Health risks and air pollution--error analysis for a cross-sectional mortality study.

Area of Science:

• Environmental epidemiology
• Biostatistics

Background:

• Multiple regression is frequently used to estimate the health impacts of air pollution.
• However, the accuracy of these estimates can be affected by various confounding factors.

Purpose of the Study:

• To quantitatively analyze uncertainties in multiple regression models estimating air pollution's effect on death rates.
• To assess the precision and robustness of these regression calculations.

Main Methods:

• Analysis of uncertainties in multiple regression estimates.
• Assessment of statistical fluctuations, local age distribution, smoking habits, pollution level errors, migration, and socioeconomic factors as sources of error.
• Illustrative calculations using UK death rates data from around the 1971 Census.

Main Results:

• Multiple regression estimates of air pollution's impact on mortality exhibit significant uncertainties.
• Factors such as statistical noise, demographic variations, lifestyle choices, and data inaccuracies contribute to poor precision and robustness.
• The reliability of regression models in this context is shown to be inadequate.

Conclusions:

• The precision and robustness of multiple regression analyses for air pollution effects on death rates are poor.
• Numerous factors introduce substantial uncertainty, challenging the validity of such estimates.
• Further methodological development is needed for reliable air pollution impact assessment.