Economic benefit-cost implications of the LNT model.

Abstract

As documented in this Special Issue, the linear no threshold (LNT) dose response model as a default model is scientifically invalid for both radiation and most chemicals. Consequently, there is no logical rationale to assume that the LNT model should be used to estimate health or safety benefits within benefit-cost analysis. Because thresholds are likely to exist for both radiation and chemicals, assuming that LNT is valid for economic analyses will lead to policy decisions with unnecessary costs imposed on society. From an economic perspective, a policy threshold is reached when the costs of decreasing exposures exceed the benefits. This paper investigates the use of LNT and policy thresholds using two examples to illustrate them, radon and formaldehyde. EPA s present LNT-driven program to mitigate indoor air radon to prevent lung cancer from homes and buildings costs billions of dollars each year while the benefits of the program are, at best, negligible. EPA s program to reduce exposure from formaldehyde in composite wood products costs about $60 million each year, again, with negligible (or zero) benefits. Both programs demonstrate that the economic threshold for decreasing exposure has, for the most part, already been attained and would have been correctly identified if a threshold model had been employed. Thus, use of a threshold model, instead of an LNT model, would have resulted in a different policy. Controversy over the use of LNT for risk assessment purposes goes back over 70 years to 1946 when Ernst Caspari reported a threshold response to radiation, based on the dose rate for gamma-ray-induced mutations in fruit flies [1]. Prior to that, ionizing radiation-induced mutations were assumed to be linear, down to zero (i.e., a single “hit” could induce cancer), with respect to the dose. Caspari s findings challenged the model originally created by Herman Muller. In fact, Muller believed he had induced mutations by dosing fruit flies with X-rays when, in fact, he made “large gene deletions and other gross chromosomal aberrations [2].” After Mullers mistakes, Caspari s findings set off a temporary alarm amongst the radiation genetics community such that it prompted one radiation researcher to ask (about the LNT), “What can we do to save the (one) hit model? [1]” Their worry was misplaced; Muller s views eventually won the day. But Ed Calabrese reports that “the LNT dose-response model, which drives cancer risk assessment, was based on flawed science, on ideological biases by leading radiation geneticists, on scientific misconduct by an NAS Genetics Panel during the atomic radiation scares of the 1950s, and on a 40 year mistaken assumption by yet another NAS Committee [3].” The controversy is important for multiple reasons. Scientifically, as extensively discussed in this Special Issue, the LNT makes no sense biologically. Second, from a risk perspective, chemicals and radiation are regulated to very low levels and those regulations often replace a very low risk with a higher risk from a substitute product or activity. This is called a risk/risk trade-off. Finally, use of the LNT inappropriately can lead to the imposition of unnecessary, and often very large, costs. It is no longer possible to ignore the costs of regulation. Today, nearly 300,000 federal workers (up from 57,000 in 1960 [4] put out 3–4000 regulations every year that have resulted in over 1 million restrictions (individual requirements) in the Code of Federal Regulations [5]. The cost of these regulations to the U.S. economy, although difficult to estimate, could be as high as $2 trillion each year [6]; nearly 11% of the U.S. GDP. Inappropriate use of the LNT can lead to spending too much on regulations and, because it results in overestimation of risks, will in turn cause benefits to be overestimated. Most benefit analyses currently use the results of risk assessments as the starting point. To be useful with estimates of cost, risk assessments need to estimate actual risks and to factor in the probabilistic information when possible to properly characterize the expected and net risks [7]. Beyond costs and benefits, there are other problems with using the LNT when it is not appropriate. In a staff report, EPA declares that it seeks to adequately protect public and environmental health by preferring an approach that does not underestimate risk in the face of uncertainty and variability. In other words, EPA seeks to adequately protect public and environmental health by ensuring that risk is not likely to be underestimated” [emphasis in original] [8]. To ensure that they do not underestimate risk, EPA staff routinely employ conservative defaults and assumptions that result in substantial overestimates of risk. Employing the LNT when there is a threshold is also conservative (precautionary). But by doing so, they have “effectively usurped risk management since managers (are) often never made aware of