Impact of Measurement Error in Air Pollution Studies
Author Information
Author(s): Gretchen T Goldman, James A Mulholland, Armistead G Russell, Matthew J Strickland, Mitchel Klein, Lance A Waller, Paige E Tolbert
Primary Institution: Georgia Institute of Technology
Hypothesis
Measurement error in air pollution epidemiology affects risk ratio estimates depending on the type of error.
Conclusion
Both the amount and type of measurement error significantly impact health effect estimates in air pollution studies.
Supporting Evidence
- Measurement error can reduce the statistical significance of health effect estimates.
- Classical-type error leads to attenuation of risk ratios, while Berkson-type error can bias them away from the null hypothesis.
- Primary pollutants showed greater measurement error impacts compared to secondary pollutants.
Takeaway
When scientists measure air pollution, mistakes can happen that change how dangerous they think it is for our health. This study shows that different kinds of mistakes can make a big difference.
Methodology
Daily measures of twelve ambient air pollutants were analyzed using Monte Carlo simulations and Poisson generalized linear models to assess the impact of measurement error on cardiovascular disease emergency department visits.
Potential Biases
Potential bias due to the use of population-weighted averages for exposure assessment.
Limitations
The study focuses on specific pollutants and may not generalize to all air quality measurements.
Participant Demographics
Residents of the 20-county metropolitan Atlanta area.
Statistical Information
P-Value
0.000009
Confidence Interval
1.0078-1.0201
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website