Antiquated Risk Assessment in the Regulatory Context--The Radon in Drinking Water Standard. Anthony J. Thompson, Shaw, Pittman, Potts and Trowbridge, Washington, DC 20037
On July 18, 1991, the Environmental Protection Agency (EPA) proposed to set a Maximum Contaminant Level Goal (MCLG) of zero and a Maximum Contaminant Level (MCL) of 300 pico curies per liter (pCi/1) for a radon in drinking water standard under the Safe Drinking Water Act (SDWA). Recently, EPA indicated it might lower the proposed level to 200 pCi/l. Many in the scientific community (and the water industry) believe these levels are too low. The primary criticism centers upon risk assessment methodologies and risk management considerations in light of the statutory and regulatory requirements associated with the SDWA.
In proposing the radon in drinking water standard, EPA used a chemical risk paradigm, as typically would be expected for limiting potential exposure to potential manmade contaminants in drinking water, instead of a traditional radiation risk paradigm which considers the potential risk from significant increments to natural background radiation levels or variations therein. This led EPA to an assessment of the potential risk that focuses on achieving zero exposure from an environmental radionuclide when the public is exposed to significantly larger background radiation levels as a natural part of their environment. The comparative risk from radon in drinking water standard was not assessed in the context of natural exposure levels nor were the potential risks assessed in the context of their potential real world effects.
EPA's Science Advisory Board (SAB) noted that generally the proposed rule would be addressing a "very small contributor" to radon risk and at various times during the development of the proposed rule characterized the quality of EPA's support documents as "not good." Moreover, the SAB also noted that EPA failed to consider adequately the full range of risks associated with its proposal including the risks attendant to the recommended control technologies---in other words, to determine the "net risk benefit" from the proposed radon in drinking water standard. Additionally, while EPA stated that it intended to use "best estimates" rather than "conservative values" in developing its regulatory limit for radon in water, the record has not reflected the use of best estimates with quantified bounds of uncertainty,
In its initial proposal, EPA failed to take advantage of flexibility in its regulatory classification scheme for setting MCLGs but instead has relied on the linear non-threshold assumption. EPA also relied to some extent on the ingestion risk from radon in spite of the lack of experimental or epidemiological data to support such an assumption. Thus, the Agency has essentially locked itself into an unreasonably low standard that will cost water systems and their users millions of dollars for little, if any, benefit.
Assuming that the final proposal is lower (200 pCi/1) than the original 300 pCi/l proposal, this rulemaking appears to demonstrate that EPA's risk management decisions may still be the prisoner of conservative assumptions masquerading as risk assessment in spite of the Agency's stated intent to update its risk assessment capabilities.