O3, SO2, CO, PM10 mass, PM2.5 mass, and PM2.5 components sulfate, nitrate, ammonium, elemental carbon and organic carbon. Semivariogram analysis was applied to assess spatial variability. Error due to this spatial variability was added to a reference pollutant time-series on the log scale using Monte Carlo simulations. Each of these time-series was exponentiated and introduced to a Poisson generalized linear model of cardiovascular disease emergency department visits.ResultsMeasurement error resulted in reduced statistical significance for the risk ratio estimates for all amounts (corresponding to different pollutants) and types of error. When modelled as classical-type error, risk ratios were attenuated, particularly for primary air pollutants, with average attenuation in risk ratios on a per unit of measurement basis ranging from 18% to 92% and on an IQR basis ranging from 18% to 86%. When modelled as Berkson-type error, risk ratios per unit of measurement were biased away from the null hypothesis by 2% to 31%, whereas risk ratios per IQR were attenuated (i.e. biased toward the null) by 5% to 34%. For CO modelled error amount, a range of error types were simulated and effects on risk ratio bias and significance were observed.ConclusionsFor multiplicative error, both the amount and type of measurement error impact health effect estimates in air pollution epidemiology. By modelling instrument imprecision and spatial variability as different error types, we estimate direction and magnitude of the effects of error over a range of error types.

Project description:Rationalepopulation-based studies have demonstrated associations between ambient air pollution exposures and mortality, but few have been able to adjust for occupational exposures. Additionally, two studies have observed higher risks in individuals with occupational dust, gas, or fume exposure.Objectiveswe examined the association of ambient residential exposure to particulate matter less than 10 microm in diameter (PM(10)), particulate matter less than 2.5 microm in diameter (PM(2.5)), NO(2), SO(2), and mortality in 53,814 men in the U.S. trucking industry.Methodsexposures for PM(10), NO(2), and SO(2) at each residential address were assigned using models combining spatial smoothing and geographic covariates. PM(2.5) exposures in 2000 were assigned from the nearest available monitor. Single and multipollutant Cox proportional hazard models were used to examine the association of an interquartile range (IQR) change (6 microg/m(3) for PM(10), 4 microg/m(3) for PM(2.5), 4ppb for SO(2), and 8ppb for NO(2)) and the risk of all-cause and cause-specific mortality.Measurements and main resultsan IQR change in ambient residential exposures to PM(10) was associated with a 4.3% (95% confidence interval [CI], 1.1-7.7%) increased risk of all-cause mortality. The increase for an IQR change in SO(2) was 6.9% (95% CI, 2.3-11.6%), for NO(2) was 8.2% (95% CI, 4.5-12.1%), and for PM(2.5) was 3.9% (95% CI, 1.0-6.9%). Elevated associations with cause-specific mortality (lung cancer, cardiovascular and respiratory disease) were observed for PM(2.5), SO(2), and NO(2), but not PM(10). None of the pollutants were confounded by occupational exposures. In multipollutant models, overall, the associations were attenuated, most strongly for PM(10). In sensitivity analyses excluding long-haul drivers, who spend days away from home, larger hazard ratios were observed.Conclusionsin this population of men, residential ambient air pollution exposures were associated with mortality.

Dataset Information

Multipollutant measurement error in air pollution epidemiology studies arising from predicting exposures with penalized regression splines.

Publications

Multipollutant measurement error in air pollution epidemiology studies arising from predicting exposures with penalized regression splines.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets