Original Article Dose-Response: An International Journal When Risk Assessment Came January-March 2019:1-15 ª The Author(s) 2019 Article reuse guidelines: to Washington: A Look Back sagepub.com/journals-permissions DOI: 10.1177/1559325818824934 journals.sagepub.com/home/dos Joseph V. Rodricks1 Abstract Federal regulatory agencies had, by the 1970s, been charged with enforcing a host of new laws requiring that they establish controls on human exposures to chemicals necessary to protect health. The agencies relied upon a methodology introduced in the 1950s to identify safe levels of exposure to chemicals known to display toxicity. During the 2 decades prior to the 1970s, federal authorities had come to treat carcinogens as distinct from other toxic agents, and to regard them as unsafe at any level of exposure, and no systematic methods had been developed to deal with the rapidly increasing numbers of carcinogens. Beginning in the mid-1970s, some scientists and policy makers in regulatory agencies, including the present author, began to propose adopting emerging quantitative methods to evaluate the risks of carcinogens and introduced new notions of safety based on explicit consideration of risk. Quantitative risk assessment rose to prominence in the decade reviewed in this article (1974-1984) and began to replace the unsystematic approaches that provided no view of how well health would be protected under various regulatory controls. This article offers the author’s recollections of that important decade. Keywords carcinogens, regulation, introduction of risk assessment, early controversies, red book I spent a good part of the period from 1974 to 1984 attempting report: “Risk Assessment in the Federal Government: Manag- to understand chemical carcinogenicity, the quantification of ing the Process.”1 The Red Book, as it came to be called, carcinogenic risk and its scientific underpinnings, and the value offered a clear path forward for risk-based decision-making, of risk assessment for public health and regulatory decision- the importance of which is still underappreciated today. making. Although scientific and policy debates pertaining to these matters continue today, they were of a somewhat differ- A Fungal Metabolite and An Endocrine ent kind during the decade in which risk assessment, as it is currently defined, was beginning to achieve prominence. The Disruptor debates that were central to the emergence of risk assessment For the first 6 years of the decade I shall discuss, I was during that critical period, and the ways in which some of them employed as a scientist at the US Food and Drug Administra- were resolved can, I believe, reveal much about how and why tion (FDA). I had entered the agency in 1965 as a lab scientist, risk assessment came to be the force it is today and about why assigned to study the important fungal metabolite and food certain difficulties pertaining to its conduct and uses persist. contaminant known as aflatoxin (actually a group of closely The editors of this journal have invited me to offer my recol- related compounds). Gerald Wogan, at the Massachusetts Insti- lections and perceptions of that somewhat combative but nev- tute of Technology (MIT), and other scientists in England had ertheless highly productive period, and I hope what I present demonstrated the animal carcinogenicity of aflatoxin and had here will contribute to an understanding of how the present world of risk assessment and risk-based decision-making came 1 Ramboll, Arlington, VA, USA to be. At the beginning of the decade about which I write, quanti- Received 29 August 2018; accepted 30 August 2018 tative risk assessment had no role in regulation or other public Corresponding Author: health efforts. When the decade ended, it had become central to Joseph V. Rodricks, Ramboll, 4350 North Fairfax Drive, Suite 300, Arlington, many programs at the federal and state levels. The decade also VA 22203, USA. ended with the publication of the famous National Academies Email: jrodricks@ramboll.com Creative Commons Non Commercial CC BY-NC: This article is distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 License (http://www.creativecommons.org/licenses/by-nc/4.0/) which permits non-commercial use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage). 2 Dose-Response: An International Journal shown that one of these substances, known as B1, was capable whatever method of analyses is used. The LOD was a de facto of producing malignancies at doses lower than those at which safety standard.4 any other known carcinogen displayed similar activity. (Even In the case of drugs used in food-producing animals, FDA is today, aflatoxin is surpassed in this respect only by 2,3,7,8- required to approve the analytical method to be used and, tetrachlorodibenzo-p-dioxin. It is interesting to compare and therefore, must specify the LOD to be achieved. As I had contrast these 2 very potent animal carcinogens, which exhibit observed in the case of aflatoxin, safety was in the hands of substantially different kinetic and dynamic behaviors.) analytical chemists. Food residues of a highly potent carcino- By 1970, aflatoxins were found to be common contaminants gen might be permitted at higher levels than those of a far less- of certain human foods, their carcinogenicity had been repro- potent carcinogen, simply because the LOD for the highly duced in several animal species, and suggestive evidence of potent compound turned out to be at a higher concentration carcinogenicity emerged from epidemiology studies in certain than the LOD for the less potent one. This made no sense and populations experiencing relatively large exposure through the LOD approach to regulation could not be claimed to satisfy foods. The FDA had placed limits on the allowable levels of any criteria for safety. aflatoxins in foods, based on what were considered to be the In the late summer of 1974, I had occasion to meet with Leo limits of analytical detection—initially 30 mg/kg and reduced Friedman, director of the FDA’s toxicology division. I raised to 20 mg/kg in the early 1970s as analytical methods improved. with Friedman my concerns about reliance upon analytical Although I was involved in implementing the regulatory detection limit as the basis for regulations. This approach made scheme, it struck me as an odd approach. I understood that the some sense for aflatoxins because they were food contaminants FDA was required to ensure the safety of food, but it was and could not be readily controlled in the way an intentionally obvious that analytical detection limits were not a measure of introduced substance such as DES could be controlled. But in safety. Moreover, detection limits continued to decline and neither case did possible risks to human health seem to enter reached 2 mg/kg by 1972. Strict application of this approach the FDA approach to regulation. I had laid out this issue in a 3- to regulation would result in loss of very large fractions of some page memo, which I delivered to Friedman. affected crops, most especially peanuts.2 Although Friedman was a scientist, he started our conver- As I began to become more involved in issues related to the sation by reminding me that the FDA’s activities were gov- uses of science in regulation, I was asked to assist with another erned by the Federal Food, Drug, and Cosmetic Act and that compound, one used in food production since about 1950, that certain amendments introduced in 1958 that pertained to food came under intense scrutiny in the early 1970s. The compound additives contained the requirement that no substance could was diethylstilbestrol (DES), a synthetic compound with estro- be legally introduced into food unless it was shown to be safe genic properties approved for use in women who were unable under its conditions of use. The language of the law then to maintain pregnancy because of natural estrogen deficiency. went on5: In 1970, DES was identified as the highly probable cause of vaginal adenocarcinomas in young women whose mothers had Provided, that no additive shall be deemed to be safe if it is used DES during pregnancy.3 The Food and Drug Administra- found to induce cancer when ingested by man or animal . . . , tion immediately acted to prohibit the use of DES in human medicine. As the evidence for DES-induced carcinogenicity This is named the Delaney Clause, introduced by Represen- emerged, it ignited new concerns regarding the widespread use tative James Delaney of New York, chairman beginning in of DES as a growth-promoting agent in animals used for human 1950 of the House Select Committee to Investigate the Use of food. This use, approved in 1949, was known to result in low Chemicals in Food Products. Although the Committee heard levels of DES residues in meat; some evidence of the carcino- from many experts before adopting the new food additives genicity of DES in animals had been developed in the 1960s, amendment in 1958, perhaps the most significant testimony but only after the human findings emerged in the early 1970s came from Arthur Fleming, then Secretary of the Department did the FDA act against the animal drug uses. The Food and of Health, Education, and Welfare (now Health and Human Drug Administration’s actions in this matter were hampered by Services). Fleming offered the following statement from a the fact that the applicable food law actually permitted the use National Cancer Institute report: of drugs such as DES if that use could be shown to result in “no residue” of the drug in human food. Here, as in the case of No one at this time can tell how much or how little of a carcino- aflatoxin, we encounter the issue of analytical detection limits gen would be required to produce cancer in any human being, as the determinant of the amount of carcinogen permitted to be or how long it would take for cancer to develop. present in food. Although the law dictates that there should be “no residue,” the only way to determine compliance with such This “no safe level” concept, although directed at the time at a standard is to sample food and apply an analytical method to substances intentionally added to food, had much broader determine whether a residue of the carcinogen can be detected. influence in the world of chemical regulation. Every analytical method has, of course, a detection limit, so I nevertheless persisted in my argument that reliance upon that, even if “no residue” is found, it can be concluded only that analytical detection limits for regulation, without considering there is no residue greater than the limit of detection (LOD) of risks to human health, seemed an inadequate approach to Rodricks 3 decision-making, especially in light of the powerful dictate of was intended to impose a “conservative” element into the pro- the Delaney Clause. The latter may not be legally applicable to cedure for extrapolation into the unknown.8 a food contaminant such as aflatoxin, but it seemed clear that it It seemed to me that the approach described by Mantel should be applicable to an intentionally introduced substance provided a way to approximate the magnitude of the health such as DES. risks carcinogens might pose at low (human) doses, and to set Leo Friedman told me that my 3-page memo echoed some standards based on the notion that once risks reached some very ideas he had been thinking about for several years. He told me low levels, we could declare that exposures in these ranges and that the method for establishing safe levels for most chemicals, below were not a health threat. Decision-making would not be devised by his predecessor at the FDA, Arnold Lehman, and deterministic (“safe/not safe”) but rather probabilistic (ie, risk- another FDA toxicologist, O. Garth Fitzhugh, was based on a based). widely accepted view, held by most toxicologists, that the toxic It would also become possible to gauge the magnitude of properties of most chemicals expressed themselves only after a risk reductions achieved as regulatory standards were tigh- threshold dose was exceeded and that concept provided a sci- tened, so that policy makers could examine the important ques- entific basis for establishing safe levels for humans. The Leh- tion of whether the public health benefits achieved by the man–Fitzhugh approach, published in the 1949 to 1955 period, imposition of various control technologies were worth the eco- relied upon the application of what were then called “safety nomic and social costs of achieving them. It appeared that a factors” to data obtained from experimental toxicology studies systematic means for dealing with the increasing numbers of and, in some cases, epidemiology studies.6 But Friedman was carcinogens that could be found in the environment was avail- aware that there was a community of experts working in the able and that it should be developed for practical application. areas of chemical and radiation carcinogenesis who had devel- Decision-making would be linked to the risk characteristics oped quite different views of the biological actions of agents specific to individual carcinogens, together with other factors having carcinogenic properties and that their notions of thresh- that dictate the practical limits of risk reduction technologies. olds, reversibility, and dose–response were radically different In late 1974, I spent several months scrutinizing many ani- from those of the traditional toxicologists.7 mal studies then available on the carcinogenicity of aflatoxins Leo Friedman pulled from his file a thin folder containing and exploring, together with 2 FDA biostatisticians, the impli- half a dozen publications he asked me to study. He felt it was cations of applying the Mantel–Bryan procedure for low-dose time for the FDA to find a more scientifically satisfactory way extrapolation; we applied other statistical procedures as well. to deal with carcinogens. I promised to return after I had gone We estimated how human intakes of aflatoxin would decline if through those few papers. regulatory tolerance levels were to be made more restrictive, and we then estimated corresponding risk reductions. I worked with FDA policy officials to begin to craft a risk-based toler- “Virtually Safe Doses” ance level for aflatoxins in peanut products.9 I studied the papers Leo Friedman had given me but never Because of my work with aflatoxins, I was asked to join discussed them with him. Leo died of heart failure not long another agency effort to move toward risk-based decisions. after our meeting, at age 52. He was one of the most thoughtful This effort concerned not food contaminants such as aflatoxins, and creative people I have known in the field of toxicology, and but rather the class of intentionally added substances repre- his early death was a significant loss to our community. sented by DES. The uses of these veterinary drugs would result The single paper in the Friedman collection that I found in their presence as “residues” in meat, milk, or eggs. Because most valuable was one published in 1961 by the prominent they were intentionally added substances, the original form of National Cancer Institute (NCI) biostatistician, Nathan Mantel, the Delaney Amendment applied to any such drug that was and an associate of his, William Bryan. I arranged through one found to be carcinogenic. During the 1960s, however, our Con- of my office directors for Mantel to present his work at a gress modified the law to allow the use in food production of seminar at our FDA offices, and he agreed to do that. Mantel animal drugs that were carcinogenic. That modification permit- demonstrated how carcinogenicity dose–response data on a ted such use if, as I noted earlier, “no residue” of the drug could series of polycyclic aromatic hydrocarbons could be described be detected in human food. This modification of the law came with a simple probit model, and how the tail of the model could to be known as the “DES Proviso.”10 be extrapolated downward from the low end of the observed It turns out that at this time, the FDA was blessed with a very dose–risk relationship (which typically describes lifetime risks astute general counsel who had great foresight, Peter Barton of tumor development no less than about 1 in 10), to estimate Hutt. Hutt had come to lead the effort to put risk assessment doses corresponding to some extremely low and completely into the regulatory equation for this class of added food ingre- unmeasurable risks. Mantel’s risk target was 1 in 100 million, dients. After extensive discussions with me and other scientists, and the calculated doses corresponding to that excess lifetime Hutt proposed that “safe doses” for carcinogens such as DES risk he labeled “virtually safe.” I should add that Mantel, in could be defined as those associated with lifetime risk levels of extending the tail of the probit model, imposed an artificial less than 1 in 1 million, when these risks were estimated using a slope on it that he thought would place an “upper bound” on linear, no-threshold model (several publications had demon- the low-dose risk, so that risk would not be underestimated; it strated that the Mantel–Bryan approach could not be counted 4 Dose-Response: An International Journal on to place an upper bound on risk at low doses but that a linear, I learned from discussions with Elizabeth Anderson, the no threshold model could—see later). Carcinogenic animal EPA scientist having the responsibility for establishing risk- drugs would be acceptable only if it could be shown that their based approaches in the agency, and the agency’s superb con- uses led to food residues at levels no greater than the level sultant on this matter, Roy Albert of New York University, that that corresponded to the “safe dose,” as defined in the fore- EPA was moving quickly on this topic. The agency had pub- going. Those seeking drug approval would be required to lished written guidelines on the conduct of carcinogenic risk develop analytical methods capable of reliably detecting the assessment12 and had major efforts underway to implement safe residue level and demonstrating that “no residue” could these guidelines. EPA was, I thought, moving in the right direc- be found in food, under the intended conditions of drug use, tion, with what seemed to me much greater internal support when that analytical method was applied. The “Sensitivity of than I and my fellow risk assessment advocates were receiving the Method” regulation became the first formally to adopt a within the FDA. Indeed, many toxicologists within FDA were risk-based approach for carcinogens.11 uncomfortable with this new approach to safety assessment. The FDA, after a protracted administrative hearing, acted to Elizabeth and Roy wanted to speak with me because they extinguish the veterinary uses of DES in 1979, based in part on were concerned about the possibility that, as different federal the fact that levels of residues detected in food did not meet agencies moved to adopt quantitative risk assessment these new risk-based safety criteria.4 approaches for carcinogens, inconsistencies might arise in the Although it was estimated in a somewhat different way methods proposed for use. Such inconsistences might weaken than that proposed by Mantel and Bryan, their “virtually safe the credibility of federal efforts to incorporate risk assessment dose” became Peter Hutt’s “safe dose” (FDA, he used to say, into decision-making. did not permit doses for added carcinogens that were only One of the reasons for this concern was the proliferation, “virtually” safe). The selection of a lifetime risk level consid- during the mid-to-late 1970s, of statistical models proposed for ered sufficiently low to ensure safety was not a scientific, but low-dose extrapolation. The various models discussed in the rather a policy choice. Thus, emerged what later came to be literature during this time equally well described the observed labeled a risk management decision, distinguishable from risk dose-risk relationships, but predicted large differences in low- assessment. dose risks for the same carcinogens, and the existence of these differences could be (and were) used to cast doubt on the validity of the risk assessment methodology. I shall deal with this important dilemma later but for now simply note that the The EPA EPA was relying upon Kenny Crump’s important 1976 publi- I do not recall how I came to be contacted by the EPA, but I cation, illustrating the use of the “linearized” multistage model believe publication of the “Sensitivity of the Method” regula- for low-dose extrapolation.13 I assured Elizabeth and Roy that tion in 1977 attracted the attention of that agency. The EPA, the FDA, after reviewing the Mantel-Bryan approach and sev- responding to the large number of new laws and regulatory eral others, had elected to adopt the linear model. requirements that emerged in the 1970s, had to deal with many Many other questions arose in connection with carcinogenic carcinogens, some, such as pesticides, intentionally introduced, risk assessment, but this one issue of low-dose extrapolation but most occurring as widespread contaminants of air, water, was, at this time, the most controversial and the most likely to and soils. The EPA and its precursor agencies had used the threaten this new and badly needed approach to dealing with traditional methods of toxicology, based on the threshold con- carcinogens. cept, to establish health-based standards but had no methodol- ogy to deal with carcinogens. The agency had, for pesticide residues in food, to enforce the Delaney Clause but was faced Interagency Collaboration 1977 to 1980 with a “no residue” requirement similar to the one I have The 5 federal regulatory agencies with responsibilities for described for veterinary drug residues. The EPA was relying chemical regulation all had new leadership under the Carter upon analytical chemistry criteria and technological achiev- administration: Douglas Costle (EPA), Eula Bingham (Occu- ability for contaminants, in much the same way FDA had for pational Safety and Health Administration [OSHA]), Susan aflatoxins and other carcinogenic contaminants of food King (CPSC), Carol Tucker Foreman (United States Depart- (PCBs became during the late 1970s another outstanding ment of Agriculture (USDA)—Food Safety and Inspection Ser- example of this problem). I learned that some EPA scientists vice), and Donald Kennedy (FDA). I moved into a science well understood that the absence of any systematic way to advisor position under Kennedy in early 1977. These new deal with the health risks posed by carcinogens was a serious agency heads got together and formed what was called the impediment to decisions having the primary purpose of pro- Interagency Regulatory Liaison Group (IRLG), having the pur- tecting human health. Earlier agency efforts proposing com- pose of ensuring consistent approaches to various common plete bans on exposures to carcinogens had incurred serious scientific and policy issues. The new regulatory leadership was criticisms from many quarters, and, in the end, were rejected in part reacting to increasing public attacks on perceived reg- as impractical. Achieving zero risk on a wide scale was not ulatory failures. In early 1977, for example, the FDA proposed an available option.1 to ban the hugely popular noncaloric sweetener, saccharin, on Rodricks 5 the basis of bladder cancer findings in rats administered record and a strong commitment to further occupational health. extremely high doses (percent levels in the diet) and the The centerpiece of her program was a new proposal to regulate requirements of the Delaney Clause. Attacks on this proposed occupational carcinogens by establishing workplace standards action played out in all major media, with much ridicule and based on best available control technology, also considering scientific criticism of animal studies. The agencies knew they costs. Dr. Bingham and her staff had put an enormous effort could not succeed in their responsibilities without reliance into this proposal and had moved ahead to impose a new Per- upon animal toxicity studies, and assembled an interagency missible Exposure Limit (PEL) on benzene, based on this new committee under the IRLG, involving experts from the reg- approach.14,15 ulatory agencies and from federal public health agencies, to During our meeting, in which I reported on the status of the develop and bolster scientific support for the use of animal Work Group effort, Dr. Bingham explained her deep concern data in decisions. Several other committees were put to work about reliance on quantitative risk assessments. The approach on various topics, including the Work Group on Risk the OSHA was proposing to manage risks from occupational Assessment.7 carcinogens was not risk-based. Rather, a finding that a sub- Eula Bingham, head of the OSHA, was the IRLG member stance was a carcinogen would be sufficient to trigger chosen to oversee this group, and I was asked to chair the technology-based controls. This approach to regulation was group. All the agencies contributed members, including Eliza- hazard-based and conceptually similar to the regulation I ear- beth Anderson of the EPA, and David Gaylor, a statistician lier described in connection with food contaminants and addi- from the FDA’s National Center for Toxicological Research, tives. Dr. Bingham was highly concerned that the IRLG effort, someone I found to be an excellent guide to my own thinking. which was to put into place quantitative risk assessment meth- Our work received superb assistance from Roy Albert, David ods for carcinogens, would necessitate risk-based decisions, Hoel (then at NIEHS), and Umberto Saffiotti and Marvin and so would undermine the new OSHA proposal. I had no Schneiderman, both of the National Cancer Institute. The Work immediate response, except to say that the Work Group’s effort Group met many times over the following 18 months, sought would be considered a guideline for the conduct of risk assess- and received advice from major science leaders in government, ment, should an agency choose to undertake such an assess- including Arthur Upton (director of NCI) and David Rall ment, and not a requirement that regulation be based on such an (director of NIEHS). The effort of the Work Group was pub- assessment. We both knew this was not an especially compel- lished in the Journal of the National Cancer Institute in a dense ling argument, and we also knew that the FDA, and most espe- and highly detailed paper entitled “Scientific Basis for Identi- cially the EPA, was committed to risk-based decision-making, fication of Potential Carcinogens and Estimation of Risks.”7 except where explicitly prohibited by law.1 Bingham and I The contention of the Work Group was that, although the fed- could not find a clear path forward on this issue, but the IRLG eral regulatory agencies had different legislative mandates, effort continued, ending with the 1979 report.7 requiring different approaches to decision-making, the agen- I recall numerous and somewhat tense calls and meetings cies could agree on common approaches to risk assessment. with the EPA on this matter, and these continued until a case Although the IRLG Work Group report certainly did not settle brought against OSHA by the American Petroleum Institute all scientific questions and disputes, it consolidated federal was decided by the U.S. Supreme Court in 1981. The Court agency thinking and set the stage for ensuring a systematic, ruled that the OSHA proposal was not consistent with federal relatively transparent, and consistent approach to evaluating law and that the agency had to demonstrate that the existing and managing the difficult problem of carcinogens in foods occupational exposure to a carcinogen carried a significant risk and consumer products, the environment, and the workplace. to health and that proposed exposure reductions would result in I and several other Work Group members and agency scien- a significant reduction in risk.16 Although factors other than tists and officials made public presentations in many venues risk would play a role in standard setting, risk and risk reduc- regarding the IRLG effort on risk assessment. I continued in my tion were essential criteria. This decision arrived well after the day job at FDA, devoting much time to the public hearing the publication of the Work Group’s guidelines but did much to agency held on DES, trying to explain the value of risk assess- ensure a permanent place for risk assessment in regulation and ment to agency scientists and policy makers, and interacting public health decision-making.17 with experts in government, industry, and the academic com- munity. I shall relate here some of the seminal events relating to the introduction into decision-making of quantitative risk The FDA and Its Critics assessment. As I have said, the DES matter was the subject of a public hearing at the FDA, beginning in 1979, and there emerged during these hearings other views of possible limits of the Concerns From the OSHA risk-based approach. There was no doubt that low levels of this Sometime during the 18 months of deliberations of the IRLG drug could be found in meat from treated cows, so it would Working Group, I was invited to visit with Eula Bingham in her seem that there should be no question that the “no residue” office at the OSHA. Dr. Bingham had come to the agency from requirement of law was being violated. Experts engaged by the the University of Cincinnati with a distinguished scientific regulated industry, however, brought forth scientific arguments 6 Dose-Response: An International Journal that the levels of those residues carried no significant health that profoundly influenced my attitude and approach to advo- risk and that they should be allowed. Professor Elwood Jensen cating increased uses of risk assessment in decision-making. of the University of Chicago, a recognized expert in hormonal I believe it occurred in the summer of 1979 and came about carcinogenesis, testified that4: at the request of another highly regarded scientist, Dr. Arthur Upton, then director of the National Cancer Institute, and a There is no evidence of any fundamental difference between the giant in the area of radiation risk. I sat for an hour with Ken- hormonal action of DES and that of the endogenous hormone nedy and Upton, and discussed risk assessment and policies in estradiol. relation to carcinogen regulation. I sensed that Dr. Upton had discussed the IRLG effort with the 2 NCI scientists who were And that he knew of: participating in the effort, Umberto Saffiotti and Marvin Schneiderman. I had earlier discussions with each of these . . . no instance in which it is established that the cancer- scientists and knew that one of them (Saffiotti) had concerns enhancing effect of DES cannot be duplicated by an appropriate of the same kind Eula Bingham had expressed. He also thought dose of a steroidal estrogen. that quantitative risk assessment was too uncertain to be used as a guide to decisions. Dr. Upton expressed similar concerns to Jensen was, in effect, saying that meat naturally contains Kennedy and me but also recognized that if carefully con- estrogen and so do the consumers of that meat, and the biolo- ducted, described, and applied, quantitative models of risk gical properties of the synthetic estrogen were in all respects could add much needed rigor to decision-making. identical to those of natural estrogens. The addition of DES The discussion that followed was at a much-elevated level; I contributes insignificantly to the normal, natural background was mostly a listener and came away with a greatly improved levels of estrogens, and whatever it contributes is identical to understanding of science in the formulation of public policies, the natural and safe estrogen exposures. This is in essence the and the need for extreme caution in the elaboration of scientific “threshold” argument that is still offered by some today in knowledge and its limitations. The temptation to leap beyond relation to what have been labeled “endocrine disruptors.”4 what is truly established knowledge can be great if that leap can A similar argument, including one based on an interesting advance some desired policy agenda, but doing so can threaten but little examined approach to quantitative risk assessment, scientific credibility and backfire. At the same time, these 2 based on biological mechanisms, was offered by Professor great minds agreed, in the area of public health protection, it Thomas Jukes of Berkeley, a prominent molecular biologist. may be necessary, for policy reasons, to introduce certain pre- Jukes contended that detectable residues of DES in meat posed cautionary elements into the interpretation and uses of scien- no significant health risk. Both Jensen and Jukes, 2 superb tific information; the goal, always, is to find the right balance in scientists, cast significant doubts on the risk assessment the context of the decision at hand. I have always wished I had a approach that had been assembled by me and other government recording of the conversation, but I did make notes to myself scientists. about it (I’m afraid I no longer have them), and I have tried to I and others at the FDA were required to evaluate these use these as a guide in my professional life. I also know that the scientific proposals, and weigh them against the counter argu- Kennedy–Upton discussion very much influenced the way I ments offered by other experts. We decided that the evidence in thought about my later role as a member of the National Aca- support of them was inadequate, and we also offered our own demies “Red Book” study.1 The conduct and uses of risk quantitative risk assessment and found that risks from detect- assessment is never to be undertaken without great care and able residues substantially exceeded those the agency consid- attention to what is known and to how well it is known. ered negligible under the “Sensitivity of the Method” The Bingham discussion, the Jensen–Jukes commentaries regulation.11 Evidence that DES had properties distinct from on low-dose DES risks, and the Kennedy–Upton discussion those of natural estrogens also influenced our thinking. Com- provide a sense of the atmosphere surrounding quantitative risk missioner Donald Kennedy, in his last official FDA act, in July assessment as the 1970s came to a close. 1979, signed a regulation prohibiting continued use of DES in animal production.4 American Industrial Health Council Perhaps the most important voice for industry during this time Lessons From 2 Superb Scientists was that of the American Industrial Health Council (AIHC), a Donald Kennedy, FDA Commissioner from 1976 to 1979, and group founded in 1977 by several major trade associations, to to whom I reported during most of that period, was an outstand- deal with OSHA’s developing cancer policy. I first encountered ing scientist (neurobiology) with a superb intellect and under- the group as a result of its comments on the IRLG efforts. The standing of the role of science in the public arena. He rapidly members and advisors of AIHC included a number of scien- developed an understanding of the problem of carcinogens and tists, and I recall many conversations and meetings with some of the essential elements of risk assessment. He provided much of them during the development of the IRLG guidelines. needed support to my own efforts and to promote the regulatory Although some members of AIHC expressed strong opposition uses of risk assessment. I recall one interaction with Kennedy to the methodologies described in the IRLG report, most Rodricks 7 favored the effort because it seemed to represent a significant found. The public discussions and debates regarding these step toward consistency in approach among federal regulatory matters were both lively and contentious. agencies18: I and my colleagues within the regulatory agencies nev- ertheless pushed ahead, primarily because the absence of AIHC supports the report’s stated objective of ensuring that any consideration of risk in decision-making (which always regulatory agencies evaluate carcinogenic risks consistently. involved setting some limit on human exposures to avoid We strongly urge that this initial step be followed up so that significant health risk) could lead in some cases to inade- a national cancer policy is developed and conflicting policies quate public health protection, in others, to unnecessarily among the regulatory agencies are minimized. restrictive limits. The debates surrounding high-to-low dose extrapolation and Although AIHC members expressed disappointment with variability were certainly critical to decisions regarding the the lack of opportunity for formal comment on the IRLG doc- preferred approaches to risk assessment, but in my experience, ument, much of what AIHC proposed did influence the initia- the most difficult issue to be resolved during this period was the tion of the Red Book effort and AIHC recommendations are question of whether quantitative risk assessment, of any kind, discussed at some length in that report. should play a role in decisions about carcinogen regulation. The proposed OSHA cancer policy, described earlier, and ulti- mately rejected by the Supreme Court, was the most visible attempt to continue hazard-based decision-making. Once con- The Foes of Risk Assessment vincing evidence of carcinogenicity became available for a It is not possible in a relatively short paper to more than briefly regulatable substance (ie, its cancer hazard was established summarize the evolution of thought regarding the conduct of with reasonable certainty), controls, and limits on exposures cancer risk assessment that made its way into scientific litera- to that substance would be established based on criteria such ture after it became clear in the mid-1970s that regulatory as “technical feasibility,” analytical detection limits, or com- officials were interested in moving to risk-based decisions. plete elimination where that was technically possible. No con- That scientific literature and the increasing numbers of confer- sideration was to be given to the magnitude of risk associated ences and workshops it engendered began to shine bright lights with the substance. on a number of scientific issues that had previously been exam- The debate over hazard as against risk-based regulation is still ined in relative isolation from one another, but which, in the with us, but it was most intense during the late 1970s perhaps evolving risk assessment context, required integration. because risk assessment had not yet been fully established as a The conduct, interpretation, and uses of epidemiological regulatory approach and many did not want to see it established. studies to identify carcinogens were the subjects of much The foes of risk assessment at the time primarily empha- increased attention, brought about in part by the initiation of sized the scientific uncertainties associated with extrapolations. the IARC Monograph program in 1971. Similar issues regard- Some focused on the high- to low-dose problem, and some on ing cancer bioassays in animals began to appear frequently in the fact that animal data were of unknown reliability for asses- scientific publications and programs, in part inspired by con- sing human hazard or risk. Umberto Saffiotti of the NCI, men- troversies such as the one, described earlier, relating to sac- tioned earlier, wrote in 1977: charin. The publication of the Ames assay in 1975 led to rapid growth in research into the utility of such assays in identifying No existing method allows us to predict precisely and reliably carcinogens and their mechanisms of action. The discussions the level of carcinogenic response in humans to chemicals and debates devoted to these topics during the 1970s did much which are known to be carcinogenic only from experimental to improve the scientific quality and reliability of evidence studies . . . relating to the identification of carcinogens and understanding how their effects were produced. Although scientific discus- Many similar statements can be found from other prominent sion of these issues and the controversies associated with them NIH scientists. continue to these days, the increased attention paid to these The issue of dose-risk modeling was a central point of the sources of scientific evidence during the 1970s and early scientific uncertainty argument. Many statistical models were 1980s laid down one of the cornerstones of risk assessment. being discussed as possible approaches to low-dose extrapola- The other cornerstones concerned dose–risk relationships and tion, and it was apparent that, although most of these models the “low-dose” problem, and cross-species and within-species could be used to “fit” the observed dose–risk data from cancer variabilities in risk. To say nothing about the issue of whether bioassays, they provided different estimates of risk, sometimes benign tumors should count! very large differences, at doses experienced by humans, which During the period when the value of risk assessment for were typically several orders of magnitude lower than the doses decisions was first given serious discussion, the scientific at which cancer responses had been observed. Because there bases for these critical issues—low-dose extrapolations and was no way to know which of these various models represented quantification of cross-species and within-species variabil- the truth about risk, there was no reliable way to decide which ities—were relatively weak and consensus was not to be predicted risk was the true one. For some critics, huge 8 Dose-Response: An International Journal differences in predicted risks meant risk assessment was a use- because approaches to them had been earlier developed for less tool for decision-making.1 other types of toxic responses. During the period leading up to the publication of the IRLG It is not surprising that many scientists frequently offered document in 1979, and in the years following that publication, I credible arguments suggesting that many carcinogens acted encountered these various criticisms of risk assessment many through mechanisms involving a threshold in the dose–risk times. (And still do, although in different contexts.) I under- relationship. Such agents, it was suggested, should be assessed stood most of these criticisms and, in fact, thought they were using the standard methods used for agents exhibiting other true. But I also thought they missed the point. I learned to forms of toxicity. paraphrase Winston Churchill’s remark about democracy being Two excellent publications in Science by Jerome Cornfield the worst form of government, except for all others. Risk on carcinogenic risk assessment and relevant dose–response assessment, with all its difficulties and uncertainties, remains models provided a clear picture of the state-of-the-science in superior to other approaches to public health protection. I shall the mid-1970s.20,21 Thresholds, concluded Cornfield, could be elaborate on this subject when discussing the National Acade- derived from the models reviewed, but only if the carcinogenic mies “Red Book” on risk assessment, but I shall make a few agent were completely deactivated prior to any initiating event. summary points here. There was no enthusiasm in our IRLG Working Group for First, it is of course the case that no methods are available to promoting the threshold concept for carcinogens; several of assess risks to human health with known accuracy, from animal us thought thresholds were likely for some carcinogens (sac- studies and even from observational studies in human popula- charin was on our minds), but thought those proposing thresh- tions that are different from the population we seek to protect. olds for carcinogens should have the burden of providing But there are ways to characterize risks that allow sufficient evidence to support such a hypothesis. The IRLG paper con- understanding of how well human populations will be pro- cluded that “any dose may induce or promote cancer.”7(p265) In tected under different risk management strategies. retrospect, I believe a better conclusion, one consistent with Second, there is no situation, short of a complete prohibition available dose–risk models, might have been that “any dose on exposure, for which we can claim that any exposure is risk- that reaches the target site could increase the risk of cancer free, whether that exposure relates to a carcinogen or to a development.” At what point that risk and the corresponding chemical displaying any other form of toxicity. In fact, the use dose became of concern is another matter. of risk assessment for carcinogens allows explicit identification Much attention during this time was focused on the Armi- of risks that are being accepted or tolerated, whereas the tage and Doll’s 1961 formulation of the multistage model of approach to establishing exposure limits intended to protect carcinogenesis.13 The model assumes that cancer originates in against all other forms of toxicity (the Lehman-Fitzhugh a cell that has undergone a series of somatic mutations, taking approach described earlier being the prototype for current place in finite steps. Each mutational stage is depicted as a approaches) provides no insight into the residual risk being Poisson process, in which transitions occur at rates that accepted, and regarding how much population protection is increase with dose in a linear fashion. Other models (including being achieved. Certainly, thresholds exist for these forms of several “time-to-tumor” models) were under much discussion. toxicity, but thresholds vary within populations, and quantita- In a now famous 1976 publication, Guess and Crump demon- tive approaches to evaluating risks for threshold agents could strated how a linear hypothesis could be incorporated into the provide some characterization of how well populations are multistage model; upper confidence limits on the linear term being protected (fractions having individual thresholds were applied to estimate upper bounds on low-dose risks. exceeded) at different labels of exposure.19 Crump later demonstrated that this mode of extrapolation pro- These and other advantages to risk-based approaches will be duced results similar to those derived from application of their covered in more detail later. These advantages depend upon the one-hit model, thought by many to be most appropriate for availability of uniform and consistent approaches to risk assess- “single-hit” carcinogenic processes.13 ment, and the 1979 IRLG guidelines set forth the principles for In 1980, I attended what I thought was one of the most achieving these needed approaches. Improved guidance on these comprehensive symposiums on health risk analysis, organized matters came with the Red Book and its sequelae. by the Oak Ridge National Laboratory (ORNL), with proceed- ings published in 1981.22 The 36 papers in the proceedings covered all aspects of risk assessment, and collectively pro- The Low-Dose Risk Problem vided, together with commentaries from the audience, an excel- For those having a generally favorable view of risk assess- lent record of the state-of-the-science in 1980; it also offered ment’s value, the dominant issue concerned models for moving perspectives on significant policy questions. from relatively high-dose observations of cancer risk to esti- With respect to dose–risk modeling, I found the broad over- mate possible risks associated with the much lower doses typi- view by Bernard Altshuler and Kenny Crump’s exposition of cally associated with exposures occurring in human the linearized multistage model to be most useful. Richard populations. Other issues requiring extrapolation—from ani- Peto’s commentary, including his deep skepticism about any mals to humans, from one human population to another, form of extrapolation and his advocacy of reliance on simple, etc—certainly arose for discussion, but received less attention data-based potency estimates, ushered in a discussion of the Rodricks 9 role of policy in model selection, based on the needs of reg- Technological Risk.”25 This article, heavily focused on safety ulators. (The role of policy in such selections is fully elaborated risks, opened up many new areas of inquiry regarding the social in the Red Book, discussed later.) Roy Albert, then at New acceptance of risk, and the important question of ensuring pub- York University Medical Center and still at this time a consul- lic safety and health without unnecessarily losing the benefits tant to the EPA’s Cancer Assessment Group, noted that “it’s of technological innovation; in the decade following Starr’s very difficult to recommend an amount of money to spend on publication, what came to be called “risk analysis” became remedial action on the basis of a potency estimate.” an organized field of study. It included not only the assessment I was also attentive to Richard Peto’s comment on Kenny of risk but also the questions of risk perception, risk commu- Crump’s presentation23: nication, public attitudes about risks of different kinds, risk– benefit and cost trade-offs, decision-making under uncertainty, If one wants to make extrapolations to get point estimates of and various risk management questions. The growing interest upper confidence limits down toward zero based on dichoto- in these broader areas of risk inspired the initiation in 1981 of mous data, I think the method you have described is obviously the journal Risk Analysis. While at the FDA in 1979, I enter- the method of choice. tained a visit from Robert Cumming of the ORNL, and he convinced me to get involved in that journal’s development. The linearized, multistage model became the EPA’s default The Society for Risk Analysis (SRA) was created in 1981, and model for cancer risk assessment, to be replaced later by the every type of threat to human safety and health is now culti- simple linear, no-threshold model proposed in 1980 by David vatable territory for members of SRA and its journal. Gaylor and Ralph Kodell. Gaylor actually commented on this An outstanding book published in 1976 by William Low- “linear interpolation algorithm” during the Oak Ridge sympo- rance, written when he was a resident fellow at the National sium. This approach yields virtually the same upper bound Academy of Sciences, and entitled “Of Acceptable Risk” was estimates of low-dose risk as does the linearized multistage my first introduction to the wider world of risk analysis. Of model, but without any assumptions regarding its biological particular interest to me was the clarity he brought to the ques- basis. Gaylor and Kodell state24: tions of risk “acceptability” and its relationship to safety26: In this paper, we have modified and developed the suggestion a thing is safe of its attendant risks are judged to be acceptable. of Mantel and Bryan. We do not extrapolate outside the experi- mental data range with the parametric model used to describe This definition raises many serious questions: (1) how do we the results in the experimental dose range. We do not use arbi- measure risks associated with the “thing,” and how good are trarily “conservative” slopes to extrapolate to lower doses, but our measurements; (2) who gets to be the judge and why should use linear interpolation to obtain an upper bound on the risk at lower dosages. The purpose of this paper is to provide a justi- the judge be trusted; and (3) why should any risk be fication for such a procedure, to provide a simple widely appli- “acceptable”? Lowrance deals with these and similar questions cable mathematical algorithm for performing low dose risk with clarity and thoughtfulness. I suspect many risk analysts assessment from dose response data, and to examine the per- would today find some of Lowrance’s views a bit antiquated, formance of this procedure on a variety of dose response curves but I think in the whole they have survived well. for toxicological data, including but not limited to Those working in the deep forests of risk assessment should carcinogenesis. always strive to achieve greater awareness and understanding of the social and policy contexts of their work. In my own case, Although I had always assumed that it was not possible to I have long believed that the methods long in place to derive claim that accurate estimates of low-dose cancer risk could be “acceptable” exposure levels for toxic agents operating through derived, the discussions at the ORNL symposium also con- threshold mechanisms, because they fail to provide any under- vinced me that upper bound estimates of low-dose risk were standing of levels of risk associated with these exposures, offer obtainable and useful for decisions. I also understood that many no opportunity for decision-makers (which would today be other decisions in risk assessments, including data review and referred to as risk managers) to evaluate and judge acceptabil- evidence weighing, selection of specific data sets for risk ity. Further, these “bright line” (safe/unsafe) models have no assessments, the value and uses of mechanistic information, value in evaluating the magnitudes of risk reduction achieved dealing with a range of issues related to biological variability, under different risk mitigation strategies, a valuable piece of estimating human exposures, and analyzing and treating uncer- information for decision-makers.19 One may question the can- tainty, all contributed to the assessment of risk and decisions cer risk assessment methodology, but it does offer quantitative resting on these assessments. Low-dose risk modeling was far measures or risks, and thus opportunity for clarity regarding from the whole story. risk acceptance. I recognize that most toxicologists are com- fortable with the traditional approach to threshold effects, but I believe a better understanding of what is required for risk Risk Analysis information to be truly useful might inspire thought regarding In 1969, Chauncey Starr of the Electric Power Research Insti- methods for quantifying risk for all forms of toxicity.27 In sum, tute published a paper in Science entitled “Social Benefit vs. distinguishing between thresholds for individuals and 10 Dose-Response: An International Journal populations, which requires recognition of the fact that individ- interests of decision-makers. To put it bluntly, risk assessors ual thresholds vary across populations, has been an overlooked were accused of making risks “disappear” if their existence problem in risk assessment that limits its utility in many cir- made decisions too difficult, and of greatly overstating risks cumstances (see the later discussion concerning the NRC’s if doing so made it easier to achieve desired outcomes. The 2009 Silver Book report). seriousness of this issue—distortion of science to achieve desired policy outcomes—during the early 1980s, when risk assessment was becoming increasingly prominent in many reg- The Red Book ulatory domains, is difficult to exaggerate.1 As I was becoming I left my position at the FDA in mid-1980, recognizing that a practitioner of risk assessment and a consultant to the EPA on quantitative risk assessment had not attracted large numbers of some new applications of risk assessment, particularly in the enthusiastic supporters at the agency. I was pleased to note that area of Superfund and hazardous waste remediation decisions, I under the leadership of Alan Rulis, Ron Lorentzen, and a few witnessed firsthand many often-bitter skirmishes relating to the others, the agency had found value in risk assessment for estab- undue influence of policy in risk assessment results, and the so- lishing criteria for acceptance of unavoidable residual levels of called “cherry-picking of data” problem. carcinogens in, for example, food-packaging materials. Mean- Congressman Donald Ritter, Republican of Pennsylvania, while, the EPA was plowing ahead and applying quantitative with a doctorate in science from MIT, introduced the Risk risk assessments in a range of areas, including the difficult Analysis Research and Demonstration Act of 1982. The pro- problems associated with remediation of Superfund and other posed Act had, I thought, many attractive features, and I testi- hazardous waste sites. The OSHA, too, began a program of fied in its favor at a hearing in 1982. Among other things, it regulating workplace carcinogens based on application of risk called for a federal study, organized by the Regulatory Affairs assessment methods according to the mandate of the Supreme Office of OMB, of risk analysis and its applications across the Court’s benzene decision. One interesting feature of the federal government, to include recommendations for research OSHA’s efforts concerned its decisions to tolerate risks for to improve risk assessment, and various projects to demonstrate occupational carcinogens significantly greater than the EPA its value and applications. Ritter’s bill went nowhere but was found acceptable for carcinogens in the environment. heavily discussed in the scientific and trade press and had many I have not mentioned one important development that features which, I suggest, would be valuable to consider even occurred while I was still employed at the FDA and that was today. I will note that the several scientists and advocates who conceived and executed by Gilbert Omenn of the President’s testified alongside me during Ritter’s hearing expressed a wide Office of Science and Technology Policy (OSTP). Omenn was range of views, some harshly critical of efforts to enlarge the a physician and geneticist who came to a high position in the application of risk assessments in regulatory discussions. The White House in 1977 and remained there until 1982. During the impassioned remarks of Nicholas Ashford, then a professor in period in which I was much involved in the IRLG program, I the Technology and Law Program at MIT, particularly caught met with Omenn and his staff to fill him in on our activities and the direction of our work. Omenn, I learned, was supportive of my attention. Ashford forcefully argued that risk assessment, a quantitative approach to risk assessment and had underway because of its highly technical nature, was simply a device for the development of a report on the topic. Omenn wisely saw an slowing down and delaying regulatory decisions. This argu- important role for the OSTP in ensuring that federal govern- ment has much power, and is still heard today, as I shall note ment agencies maintained strong scientific support for their in the closing section. activities and that consistency across those agencies in the use “Risk Assessment in the Federal Government: Managing of scientific information was also achieved. He realized that the Process” was published in 1983 by the Committee on the risk assessment was becoming central to many areas of Institutional Means for Assessment of Risks to Public Health of decision-making, and he and his staff published, in 1978, an the National Research Council.1 The Committee was chaired important report on the topic. by Reuel Stallones (“Stoney”), an epidemiologist at the School The report focused on the content of risk assessment but was of Public Health of the University of Texas at Houston. I was not highly directive on the specific methods to be used. It invited to serve on the Committee and I was more than eager to emphasized the need for transparency and consistency, and also do so. The committee had only 14 members and included set forth the content of and distinction between the scientific health and social scientists and policy and regulatory experts, assessment process and the development and implementation who came from positions in academia, industry and consulting, of regulatory approaches.28 Much of the debate that arose dur- nongovernmental agencies, and government research institu- ing the early 1980s concerned what came to be called the tions. A few members had extensive experience in risk assess- “separation” of science and policy. Many critics of regulatory ment, but most, including the Chair, did not. But those without practices claimed that the science behind regulation was too specific experience in risk assessment brought to the Commit- often twisted to achieve predetermined policy (or “political”) tee deep knowledge and understanding of science and the role outcomes. The most common criticisms related to alleged of science in forming public policy. The Committee was well biases in the selection of data and models in risk assessments, served by a superb staff, headed by Lawrence McCray. McCray designed to achieve results that would best serve the political was a superb social scientist who provided much needed Rodricks 11 guidance and support to our group; he went on to become a according to their relative scientific merits, and the option to faculty member at MIT. be used—across all risk assessments—is selected. The selec- It may seem odd that a Committee assigned to examine risk tion involves an unavoidable policy choice. It is not the kind of assessment was not overloaded with risk assessment experts. I policy choice associated with risk management; it is rather a do not intend to repeat here the major findings of the 1983 Red “science policy” choice. These “inference” options and their Book, already discussed thousands of times in the literature, but selection do not apply only to carcinogen risk assessment but to I will emphasize a matter that is not often discussed—the all assessments of toxicity risks. The Committee offered gui- objectives the Committee was asked to fulfill. The Committee dance on how the inference options to be used, at each step of was formed in response to a congressional directive to: risk assessment where they are needed, was best selected. Uncertainties are inherent in the process and are to be explicitly (i) Assess the merits of separating the analytic functions of considered. developing risk assessments from the regulatory func- The second area of major importance is the Red Book’s tions of making policy decisions. efforts on risk assessment guidelines. Guidelines on the con- (ii) Consider the feasibility of designating a single institu- duct of risk assessment were necessary to ensure consistency tion to do risk assessments for all regulatory agencies. and uniformity in approach across all agencies. The guidelines (iii) Consider the feasibility of developing risk assessment should, according to the Red Book Committee, include guidelines for use by all regulatory agencies. approaches for identifying the data and methods to be used in risk assessment, but also the options for making inferences These objectives all relate to the raging debates I have beyond the data and the choice of options to be used (“the described pertaining to the alleged improper incursions of pol- defaults”). The Red Book Committee emphasized that, in spe- icy makers into the conduct of risk assessment or the tendencies cific cases, in which well-developed scientific support became of risk assessors to offer up risk findings they believe policy available, it should be possible to deviate from the guidelines. makers would prefer to have. The “solutions” to this problem In effect, scientific data might become available in specific envisioned in the Committee’s legislative mandate can be read- cases to reduce the need for the defaults. ily inferred from these objectives. Risk assessments might be These 2 critical recommendations were developed to deal kept “pure” by ensuring the complete insulation of risk asses- with the Committee’s charge regarding institutional separation sors from regulatory policy makers, perhaps even by creating a of risk assessment activities from the regulatory decision- “National Institute of Risk Assessment” (see (ii) above). The making agencies. The Committee believed that the use by all guideline questions indicate the value of having “blueprints” all regulatory agencies of uniform guidelines which contained risk assessors must follow, without interference from decision- agreed-upon inference options and defaults would be adequate makers. These were the questions the Committee was asked to to resolve the problem of case-by-case interjections of policy consider. It was not asked to offer opinions on how risk assess- biases into the risk assessment process. Separating scientists ments were to be conducted, and it did not do so. But it did involved in risk assessment from those who had to use assess- make many important recommendations. ments for decisions would, the Committee thought, lead to Almost all public discussions about the Red Book begin and serious inefficiencies in communication and lack of clarity end with a presentation of its risk assessment framework and regarding the specific problems to be examined and addressed. the 4 steps of risk assessment. This was no doubt important Assessors and managers should be able to engage in useful (and and helped to advance the field but resolved no significant controversies. Two other sets of recommendations and their necessary) discussion without undue influence of the managers support did help to resolve controversies, but their importance on the conduct of risk assessments; guidelines for risk assess- in advancing the field has not been given enough attention. ment would be essential to ensure the success of this approach. First, the Committee made it clear that risk assessments No separate “risk assessment agency” should be considered. could not be completed without the inclusion of certain The 3 major recommendations of the Red Book Committee, “science policy” decisions to deal with ever-present gaps in out of a total of 10, were set forth in the Red Book’s summary. data and, most importantly, basic knowledge. It was critical, The third major recommendation, about which I have written the Committee noted, that the best possible scientific basis be elsewhere,29 was directed at the Congress and urged the devel- developed for decisions aimed at public health protection and opment of a Board on Risk Assessment methods. The Board that risk assessments were the necessary basis. But unless would track developments in risk assessment and periodically uncertainties were dealt with, risk assessments could never revise inference guidelines, study the usefulness of agency be completed—a completely unsatisfactory outcome. approaches, and identify research needs. This important rec- At many steps of risk assessment, including the all- ommendation, if followed, would have done much to help important dose–risk relationship in dose ranges for which ensure the efficient and scientifically adequate functioning of empirical data are generally unavailable, inferences beyond the risk assessment, both within and outside of government, on a available data must be made. The options available to make continuing basis. these inferences are to be examined using the best available methods of science, they are to be ranked, if possible, No one paid attention to this recommendation. 12 Dose-Response: An International Journal Aftermath implementing guidelines.30 I became aware of the important role of states in chemical regulation when I was invited by Bernard Goldstein became chief scientist at the EPA at about James Solyst, soon after the Red Book’s release, to brief the the time the Red Book was released. He was an enthusiastic board of the National Governor’s Association on its purpose supporter of its findings and recommendations and wanted to and content. I also attended a meeting of state agency scientists see recommendations followed at his agency. Goldstein also and administrators held at Times Beach, Missouri, a site of understood that risk assessment was both controversial and significant dioxin contamination that had received national not well understood. I was at that time advising the EPA’s attention. The Red Book was a major topic of discussion at the Superfund office on how risk assessment might be applied to meeting, and it was clear that some states were ready to imple- decisions about site remediation, and one day ended up in ment its recommendations with high enthusiasm. Goldstein’s office, along with other senior EPA scientists. My most memorable encounter at the meeting was with Goldstein had in mind an ambitious program directed at train- Thomas Burke, then a senior regulatory and public health offi- ing stakeholders from other federal and state agencies, regu- cial in New Jersey. Tom has, for many years, been on the lated industries, NGOs, and academics who were serving as faculty of the Johns Hopkins Bloomberg School of Public advisers to the EPA and related agencies. The list of stake- Health and serves as director of the School’s Risk Sciences holders Goldstein had in mind included not only experienced and Public Policy Institute. At our very first meeting in Times scientists but also scientists with little experience and policy Beach, I found Tom to be not only a quick learner but someone and legal experts. The program was implemented and I recall with significant understanding and foresight regarding the a 2-day training exercise (I think at a hotel in or near Anna- importance of risk assessment, not only to regulation, but in polis, sometime in 1984) for about 120 people. the broader world of public health science and policy. Tom has I was asked to help with the training, developed an exercise had major roles in many National Academies efforts, the most based on a hypothetical carcinogen (dinitrochickenwire notable of which was his role as Chair of the Committee that [DNC]), and invented data regarding its hazards, dose–risk produced the extraordinary 2009 report “Science and Deci- relationships, and human exposures. Trainees worked in groups sions: Advancing Risk Assessment.” The recommendations of about 8 people and were required to work through the data of this report have not received sufficient attention, but I con- and various options for evaluating the data and reaching con- tend that if risk assessment has a future, much of that future is clusions about DNC’s risks. Each of the 20 or so groups was to be found in this report.19 then required to present its conclusions to the entire group of Tom served for about 2 years under the Obama administra- trainees. The exercise was quite well received, and most people tion as the EPA Science Advisor and Deputy Assistant Admin- left the meeting with substantially improved understanding of istrator for Research and Development. I have been a friend of the content of risk assessment, the significant impediments to Tom since the Times Beach meeting and agreed to meet him achieving highly certain results, and even regarding the borders for a drink the day after the presidential election of 2016. Like between assessment and management. (I have been involved in most Americans we had been shocked by Trump’s victory. We offering the DNC exercise to many different groups, most also knew that Tom would not remain much longer at the EPA recently in China and Australia.) Bernard Goldstein deserves and we thought that Trump’s plan for the agency was likely to much credit for creating an environment in which risk assess- be destructive. We had our drink, but it was a gloomy evening. ment would be recognized as an important basis for regulatory We had our drink at the Willard Hotel, where Abraham and similar public health decisions. Lincoln had stayed when he arrived in Washington after his The publication of the Red Book generated immense interest election to the Presidency. in risk assessment, and I and many others were involved for a year or more in seminars and workshops devoted to the subject. The EPA got busy developing the recommended guidelines and Some of the Missing Parts, Including Ed also incorporating them into its many regulatory programs. Calabrese While the publication of the Red Book brought much clarity There are many omissions in my story, both because there were and understanding and opened many pathways for its applica- during the decade I cover many activities I had little or no hand tion, controversies over the conduct of risk assessment did not in, and also because I had to place some limits on what I chose vanish. This is not in the least surprising, because uncertainties to relate. Readers should not, therefore, assume that what I are ever present in risk assessment, and scientific debate have offered is anything like a complete history of this impor- regarding them is not only expected but also essential. The tant time. In addition, I have omitted the contributions of many existence of guidelines, including the establishment of infer- individuals, including those who were important influences on ence options and default assumptions, certainly placed some my own thinking. bounds on these debates, and this was necessary to allow agen- I am not an adequate writer to convey the nature and inten- cies to complete assessments without endless debates over sities of the many debates and disagreements that accompanied scientifically irresolvable issues. my journey, but I must say I found most civil and good- By the time of the Red Book’s publication some state agen- intentioned. I did suffer a few ad hominem attacks, interestingly cies had undertaken efforts to develop their own policies and from some traditional toxicologists who thought I was both Rodricks 13 trying to wreck their traditions and that I was inadequately publication of the Red Book, a long series of reports relating educated in those traditions, but these were not at all to risk assessment and other elements of risk analyses.19 Most representative. of these reports have been sponsored by the EPA, but several I do want to offer a few observations about the work of Ed other agencies of the federal government have also sought Calabrese, a scientist well-known to the readers of this journal advice from the NASEM. A number of these reports have been and to the world’s community of dose–response scientists. Ed, devoted to identifying approaches to improve the practice of in a series of extraordinary review articles, has researched and risk assessment, and many focus on new applications. What can documented serious concerns about the scientific basis for the be found in all of these reports is a commitment to the princi- assumption of linearity at low dose, first in connection with ples and analytical frameworks first laid out in the Red Book. radiation-induced cancer and then in connection with its adop- The 2009 report “Science and Decisions” does create a new tion, by the EPA in particular, for chemically induced cancers. and valuable decision-making framework, in which the con- Calabrese has painstakingly reviewed many publications on duct of risk assessment follows a “problem formulation and radiation-induced mutagenicity and found that much of it falls scoping” exercise, designed to ensure that the risk assessment short of providing evidence to support a relationship between will be directed at the questions important to ultimate somatic mutation theories of carcinogenesis and low-dose lin- decision-making. The committee that produced this report earity for carcinogens, whether induced by radiation or by recognized that many programs lacked emphasis on maximiz- chemicals. ing the utility of risk assessments for ultimate decisions. This I have not devoted sufficient effort to provide an adequate and other recent NASEM reports have also placed emphasis appraisal of Calabrese’s work, but it does deserve, I suggest, far on ensuring that scientific uncertainties, inherent to all risk more attention from the mainstream risk analysis community assessments, are adequately described and considered in the involved in regulatory and public health practice. Calabrese risk management process.32 draws from his work the conclusion that what he terms “the In addition to a wealth of guidance from committees of flawed acceptance of linearity at low dose” by a 1977 Com- NASEM, there are of course studies from many other institu- mittee of the National Academies that was providing guidance tions, both governmental and nongovernmental. One can read- to the EPA on carcinogenic contaminants of drinking water31 ily perceive that much of this guidance has penetrated the was of signal importance in driving regulators to the low-dose major risk assessment programs of federal and state govern- linearity model. This report may have been important to the ment, not perfectly but nevertheless usefully and effectively. EPA, but in my experience and as suggested in this article, In my view, however, significant issues that arose during there were many other forces at work that moved regulators the decade about which I have written remain unattended or in this direction. inadequately addressed. Some have been the subject of expert Calabrese is also apparently dismayed that precautionary studies; but actual practice remains largely unaffected. I note policies have been at play in some of the defaults used by 3 such issues. regulators. But, as I have tried to make clear in this article, such policies are inevitable when science is uncertain and decisions 1. Guidelines. I have emphasized the need for continuing have to be made. I do agree with Calabrese’s contention that review and updating of risk assessment guidelines. errors in science should be corrected when they are found and The EPA has been, by far, the most diligent in produc- that there may have been inadequate attention to some of the ing guidelines, but even that agency’s efforts have scientific foundations for “one-hit” and “irreversibility” and fallen short, and other important agencies have pro- other hypotheses that moved carcinogen risk toward low-dose duced little. The difficult but essential issue of justify- linearity and that “science policy” (as in the Red Book) should ing default assumptions, and criteria for moving away not be decided until what science can and cannot tell us has been from them in specific cases, has received insufficient made clear. But I do suggest that it would take an enormous attention. Without adequate guidelines, the production effort to revisit, verify, or reject all of the science upon which of risk assessments having adequate transparency is concepts of thresholds and low-dose linearity are based, and it is hampered. Risk assessments also become subject completely unclear how such an effort might be accomplished unnecessarily to controversies that should be resolved (of course, this is not a problem Ed should be asked to solve). within guidelines, and so their timely production is Nevertheless, Calabrese’s work is powerful and reflects the impeded. As I noted earlier, the failure of government determination to uncover facts that, were they found to be true, institutions to support the Red Book’s recommenda- would upset decades of risk assessment work and regulation. tion regarding a standing Board on Risk Assessment Methods has been an impediment to progress. 2. Quantification of Risk. Cancer risks are generally sub- A Bit of Summation jected to quantitative assessment, but quantification of risk for all other forms of toxicity has not been signif- Finished and Unfinished Business icantly pursued. Instead, “bright-line: models,” such as What is now called the National Academies of Science, Engi- those represented by the RfD, ADI, and TDI, are the neering and Medicine (NASEM) has issued, since the norm. These approaches provide no guidance to 14 Dose-Response: An International Journal decision-makers on the risks associated with agents began to achieve recognition remains in force. It seems acting though threshold mechanisms or to agents other unlikely that the risk-based requirements of our laws than carcinogens acting through nonthreshold mechan- will be reversed, so that risk assessment will remain a isms. Methods for probabilistic assessments have been priority. The arguments for the much simpler hazard- suggested by the WHO/International Program on based approaches are, however, appealing in many Chemical Safety and demonstrated by others.27,33 The ways and will likely influence the policies and deci- “Science and Decisions” report mentioned earlier sions that can be taken on a voluntary basis. Although called for similar quantification efforts, heavily depen- the 2 approaches do conflict on a scientific level (sim- dent upon modes of toxic action. Probabilistic plistically, one divides the world of chemicals into approaches are both more revealing regarding the risks “the toxic” and “the non-toxic;” and the second into that remain under different exposure scenarios and are “the risky” and “the not so risky”), it is a conflict that useful (necessary?) for the many decisions requiring may not require a resolution unless it leads to unfore- trade-offs of one kind or another. Resistance of many seen and unacceptable consequences. toward probabilistic assessments for all forms of toxi- city remains strong; indeed, the matter is hardly I am highly grateful for having been given the opportunity to discussed. offer my recollections and perceptions of risk assessment’s rise during the period I have covered, and I hope this subjective Risk assessments should contain assessments of risk! historical summary will somehow help support continued review and improvement of assessment and its uses. 3. Risk-Based Decision-Making. I earlier described Many of us have concerns that, at present, the federal gov- some of the opposition to risk-based decision- ernment does not seem interested in devoting efforts to making that emerged from important stakeholders improve and foster science-based decisions. But risk assess- during the decade when risk assessment rose to pro- ment is now a worldwide project that should see continuing minence. The alternative approach focuses primarily support. And interest in science-based policies is certainly not on the identified hazards associated with a chemical forever lost in the United States. I was extremely fortunate to be (the type of toxicity it can cause) and seeks to reduce present at the creation and hope to be able to witness the con- exposures to substances having what are judged to be tinuing incorporation of these types of innovative approaches particularly serious hazardous properties (eg, carci- to both data development and risk assessment, much discussed nogens, reproductive and developmental toxicants, now among scientists from an extraordinary range of disci- endocrine-disrupting substances). In many cases, plines. Maybe policy-free (default-free) risk assessment is too total elimination of the hazardous substance is sought much to hope for, but current research tools and methods seem and in others reductions to the lowest levels thought to point to that possibility. to be reasonably achievable; in neither case is the risk posed by the substance considered. Declaration of Conflicting Interests Arguments for hazard-based decisions typically The author(s) declared no potential conflicts of interest with respect to point to the uncertainties associated with risk assess- the research, authorship, and/or publication of this article. ments, the fact that risk assessments require much more data (eg, on dose–response and exposure) and Funding take much more time to complete. These powerful The author(s) received no financial support for the research, author- arguments do, however, seem counter to the fact that ship, and/or publication of this article. almost all laws governing chemical regulation call for risk-based approaches. They also seem counter to the Note fact that hazard-based approaches can often lead to the 1. The strict prohibition of the Delaney Clause continued to apply, as unnecessary elimination of valuable substances with- it does today, to any intentionally and directly added food sub- out achieving significant health benefits. Introducing stance that is carcinogenic. new chemicals to replace eliminated chemicals also raises the possibility of introducing unanticipated References risks. The hazard-based approach to decisions has 1. NRC. Risk Assessment in the Federal Government: Managing the many advocates and is most seen in the voluntary Process. Washington, DC: Author; 1983. actions of manufacturers and users of targeted sub- 2. FDA. Assessment of Estimated Risk Resulting from Aflatoxins in stances because risk-based laws and regulation do not Consumer Product Peanut Products and Other Contaminants. apply in such circumstances. Substitutions to replace Rockville, MD: Author; 1979. chemicals in products can be perilous if not carefully 3. Herbst AL, Ulfelder H, Poskanzer DC. Adenocarcinoma of the done, with proper attention to risk. vagina. N Engl J Med. 1971;284(16):878-881. Thus, the question of risk-based approaches which I 4. Zervos C, Rodricks JV. FDA’s ban of DES in meat production. encountered in the years during which risk assessment Am Stat. 1982;36(3):278. Rodricks 15 5. 21 U.S.C. Federal Food, Drugs, and Cosmetic Act. Food Addi- 20. Cornfield J. Carcinogenic risk assessment. Science. 1977; tives Amendment of 1958. 198(4318):693-699. 6. Lehman A, Fitzhugh O. 100-fold margin of safety. Quarterly 21. Cornfield J. Models for carcinogenic risk assessment. Science. Bulletin. 1954;18:33-35. 1978;202(4372):1107-1109. 7. IRLG (Interagency Regulatory Liaison Group). Scientific bases 22. Richmond CR, Walsh PJ, Copenhaver ED. Health risk analysis. for identification of potential carcinogens and estimation of risks. In: Proceedings of the Third Life Science Symposium. Philadel- JNCI J Natl Cancer Inst. 1979;63(1):241-268. phia, PA: Franklin Institute; 1980. 8. Mantel N, Bryan WR. “Safety” testing of carcinogenic agents. 23. Peto R. See Ref.22:391. JNCI J Natl Cancer Inst. 1961;27(2):455-470. 24. Gaylor DW, Kodell RL. Linear interpolation algorithm for low 9. Rodricks JV. Regulation of carcinogens in food. Ann N Y Acad dose risk assessment of toxic substances. J Environ Pathol Tox- Sci. 1981;363(1):29-35. icol. 1980;4(5-6):305-312. http://www.ncbi.nlm.nih.gov/pubmed/ 10. Hutt PB, Merrill R. Regulation of carcinogens. In: Food and Drug 7217854. Accessed August 17, 2018. Law. Westbury, NY: The Foundation Press; 1991:863-963. 25. Starr C. Social benefit versus technological risk. Science. 1969; 11. FDA. Chemical compounds in food-producing animals; criteria 165(3899):1232-1238. http://www.ncbi.nlm.nih.gov/pubmed/ and procedures for evaluating assays for carcinogenic residues. 5803536. Accessed August 17, 2018. Fed Regist. 1977;42:10412-10429. 26. Lowrance WW. Of Acceptable Risk: Science and the Determina- 12. Albert RE, Train RE, Anderson E. Rationale developed by the tion of Safety. Hebden Bridge, England: Will Kaufman; 1976. Environmental Protection Agency for the assessment of carcino- 27. WHO (World Health Organization). Harmonization Project Doc- genic risks. J Natl Cancer Inst. 1977;58(5):1537-1541. ument 11 Guidance Document on Evaluating and Expressing 13. Crump KS, Hoel DG, Langley CH, Hod DG. Fundamental carci- Uncertainty in Hazard Characterization. Geneva, Switzerland: nogenic processes and their implications for low dose risk assess- World Health Organization; 2014. ment fundamental carcinogenic processes and their implications 28. OSTP (Office of Science and Technology Policy). Potential for low dose risk assessment. Cancer Res. 1976;36(9 pt 1): Himan Carcinogens: Methods of Identification and Characteriza- 2973-2979. tion. Part 1; USA: US Government’s Office of Science and Tech- 14. OSHA (Occupational Safety and Health Administration). Pro- nology Policy; 1978. posed rule: identification, classification and regulation of poten- 29. Rodricks JV. What happened to the Red Book’s second most tial occupational carcinogens. Fed Regist. 1977;42:54148. important recommendation? Hum Ecol Risk Assess. 2003;9(5): 15. OSHA (Occupational Safety and Health Administration). Final 1169-1180. doi: 10.1080/10807030390240373. rule: identification, classification and regulation of potential 30. State of California - Health and Welfare Agency. Sacramento, occupational carcinogens. Fed Regist. 1980;46:5001. CA. October, 1982. Carcinogen Identification Policy: A State- 16. Industrial Union Dept. v. Amer. Abu Dhabi, UAE: Petroleum ment of Science as a Basis of Policy; Section 2: Methods for Institute; 1980. Estimating Cancer Risks from Exposures to Carcinogens. 17. OSHA (Occupational Safety and Health Administration). Identi- 31. National Academy of Sciences. Drinking Water and Health. fication, classification and regulation of potential occupational National Academy of Sciences: Washington, DC; 1977. carcinogens. 1982;47:187. 32. IOM (Institute of Medicine). Environmental Decisions in the 18. AIHC (American Industrial Health Council). Recommended Face of Uncertainty. Washington, DC: National Academies Alternative to OSHA’s Generic Carcinogen Proposal. Scarsdale, Press; 2013. NY: Author; 1978. 33. Chiu WA, Slob W. A unified probabilistic framework for dose– 19. NRC (National Research Council). Science and Decisions. response assessment of human health effects. Environ Health Washington, DC: National Academies Press; 2009. Perspect. 2015;123(12):1241-1254.