The Radon Project Web Site
Radioactive radon gas, seeping into houses from the soil below, poses a health risk to humans in certain parts of the United States. A new Web site, developed jointly by EETD and Columbia University's Department of Statistics, uses advanced mathematical methods, research on radon gas infiltration, and geologic data to help homeowners determine when and how to take action to reduce health risks from radon exposure.
Developed by EETD's Phil Price and Andrew Gelman, along with researchers at Columbia, the Radon Project Web site resides at the Columbia University Department of Statistics and Lawrence Berkeley National Laboratory Radon Project web site.
Radon is a naturally occurring radioactive gas, a product of radioactive decay of radium. Studies of uranium miners have shown that breathing high concentrations of radon can increase the risk of getting lung cancer. Long-term exposure of more than 20 picoCuries per liter (pCi/L) has been shown to cause lung cancer in people. Enough radon gas seeps out of soil into basements and crawlspaces in parts of the United States to pose a real health risk to people living in homes built on that soil.
The Web site can help homeowners determine what the chances are that they are breathing potentially dangerous levels of radon, and whether those risks are high enough to warrant measuring their actual exposure or having a contractor come in right away to install measures to reduce radon's infiltration in the house. Policymakers can also use the site as a tool to determine optimal policies for addressing radon health risks.
The site prompts the user to choose his or her home state and county from a U.S. map, then asks a few questions about the house's basement, whether any known measurements of radon were made in the area, and how many smokers and nonsmokers live in the house. Radon exposure increases the risk of lung cancer more for smokers than for nonsmokers.
The site then automatically inputs default values for acceptable risk levels and the estimated costs of reducing exposure to acceptable levels (assumed in this software to be 2 pCi/L or less). With this information, the software calculates the probable level of radon exposure in the home and provides the user with an exposure probability curve and a simple table showing the costs of various actions (doing nothing, short- or long-term testing to measure radon level accurately, or immediate remediation). It also recommends one of these four steps, based on the exposure level calculated.
The Environmental Protection Agency suggests that people with living-area concentrations over 4 pCi/L should remediate their homes to reduce this risk. Radon levels between 2 and 10 pCi/L probably pose some risk—the higher the level, the greater the risk—and levels above 10 pCi/L are considered definitely dangerous by experts in the field. Research suggests that in 50,000 to 100,000 homes, radon concentrations in living spaces exceed 20 pCi/L. This level is roughly equivalent to the lifetime exposure of a uranium miner and increases cancer risk among nonsmokers appreciably, among smokers by considerably more. A unique scientific attribute of the software is its use of a method called hierarchical modeling to help decide what step to take (radon measurement or immediate remediation) given the uncertainty about how much radon is present. Radon varies tremendously throughout the United States, ranging from almost none in some areas to more than 30 pCi/L in others. Even within a single U.S. county, radon can vary within a similar broad range. Hierarchical modeling allows the software to take the spatial variability of radon into account when calculating the odds of radon exposure where the home is located.
Price and Columbia's Gelman co-developed the computational theory underlying the software. Berkeley Lab provided the data and model parameters for the Radon Project site. Maps of soil radon concentration in the United States used in the Berkeley Lab data were developed by the U.S. Geological Survey. Price wrote the computer code for using the data in the site's software, while Columbia researchers wrote the FORTRAN program to implement the computational theory.
For more information, contact:
- Phil Price
- (510) 486-7875; fax (510) 486-6658
The Radon Project
Lin, Chia-yu, Andrew Gelman, Philip Price and David Krantz, 1999. "Analysis of Local Decisions Using Hierarchical Modeling, Applied to Home Radon Measurement and Remediation," Statistical Science, v. 14, no.3, pp 305-337.
This work is sponsored by the U.S. Environmental Protection Agency, and at Columbia University by the National Science Foundation.