Jan. 24, 2000


Columbia Web Site Helps Homeowners Evaluate Radon Risk

By Hannah Fairfield

Phillip Price, left, with Andrew Gelman in front of their new radon Web site.

If you've ever wondered if radon exposure is a risk in your home, a new Columbia University Web site might ease your fears. The site provides a county-specific map of the contiguous United States and a risk test to help homeowners estimate the relative radon exposure risk in their homes.

Columbia University Professor Andrew Gelman, a statistician, worked with radon expert Phillip Price of the Lawrence Berkeley National Laboratory to design the site in order to tell homeowners what to do about radon, which is a colorless, odorless, naturally occurring radioactive gas that can cause lung cancer if breathed in large concentrations over time. Their work has yielded the first local cost-benefit analysis for radon exposure risk.

The Environmental Protection Agency (EPA) recommends radon testing for every house in the United States, but Gelman and Price believe this to be unnecessarily broad and expensive. They believe that the decision to test should be made based upon local information.

"The goal is to have people think about radon in terms of local decision-making guided by professional expertise," Gelman said. "It's a step up from the EPA's recommendation, which would have many people testing in areas where the risk is in fact very small."

Gelman asserts that the total cost of lowering household radon levels could be reduced by 40 percent, while saving the same number of lives as the EPA's recommended policy, if homeowners used the new Web site to evaluate local risk rather than using the EPA's blanket statement.

In order to create a model for local recommendations, Gelman and Price had to first compile a map of average radon concentrations in the more than 3,000 counties in the United States, using radon measurements along with geologic and climate data. The researchers then developed a risk test that used such variables as the number of smokers and non-smokers in the house, and whether the basement is used as a living space. If a neighbor's radon measurement is known, that information can be used as well. The program establishes averages for any chosen county if homeowners don't know the answers to some questions. At the end, it recommends whether to measure the radon level in the house and gives the probability that the measurement will be high enough to recommend home remediation, which usually consists of installing a system that sucks air from beneath the house. This feature could help eliminate a problem faced by many people purchasing a new home.

"Most measurements are made as part of a home sale, so most people make short-term measurements," Price said. "Those are terribly inaccurate, but buyers think they need an answer quickly. Unfortunately, this leads to lots of unnecessary hassle and wasted money. Another option is to use our Web site to determine the expected cost associated with radon remediation, and then subtract that from the home price."

Gelman and Price, with Columbia University co-authors Chia-yu Lin and David Krantz, published a recent paper outlining their statistical method in Statistical Science, and they hope that other scientists will soon use the method for similar risk assessment problems.

"Evaluating radon is only the beginning," Gelman said. "We can use the same strategy to evaluate the cost-effectiveness of remediating many public health concerns, like urban causes of asthma. Using local data to evaluate public health risks is just a better way to go about making decisions."