Test of the linear-no threshold theory: rationale for procedures

Area of Science:

• Environmental Health
• Epidemiology
• Radiation Science

Background:

• The linear-no threshold (LNT) theory is a model for radiation carcinogenesis.
• Ecological studies comparing population-level radon exposure and lung cancer rates face challenges with confounding factors.
• Accurate assessment of radon's carcinogenic effects requires robust methods to control for confounding variables.

Purpose of the Study:

• To rigorously test the linear-no threshold theory of radiation carcinogenesis.
• To evaluate the relationship between average radon exposure and lung cancer rates in U.S. counties.
• To develop and apply methods for effectively addressing confounding factors in ecological studies.

Main Methods:

• Employed ecological variables, comparing U.S. county-level lung cancer rates with average radon exposure.
• Introduced and utilized the method of stratification, demonstrating its superiority over multiple regression for confounding factor evaluation.
• Quantified requirements for confounding factors to influence results based on correlations with lung cancer rates and radon levels.

Main Results:

• Demonstrated that confounding factors are unlikely to invalidate the observed relationship between radon exposure and lung cancer rates.
• Showed that numerous confounding factors, when utilized, reduce the issue of representing complex confounding relationships with single average values.
• Found that the existence of an unknown confounder meeting the necessary correlation thresholds is highly implausible.

Conclusions:

• The study provides a strong justification for the procedures used in testing the linear-no threshold theory.
• The methods employed effectively resolve the problem of confounding factors at the individual level.
• The findings support the validity of the linear-no threshold model in the context of radon-induced lung cancer.