Dioxin and the EPA: The Science and Politics of Regulation
From the Environmental Review Newsletter
Volume Two Number Five, May 1995
The U.S. Environmental Protection Agency has conducted a two-year reassessment of the effects of dioxins on human health. EPA has actively sought information from all interested parties: academic and medical researchers as well as environmental and industrial activists. In an unprecedented commitment to openness in government, the EPA has released preliminary drafts of the dioxin health assessment documents for comment and criticism. Industries likely to be adversely affected by stricter regulation of dioxins - car makers and electric utilities for example - have hired consultants to lobby congressional staff, testify before committees and respond to the preliminary EPA drafts. In December 1994, the editors of Science magazine published a letter signed by eighteen scientists which expressed unspecified reservations about the quality of the science in the final EPA health assessment document. The letter, generated by the consulting firm Environ, published in a prestigious scientific journal, is a public relations coup and should cause concern about the efforts of commercial interests to affect public policy and the science which informs it. We discussed this letter and its implications with one of the signatories, Dr. Donald Mattison, Dean of the School of Public Health at the University of Pittsburgh.
ER: Dr. Mattison, how would you characterize Environ?
DM: It is a consulting firm involved in a range of issues in which they provide scientific and other kinds of consultation to industry or to government or to whomever is interested in retaining their services. There is a disclaimer in the letter that the group of people that signed the letter were convened by a consulting firm - Environ - and expenses and an honorarium was received by everyone that participated.
I have participated on this panel intermittently over the past several years and the guidelines under which we have operated are that we are asked to comment on areas that represent our expertise. We give our opinions and if they like them or not that is fine. We say what we think, and that is provided as feedback both to the industry and to the EPA. In that context we were asked to review and comment on drafts of the dioxin risk assessment, on the discussions around the dioxin risk assessment, about interpretation of data. And we have worked in a very unconstrained way. It is enjoyable, in that they bring me together with a group of my colleagues who represent different disciplines and we argue and discuss and debate issues that relate to the risk assessment of dioxin. Our consultation has functioned in the best sense of how a scientific process should function: questioning, disagreeing; if you say something, what is the data that supports it? How strong is that data? What are the data gaps? I do not know how the people in the EPA view this, but if I were in their position, I would appreciate a rigorous critique of a risk assessment document before it is released. And I think that is one of the purposes this committee serves. If we are doing our job, the EPA is going to hear from us the most credible but severe criticism of their risk assessment. Then it is a risk manager's decision how to weigh this critique against the arguments the internal EPA scientists are making. I am sure that even within EPA, among some of the issues that end up being placed in a risk assessment document, there are internal discussions and debate. Within the EPA itself there are many credible scientists but they often bring different perspectives to what they do.
ER: Different in what way?
DM: Differing views of how data should be interpreted and extrapolated
ER: Different scientific style?
DM: Yes. Or perspectives. I have been extremely pleased with the openness of the EPA; I think the way they have invited public comment in this process has been superb. We have had the opportunity to comment and discuss the first eight chapters of the risk assessment document on at least a couple of occasions over the last few years, and it is very clear to me that at EPA they have thought about what we and others have said and that they have modified the risk assessment document in the context of those comments. It would have been nice if the capstone chapter, which is chapter nine had also been circulated through that same kind of process earlier.
ER: It was circulated?
DM: It is circulated now, but had the document been available in toto, then all of the discussion could have taken place, about the science in each of the chapters: the cancer chapter, the human effects, the non cancer effects, the reproductive effects, the endocrine effects and so on. I think the exchange of views, the scientific debate could have been a lot better. But my impression is that this is the most open of any of the risk assessment activities that EPA has done. So in a sense I almost hate to have to criticize them.
ER: It seems like we have made some progress in risk assessment. So I was puzzled by this letter in Science.
DM: The letter is rather obtuse but there is a lot of science behind it. The area I have the strongest background in is reproductive and developmental biology and thinking about how you translate animal data to humans in that arena. I have many concerns about the way that data was gathered and interpreted. And I was pleased that what I thought to be an inappropriate extrapolation clearly has evolved in succeeding drafts of the risk assessment.
ER: For instance?
DM: The initial document put a lot of weight on altering central nervous system function as it controls reproductive performance and behavior; using data from experimental animals and then extrapolating from that data to humans. There are a host of factors that alter reproductive performance and behavior in rodents that are extremely difficult to understand how you would extrapolate to humans.
ER: Dioxin has a specific receptor in virtually all mammalian cells and adversely affects cells in every mammalian system that has been looked at. I don't understand how these other factors can confuse the issue.
DM: One approach to risk characterization is to say that we do not care whether the end point does or even can occur in a human population; we will pick the most sensitive endpoint in any animal model and we will use that endpoint to set allowable human exposure levels, to this particular chemical. But there are some problems with that; for example, for some chemical compounds the kind of endpoint that is observed in experimental animals, cannot occur in humans, and as a matter of fact if you understand the mechanism of toxicity for some of these endpoints, you can demonstrate that the endpoint or toxicity could never occur in humans. We are now at a watershed of how much mechanistically based information we require and how we use it in characterizing and setting allowable exposure levels.
I have no problem with an argument that says we do not understand how this compound works but we are so concerned about potential for toxicity that we will take a very conservative approach. But you have to recognize that approach leaves a lot to be desired, and as you develop a better understanding about mechanism of action, you may need to step back from that. In this context, I would argue that this endpoint that we are looking at, a developmental effect, is alteration of development of the rodent brain that cannot occur in humans. That is to say that we cannot produce this kind of effect in humans, even in settings where we know that the human is exposed to high levels of toxicant that produces the same effect in experimental animals. Then I have to ask, What is the science base for making the extrapolation? At some point you have to make a decision that goes beyond the science that supports that decision. What we are arguing for is to be as explicit as you can. If you say that we are so worried about this compound that we are going to do this extrapolation even though we know that all of the things that produce this effect in experimental animals do not do the same thing in humans, is not that a risk management decision about the use of science to support a policy? I think, just be explicit. That is the only thing we have been arguing about.
ER: Is threshold doses what you are addressing in the letter?
DM: It is part of the issue. I have some concern about what actual body burdens are and what those body burdens translate into in terms of biological responses. For example, the current risk assessment argues that body burdens are at or above the threshold for biological response. But if that is the case, we ought to be able to demonstrate that. This is an idea developed by Allan Oakey. He reviewed the data on induction of Ah responsive processes like P450 in human placenta. He said, if that is the case, here is the data on humans for non-smokers and smokers, and in human populations that are exposed to dioxins, and among the non-smokers it is hard to find any P450 above the basal level, whereas you can see induction in smokers, and you can see induction in populations that are exposed to dioxins. Well if everything is down at the basement level for these cytochrome P450s in the non-smokers but we are all carrying around body burdens that are close to the threshold, should not we see some distribution, some evidence. That is a doable project, and it is doable in a short period of time. That is the kind of science base that would lend credence to the argument EPA is making.
ER: Argument and examination of assumptions is part of what makes science effective but I am concerned that people will be confused by this valid, useful discussion. When Dixy Lee Ray was lobbing grenades into the EPA bunker about dioxin, people were confused at least, and misinformed.
DM: This is a difficult subject. We have very limited resources, in both time and money, to direct toward public health concerns, and when a regulatory decision is made I think public health practitioners and officials need to be as strongly behind the decision as we can. I also think we need to be as forceful in our argument as possible for the quality of analysis that goes into the making of public health decisions. I get especially concerned because of the limitations in resources, that we do not let the science that goes into making public health policy be swayed by fashion or fad. And despite the disclaimer I gave at the beginning that we are supported by an industry association, some of the conclusions and some of the outcome of the deliberations that we do may end up having an adverse impact on industry. That has never been a part of any of the discussion on the panel. First of all readers need to know that this was funded by industry, but I think all of us approached the issue from as critical a science base as we could. I do not want to minimize what might be the potential human health effects of exposure to dioxins but I also think that we have to ask our regulatory agencies for their best effort in terms of credible science in supporting that. I think in the long run, that rigor goes to the credit of the agency because the harder it becomes to discredit their risk assessment in a scientific community, the stronger the agency's position.
ER: Doesn't that reflect the maturity of the science or funding priority? Environmental toxicology is both young and underfunded compared to pharmacology for instance.
DM: It is important to remember that this a young agency, it is twenty years old; EPA is about as old as the discipline of risk assessment. I think EPA has done remarkably well considering its youth and the brickbats that have been directed against it and its waxing and waning in a changing political environment.
ER: How vulnerable is the EPA to politics for its funding or in its approach to regulations?
DM: It is extremely vulnerable. And that is why I think the stronger the science the more diminished the vulnerability will be. A risk management decision is essentially a political decision; it is political, it is economic, it is science, it is all of those things thrown into the pot. But the stronger the science that drives these decisions, the harder it is to apply a political influence.
ER: Is part of your concern that perhaps we are setting a wrong priority in public health? That we could spend our money more effectively doing something else?
DM: In the scheme of things, how does human health alteration from the dioxins compare with non-immunized populations? With lead? With emerging infectious diseases? With the quality of the food supply? How would you rank dioxin against violence? That is not a fair comparison perhaps, but in public health we deal with informing people about things they need to be worried about. On the scheme of things to be worried about in life, how does dioxin fall? Is this a big issue, or is it an issue that will be a big issue today and tomorrow it is going to be something else? How can we develop some consistent way of choosing things for attention? How does dioxin stack up in terms of ecological health? Is this an issue we need to be concerned about for wild animal populations or biological diversity? Or are there other issues that should be capturing our attention?
ER: There are people in EPA who are pretty worried about dioxins in the environment but I don't think there is a clear consensus what that priority should be. Would you care to hazard a guess?
DM: The only thing I could do is defer because I am not sure that I know. Maybe we should think about the relative use of risk assessment and the relatively short time span we have focussed our attention on a comparative analysis of risks.
ER: As a scientist I worry about mixing science and policy.
DM: I think this is always going to be the issue of risk assessment, because we are always going to have to make decisions that go beyond what we understand or can demonstrate. So it is almost a given that there are going to be arguments and discussions and disagreements.
DM: It has been argued that we are not very good globally, world wide, in translating scientific knowledge into policy. And I think scientists both have to be careful of and aware of the fact that the insights they provide should appropriately be used to influence policy. However, when policy is developed, the way that knowledge is used may have different ends. I think that is what is playing out here with dioxin. It could be argued that all of our views are influenced by the groups that we work with, or perhaps the choice of a group to work with, is influenced by the perspective the scientist brings to the deliberative process.
ER: The eighteen members of your committee were selected by the consulting firm for a specific purpose, to support their position.
DM: There has been no attempt to influence what each of us thinks or says, by industry. But that does not mean that industry has been unbiased in the choice of individuals it has brought together.
ER: Or how they used your input.
DM: I look at this debate as a way of strengthening the science base for policy development around environmental risk. My sense from colleagues at the EPA was that they welcomed the input that this and other critical groups have provided because it allows them to play out before the regulations are actually put into place, all of the criticisms from as many different perspectives as possible on the regulations they are proposing, and allows them to fine tune or hone the regulatory process. I see it as being very constructive.
ER: Isn't it naive to say you are trying to provide the best science possible when this information is being used by industry for their own purposes.
DM: But this information is being used in different ways by all of the groups that latch on to it. Each group looks for something that supports its perspective in this deliberative process. I could turn around and apply the same sets of comments to the Natural Resources Defense Council or Environmental Defense Fund. All of us bring a perspective or a bias to this process of discussion. What I hope does not get lost is the collegiality of the debate. It is not good guys and bad guys, it is trying to understand how best to use what we know and what we do not know in developing policies that protect the health of the public.
ER: But that is applying the scientific model to the political debate and I don't see where that holds. Science is directed towards objective facts and our political culture is adversarial.
DM: There is no question politics is adversarial but that is also an element of science. If there is no adversarial basis in science then why is there debate and discussion? Why are there different models? I think risk assessment works like science at its edge: where our understanding is incomplete, where there are differing views or perspectives or perceptions about what the data actually reveal as being the truth. The challenge comes in translating both what is known and what is not known into policy. I think that is always going to be characteristic of risk assessment. The deliberative process has got to be part of it. I recognize all of these decisions ultimately get carried out in a context that is above science. Risk management is an attempt to balance these differing perspectives, their economic impacts, how can we best use the resources we have to enhance the health of the public.
It is important for us to recognize that despite the fact that science as a method has as its goal the development and the testing of hypotheses with continuous refinement, that it is being done by all of us who are human beings, and we bring to it our own personalities, characteristics or traits. The better we can inform the public about the value and the strengths and the weaknesses and limitations of science as it influences policy, the more realistic people are going to be in their understanding of how or why decisions are being made.
Copyright 1995 Environmental Review