The Case for Public Access to Federally Funded Research Data

DOCUMENT RESUME

ED 450 184 UD 034 005

AUTHOR Gough, Michael; Milloy, Steven TITLE The Case for Public Access to Federally Funded Research Data. Policy Analysis. INSTITUTION Cato Inst., Washington, DC. REPORT NO PA-366 PUB DATE 2000-02-02 NOTE 22p. AVAILABLE FROM Cato Institute, 1000 Massachusetts Avenue, N.W., Washington, DC 20001. Tel: 800-767-1241 (Toll free); Fax: 202-842-3490; Website: http://www.cato.org. PUB TYPE Reports Descriptive (141) EDRS PRICE MF01/PC01 Plus Postage. DESCRIPTORS Asbestos; Environmental Research; *Federal Regulation; Government Role; Public Policy; *Scientific Research IDENTIFIERS Endocrine Changes; Environmental Protection Agency; Food and Drug Administration; National Cancer Institute; National Institutes of Health; *Public Awareness

ABSTRACT This study examines the importance of public review of federally funded scientific research by looking at several case studies. It shows that independent, nongovernmental review of federal scientific research has had a major positive effect on knowledge in many areas. The study focuses on: the Environmental Protection Agency and airborne asbestos; the panic over endocrine disrupters; the National Cancer Institute and the herbicide 2,4-D; the National Institutes of Health and the Dalkon Shield birth control device; and the Federal Drug Administration and the diet drug fen-phen. In many of those cases, third party review served to correct or prevent costly regulatory mistakes. In some cases, however, independent review of federally funded science occurred too late to prevent significant economic and consumer harm. The report considers the political, regulatory, and theoretical issues surrounding the Shelby Amendment, which requires federal agencies to ensure that all data produced under a grant be made available to the public through procedures established under the Freedom of Information Act. It concludes that the amendment, if fully complied with by federal agencies, will improve the quality of federal scientific research, and accordingly, the quality of federal regulation. (SM)

Reproductions supplied by EDRS are the best that can be made from the original document. 1 0

THE CASE FOR PUBLIC ACCESS TO FEDERALLY FUNDED RESEARCH DATA Policy Analysis, No. 366 Michael Gough and Steven Milloy CATO Institute February 2, 2000

U.S. DEPARTMENT OF EDUCATION Office of Educational Research and Improvement EDUCATIONAL RESOURCES INFORMATION CENTER (ERIC) This document has been reproduced as received from the person or organization originating it. Minor changes have been made to improve reproduction quality.

Points of view or opinions stated in this document do not necessarily represent official OERI position or policy. BEST COPY AVAILABLE

11) PERMISSION TO REPRODUCE AND DISSEMINATE THIS MATERIAL HAS C BEEN GRANTED BY G.C.R10)r 2 TO THE EDUCATIONAL RESOURCES INFORMATION CENTER (ERIC) 1 O

No. 366 February 2, 2000 Routing

The Case for Public Access to Federally Funded Research Data by Michael Gough and Steven Milloy

Executive Summary

Advocates of laws, regulations, and policiesthe provisions of the Freedom of Information often use scientific data and analyses to advance Act, grantees to make data that result in a pub- their claims. In the normal course of science,lished report or that are cited in a federal rule controversies about data and analyses areor regulation available to members of the pub-. resolved by independent researchers who lic on request. attempt to replicate the data and redo the analy- Although the Shelby Amendment has drawn ses. To facilitate independent review, the scientist criticism from many policy activists and scien- who produces the data and analyses is generallytists, public review of data and methodology is expected to disclose his data and his methods to crucial for both good science and good public potential reviewers. policy. Scientific data collected by federal agen- During the past three years, access to scien-cies have often been subjected to independent tific data collected by federal grantees hasreview and found to be in error. Scientific become a major political issue. For instance,research undertaken by nongovernment scien- controversial scientific evidence was used totists and financed with public money should justify onerous regulation of particulate mat- similarly be available for review by public watch- ter (a prominent air pollutant) and urbandogs to ensure that any new laws and regulations smog, but when Congress requested the con- based on the research are merited. If the research troversialdata,the granteesrefused.In is soundly grounded, then independent review October 1998, Congress passed, and the presi-will underscore the merits of those laws and reg- dent signed, Public Law 105-277, known as the ulations. If not, then independent review will Shelby Amendment. The law requires, throughhelp society avoid costly policy mistakes.

Michael Gough is an adjunct scholar at the Cato Institute. A biologist by training, he is the author of Dioxin, Agent Orange (1986) and a former staff member at the National Institutes of Health and the congressional Office of Technology Assessment. Steven Milloy, also an adjunct scholar at the Cato Institute, is a lawyer and the publisher of the Junk Science Home Page, www.junkscience.com. He is the author of Science without Sense: The Risky Business of Public Health Research (1995). 11 NSTMITE

3 Science flourishes Breast Implant Case,3 said, "Many reports are in an atmosphere The Nature of Science trivial, many more are just plain wrong, and a lot of what we're told is inconsistent."4 The of free exchange The late Sir Karl Popper delved deeply into accepted way to sort through reports of dif- of information. questions about the nature of science andferent value is through the process of critical how it works. He concluded that the scientif- review and attempted replication. ic process, for all its accouterments of math, There's nothing shameful about a mis- instrumentation, and specialized knowledge, take, but scientists seek to avoid incorporat- can be divided into two parts. The first part is ing mistakes into accepted science. Addi- the formulation of an idea, hypothesis, or the- tional ideas and hypotheses that are based on ory (the terms may be used interchangeablymistakes are almost certain to be wrong, and for present purposes) about how some part ofthe time and effort expended on developing the physical universe works. The second part and testing those ideas and hypotheses are is the design and execution of an experiment lost. It is far better to review, analyze, and or test to discover whether the idea, hypothe- attempt to replicate a new finding before sis, or theory is corroborated or falsified. If itaccepting it. is corroborated, then the idea or hypothesis Scientists have developed myriad methods may be incorporated into scientists' knowl- for review. Scientists are expected to talk edge of the universe, and it can be used in theabout their results and insights with their construction of other ideas and hypotheses.' peers, and scientists who don't soon pass into The distinguishing mark of science, accord- obscurity, along with their work. ing to Popper, is the formation of testable The primary method of dissemination hypotheses. and review of scientific results is through the The physicist Paul Davies underscored thepublication of papers in scientific journals. importance of hypothesis testing: Although journals have varying standards for review of papers submitted for publication, A powerful theory is one that is high- those with the most rigorous review process- ly vulnerable to falsification, and so es tend to be the most prestigious. it can be tested in many detailed and Nevertheless, no matter how complete specific ways. If the theory passes any published information is, a scientist those tests, our confidence in the interested in replicating or examining theory is reinforced. A theory that is another scientist's work may have to call or too vague or general, or makes pre- write for additional details. There's no law dictions concerning only circum- against scientists' refusing to cooperate stances beyond our ability to test, is with requests for additional information; of little value.' however, scientists who block the flow of information risk their reputations and In the process of testing hypotheses, scien-undermine the credibility of their work. tists produce data, usually measurements or Science flourishes in an atmosphere of counts. To obtain recognition for theirfree exchange of information.5 Of course, hypotheses and data, scientists must publishsome scientistsby their choice of employ- their findings, or "show their work," so thatmentcut themselves off from completely other scientists can examine and criticize open science. Scientists who work for defense those findings. companies, for example, can share informa- Scientists, like everyone else, can maketion only on a need-to-know basis, and mistakes. As Marcia Angell, a physician and industry scientists may not participate open- executive editor of the New England Journal ofly in scientific exchanges because of concerns Medicine and author of Science on Trial: Theabout patents, trade secrets, and other intel- Clash of Medical Evidence and the Law in the lectual property.

4 Government-Funded Science to have access to the published results of an Scientific work that is used to supportepidemiologic study, because the underlying regulations, as is scientific work that is per-data sets can be too large to be incorporated formed for industry, is done for a client. Theinto normal-length scientific articles. Similar industrial client may be more interested inproblems with large data sets are common in using the science to foster the development economics and policy studies, and journals in of a product than in examining the sound-those disciplines often require the authors of ness of the science, and the governmentthe papers to tell readers where the entire client more interested in justifying a regula- data sets are available. tion. Unlike a business firm, which can face a This study examines the importance of disaster if it prematurely rushes a product topublic review of federally funded scientific market, state regulators pass on to the publicresearch by examining a number of case stud- the losses entailed by poor regulation. Thus,ies. In the next six sections, we show that because regulatory agencies are rarely penal-independent, nongovernmental review of ized for erroneous science, they are less moti-federal scientific research has had a major vated to ensure that the science they use ispositive effect on our knowledge about air- valid. This leaves the process of checking theborne asbestos, endocrine disrupters, the her- science, if it is to be done at all, to the public.bicide 2,4-D, the Dalkon Shield birth control Because regulato- A provision in Public Law 105-277, knowndevice, and the diet drug fen-phen. In many ry agencies are popularly as the Shelby Amendment andof those cases, third-party review served to rarely penalized enacted by Congress in October 1998, guar-correct or prevent costly regulatory mistakes. antees, through the Freedom of InformationIn some cases, however, independent review for erroneous sci- Act, public access to grantee-collected dataof federally funded science occurred too late ence, they are less that are used in support of rules or regula-to prevent significant economic and con- motivated to tions. The history of public review of federal-sumer harm. ly collected data and analyses reveals the In the final section, we consider political, ensure that the importance of such review in correctingregulatory, and theoretical issues surround- science they use is errors made by government officials. ing the Shelby Amendment. In sum, we conclude that the amendment, if fully complied valid. This leaves Data Used to Formulate Public Policy with by federal agencies, will improve the the process of Government data used to set regulationsquality of federal scientific research and, checking the sci- and to issue warnings and announcements accordingly, the quality of federal regulation. about various products are of two types: (1) ence, if it is to be laboratory data and (2) epidemiologic data. done at all, to the In general, it's possible to conduct additional The EPA and experiments and to attempt replication of Airborne Asbestos public. laboratory studies. It is impossible, however, to replicate epi- In the last quarter century, "asbestos" has demiologic studies, which provide informa-become synonymous with cancer, premature tion about unique populations of people, death, long-running lawsuits brought by for- subject to unique conditions, over particular mer asbestos workers against the companies time periods, but the data from such studies that mined and supplied asbestos, and billions can be reviewed and reanalyzed by indepen- of dollars in settlements for disease and death. dent scientists. The only way to determineThat perception of asbestos is far different the validity of those studies is through care-from the one that prevailed during World War ful analysis of the techniques that were usedII, when asbestos was used in great amounts in to produce the data and through carefulthe construction of both warships and mer- review of the analytic methodology adoptedchant ships, which were subject to airplane by the study's authors. Often it's not enoughand submarine attacks. Asbestos saved the

5 lives of many sailors. The building boom after for in the previous documents. EPA has never World War II found more uses for asbestosexplained the reason for its reversal, but, fireproofing, insulation for heating and cool- from the record, it is clear that entities that ing systems, soundproofingand it was used bore the costs of asbestos removalformer in buildings of all sizes. manufacturers of ACM and their insurers Before 1940 the asbestos industry washad the capacity to generate their own mea- aware of the cancer-causing properties ofsures of exposure and to attack EPA's previ- asbestos. In the United States, the industryous misguided guidance documents. succeeded in keeping that information from the general public and health authorities.6Asbestos Measurements (German health officials, however, instituted In 1982 EPA published measurements measures to reduce asbestos exposures inincorrect asit turned outof airborne workplaces hefore and during World War II.)' asbestos. EPA contractors had collected sam- In 1955 Sir Richard Doll published aples of airborne asbestos in Houston school- paper that clearly showed that exposure to rooms and, using unconventional and invali- asbestos greatly increased the risk of lungdated methods, reported levels of airborne cancer and other deadly diseases.8 After 1955, asbestos that approached levels found in exposure to asbestos in U.S. workplaces was workplaces where asbestos was mined or reduced. (There is an extensive and signifi-milled; used in the fabrication of asbestos- cant literature about the importance of inter-containing products; or installed as insula- actions between smoking and asbestos expo-tion or fireproofing in ships, buildings, and sure in causing lung cancer, but there is no some appliances.1° Former manufacturers of doubt that exposure to asbestos in the work- ACM used standard methods for collecting place was a health risk.) and measuring airborne asbestos and found The realization that asbestos-containingmuch lower levels. Moreover, measurements materials (ACM) were present in many schools of airborne asbestos in buildings around the led parents in 1978 in some school districts in world revealed low levels of asbestos that were New Jersey to demand the removal of those often no different from levels in outdoor air EPA has never materials. The issue came to the attention ofor in buildings with no ACM.11 (Asbestos is the Environmental Protection Agency whenpresent in many minerals, such as "cats-eye" explained the the governor of New Jersey petitioned thesemiprecious stones and "rocks" of various reason for its agency to issue regulations for the control ofkinds; in addition, natural processes release reversal, but it is asbestos in public buildings. Shortly there- small amounts of asbestos into the air.) after, the Environmental Defense Fund peti- Faced with measurements that did not clear that entities tioned EPA to issue regulations on control ofsupport its own, EPA investigated the that bore the asbestos in schools. method that had been used in the Houston costs of asbestos Between 1978 and 1990, EPA issued aschools. EPA decided that the method was series of rules and guidance documentsimprecise and unreliable and discarded it.12 removal had the about the management of exposure toHad EPA listened to outside experts on mea- capacity to gener- asbestos in schools. The rules and guidancesuring asbestos levels in the workplace, the documents, all of which suggested that expo-agency would never have used the method ate their own sure to ACM was always a risk, caused many that led to the mistaken measurements in measures of expo- schools to rip out ACM, at costs from $7.5 Houston. sure and to attack billion to $16 billion.9 In the late 1980s, Congress ordered EPA In 1990 EPA reversed course. It issued a to measure asbestos levels inside and outside EPA's previous guidance document that called for leaving public buildings. When it did so, using stan- misguided guid- ACM alone"managing asbestos in place."dard measurement techniques, EPA found That document provided for protectionthere was essentially no difference between ance documents. from asbestos without the huge costs calledthe two environments."

6 EPA did not conclude, however, that lowasbestos in buildings. In this case, it was measured levels of exposure meant that the In this case, it was the availability of mea- the availability of risks were low. EPA continued to argue thatsurement methods and opportunities for all ACM were a time bomb that would go offother parties to make independent measure- measurement unexpectedly and shower building occupantsments that discredited EPA's earlier rulings. methods and with deadly concentrations of asbestos. The Access to the agency's data was not necessary opportunities for agency also argued that EPA-certified inspec- because methods for replicating those data tors could somehow predict future releases were available and used. other parties to on the basis of visual inspections. make indepen- By 1990, standard measurements revealedBenefits and Costs convincingly that the reports of EPA-certified If there are any health benefits that result- dent measure- inspectors about the danger posed by ACM bore ed from the furor over asbestos in schools, ments that dis- no correlation with asbestos concentrationsthose benefits are so small as to be unde- credited EPA's determined by measurements: the asbestos intectable!' And they may well be zero. ACM stayed put. The exception to that state- School systems and the taxpayers or the earlier rulings. ment occurred when ACM were ripped out.tuition payers that support the schools paid Then exposure levels soared. The best manage- a very high price (between $7.5 billion and ment solution for ACM was to leave them alone $16 billion) for ripping out ACM.19 Former until the building came down." ACM manufacturers and their insurers paid Also in 1990, Science, one of the mostlarge proportions of that money to school respected scientific journals in the world, districts that successfully sued for recovery of published a paper that showed that the riskstheir costs, and some former ACM manufac- from asbestos in buildings is very low.15 Forturers were plunged into bankruptcy. whatever reasonsmaybe because of the pres- No one was a winner except, maybe, own- tige of Science or maybe because the readers ofers of ACM inspection and removal compa- Science had read enough about the actualnies. Even with the collapse of.the asbestos- measurements of asbestos that they werein-buildings scare, asbestos removal goes on. ready to believe the articlethe Science paperBanks often refuse loans for the purchase of had a major impact and was widely reported buildings with ACM unless the seller or buyer in the newspapers and cited in Congress. agrees to remove the material. One estimate projects that, by 2020, an additional $50 bil- EPA Changes Its Mind lion will have been spent on asbestos In late 1990, EPA issued another in its removal.2° series of guidance documents about asbestos With the publication of its 1990 guidance in buildings.16 The document contained thedocument, EPA tacitly admitted its mistakes. following conclusions: health risk is relatedHowever, EPA did not bear the costs of its to the levels of airborne asbestos; those levelsmisguided policies. School districts, ACM are low in buildings with ACM; the healthmanufacturers and their insurers, and tax- risk "appears to be very low"; the removal of payers, none of which benefited from EPA's ACM is often not the best course of action policies, paid the price. and can increase exposures and risks; and building owners should keep an eye on ACM and repair or replace them when damaged. The Panic over Endocrine Finally, EPA required removal of ACM only Disrupters when buildings are being renovated or demolished." Thus, after a decade of arguing Theo Colborn, a scientist employed by the otherwise, EPA came around to the positionWorld Wildlife Fund, was convinced that taken by many occupational health expertsindustrial activity around the Great Lakes from the beginning of the debates aboutand chemical contamination of the lakes

5 7 were the cause of a cancer epidemic. In herThe Tulane Study own words, it was a "great setback"21 when The endocrine disrupter theory needed evi- she found no evidence of an epidemic. Ofdence to supportit, and that evidence course, it was no setback for the people living appeared in June 1996, when five scientists around the lakes, only for her credibility as anassociated with John A. McLachlan at Tulane expert on the effects of trace amounts ofUniversity published a paper in Science,23 chemicals on human and animal health. which has a reputation for rigorous prepubli- Colborn then turned to the literature oncation review of papers. Those scientists reports of deformed wildlife in the Greatreported that tiny amounts of pesticides, Lakes area. Such deformities have beenpresent at concentrations that are now per- reported for centuries, but Colborn theorizedmitted under EPA regulations, could interact that industrial chemicals were to blame. and bind very strongly with hormone recep- To explain the many different purportedtorsa first step toward affecting hormone- effects of the chemicals she observed, Colborn modulated biological systems. The researchers and others suggested that trace amounts ofreported, for instance, that two pesticides chemicals were affecting the functioning ofpresent at concentrations at which neither Five months after endocrine hormones such as estrogen andhad much effect independently were up to testosterone. Those hormones play major 1,600 times as potent in combination. the publication of roles in the development and functioning of On the basis of their results, McLachlan the Tulane many organ systems, and the hypothesis thatand his coauthors suggested that regulations results, scientists chemical disruption of those hormones might controlling pesticides needed to be tightened produce a panoply of adverse effects wasn't up. Science, which makes little or no effort to at a national without some merit. However, there was no publicize most of the hundreds of papers it meeting reported evidence to support the hypothesis. The onlypublishes each year, pulled out the stops with experimental or clinical evidence for hormon- the McLachlan paper. It ran a news article that they could al effects came from the study of exposure to about the experiment, complete with a pic- not replicate powerful drugs that affect the endocrine sys- ture of the Tulane researchers. It invited a sci- those results. tem in human beings or livestock. However, entist from the National Institutes of Health those exposures weren't to run-of-the-mill to write an article clarifying how the difficult- "chemicals." They were exposures to drugs to-explain results might have come about. selected because they cause those effects.Those efforts paid off. The Tulane results Moreover, the exposures weren't to low orwere reported in major newspapers and on trace levels of the drugs but to doses sufficientthe TV evening news. to cause observed clinical effects. The Tulane results surprised many scien- Colborn's coauthored book, Our Stolentists. How could they have had no inkling of Future,22 which appeared in March 1996,was the synergistic effects of pesticides? After greeted with reservations by science writers inall, pesticides have been subject to thousands the major newspapers. Her explanation thatof tests. It turned out that those scientists trace amounts of chemicals were the cause ofhadn't missed the synergistic effects. almost every adverse health effect ever Five months after the publication of the describedfrom deformities to cancer to infer-Tulane results, scientists at a national meet- tility to attention deficit disorder, froming reported that they could not replicate the wildlife population explosions to wildlife pop- results of the Tulane study. In January 1997, ulation collapseswas generally thought tooscientists from universities, the federal gov- glib. And how, if she was correct, had legions ernment, and industry published a letter in of toxicologists and epidemiologists missed Science that detailed their failures to replicate the pivotal role played by endocrine-disrupt-the Tulane results.24 A month later, another ing chemicals in the wide-ranging investiga-group of scientists published a report of its tions of the toxic effects of chemicals? attempts to replicate the Tulane results. In Nature, those scientists dismissed the Tulane ments in enough detail that others can try to findings and their implications: "Our results replicate them led to discovery of the mistake. do not support the assertion that synergism between estrogens is likely to present a majorThe Legislative Legacy of the Endocrine human or wildlife health concern."25 Disrupter Scare Confronted with the repudiation of their Before he published his paper, McLachlan results and conclusions, the Tulane scientists leaked the findings to EPA, and rumors first reacted by sticking to their guns. They about the paper's results were promoted by suggested that minor differences betweenEPA staff in discussions with congressional conditions and procedures in their laborato-staff. Those rumors were instrumental in the ry and the other laboratories explained thepassage of the Food Quality Protection Act discrepancies.26 But 13 months after they of 19963° and the Safe Drinking Water Act of published their alarming report, the Tulane 1996. Colborn's book and the Tulane results scientists threw in the towel. In a letter inwere especially instrumental in Congress's Science, McLachlan said: directing EPA to require new tests of commercial chemicals for endocrine disruption.31 I write to formally withdraw the The added testing may have a cost well report "Synergistic activation of beyond the economic one: it may cause dis- estrogen receptor with combinations ease and death. Bruce Ames, the recipient of of environmental chemicals." the 1999 National Science Medal, has written We have conducted experiments extensively about the importance of eating duplicating the conditions of our fresh fruits and vegetables in preventing can- earlier work, but have not been able cer.32One consequence of regulation of pesti- to replicate our initial results. cides is that fruits and vegetables become

Also...others have been unable more expensive and less available. Although to reproduce theresults. . . . the increases in price will not stop middle- Meanwhile, people in many walks of class shoppers from buying fresh produce, life have, on their own, put great low-income shoppers, whose diets are already weight on this report as the basis of far from optimal, may buy fewer fresh fruits much discussion, thought, and even and vegetables. If that is the case, their cancer public policy.27 risks will increase. Perversely, laws and regulations that are designed to protect con- The last quoted sentence is disingenuous. sumers from the hypothetical risks that may Colborn's book In fact, in the original paper, McLachlan andaccompany pesticideusecanactually his coauthors had suggested that their results increase health risks. and the Tulane would have public policy implications. When one considers all the adverse eco- results were Tulane University investigated whethernomic, and possibly adverse health, effects especially instru- any fraud had been involvedinthe that stem from the Tulane study, the reaction McLachlan paper. The university committeeof one government scientist to the study's mental in cleared McLachlan, but it concluded that unmasking is remarkable. Earl Gray, a scien- Congress's direct- Steven Arnold, the first author of the paper, tist from an EPA laboratory, remarked, "I'm "provided insufficient data to support thejust glad it's starting to clear up for John ing EPA to major conclusions of the Science paper"28 andMcLachlan."33 McLachlan's reputation is an require new tests that "independent review of Arnold's data issue, but he could have avoided tarnishing it of commercial does not support the major conclusions."29 by closer supervision of the work in his labo- Science worked. Even though the reviewers ratory. Long after McLachlan's reputation chemicals for for Science had not caught the faulty sciencestarts "to clear up" for him, the country will endocrine before the paper was published, the require- still face billions of dollars in costs of chemi- ment that scientists describe their experi-cal testing. Gray acknowledges that there is disruption.

7 9 Manufacturers of no justification for the testing that wasand published a corrected table," but they 2,4-D found a sig- added because of the Tulane study. "As forhave never explained how the error came the hypothesis that hormonelike chemicals about and how a question about herbicide nificant flaw in are dramatically more potent in combina-was turned into a conclusion about 2,4-D. NCI's study tion, 'it's kind of fallen by the wayside since Moreover, NCI depended on flimsy data. design when they the paper was retracted,'" says Gray.34 The only statistically significant increase in At the beginning of the endocrine dis-cancer was reported among workers exposed were allowed rupter controversy, Congress commissionedto herbicide (initially misreported as "2,4-D" access to it. the National Research Council to evaluateexposure) on more than 21 days per year. theevidence.TheNationalResearch How reasonable is it to assume that any indi- Council's August 1999 report, issued long vidual can remember whether he used an sher- after Congress had passed the Food Qualitybicide for 1-2 days, 3-5 days, 6-10 days, Protection Act and the Safe Drinking Water 11-20 days, or more than 21 days per year Act, declared that the endocrine disrupter over the past 15 or more years? According to hypothesis was "rife with uncertainty" andthe NCI analysis, the increased cancer risk that it was without clinical or experimentalfrom using herbicides for 1-2 days was support.35 greater than the risk from use for 3-5 and Gregg Easterbrook, commenting in the6-10 days. The reported risk at 11-20 days New Republic on the collapse of the endocrine was slightly greater than the risk at 1-2 days, disrupter scare, blistered Theo Colborn forand the risk at more than 21 days was highest the looseness of her reasoning.36 As the evi-of all. But only 11 workers (5 with cancer and dence for her theory collapsed, Colborn6 without) reported exposures of greater declared that evidence isn't important: "Justthan 21 days, and the entire NCI analysis is because we don't have the evidence doesn't based on the recall about herbicide exposure mean there are no effects."37 of 37 workers with cancer (or their next of kin). To draw any conclusion from those data appears to be a reach. The National Cancer Institute and the Follow-Up Studies Come Up Empty Herbicide 2,4-D Alarmed by the Kansas findings, a team of NCI scientists that included several authors In 1986 the National Cancer Instituteof the Kansas study decided to investigate published a study of Kansas farm workerscancer incidence among farmers exposed to that concluded that exposure to the herbi-2,4-D in Nebraska.4° In the Nebraska study, cide 2,4-D increased cancer rates.38 Thepublished in 1990, the NCI scientists did ask finding was widely reported; homeowners about 2,4-D use, but they found no statisti- who used 2,4-D on their lawns were fright-cally significant cancer increase, even in ened, and neighbors of those users were workers who reported use of 2,4-D on more alarmed as well. than 21 days per year. Manufacturers of 2,4-D found a signifi- The NCI reexamined its study of male farm cant flaw in NCI's study design when theyworkers and cancer incidence in Iowa and were allowed access to it. The NCI scientistsMinnesota. In that study, farmers were inter- had asked farmers a question about 2,4-D use, viewed about chemical exposures between and then asked only those who had reported 1981 and 1984 but were not asked specifically 2,4-D use additional questions about herbi- about exposure to 2,4-D. NCI interviewers sub- cide use. When the NCI scientists stated the sequently telephoned the Iowa and Minnesota results of those questions, they reported participants in 1987-88 to inquire about 2,4-D information about "herbicide" use as "2,4-D" usefour to seven years after the initial inter- use. The NCI scientists admitted their error views.41 For cases in which the participants had

10 died, the NCI scientists had to interview next between cancer in dogs and the dog owners' of kin or neighbors to obtain proxy informa- use of 2,4-D.47 Like the NCI studies of farm- tion about exposures. Although NCI steadfast-ers, the dog study attracted a lot of attention, ly considers first-hand information gatheredand editorials drew attention to the similari- from workers and proxy information gathered ties between the cancers reported in the farm- from next of kin equally valid, comparison ofers and in the dogs. directly reported and proxy data showed that Industry officials had some doubts about proxy data were more likely to report exposures the methods of analysis used by the authors to widely used substances, such as 2,4-D, andof the dog study, and the officials requested less likely to report exposures to less widely the underlying data from the NCI. That orga- used substances.42 That's no surprise. People nization stonewalled release of the data for who work with chemicals are more likely tomore than 18 months. Although the dog know the identity of the lesser-known sub-owners' names had already been removed stances than are people who do not. from the data, the NCI said that it was con- Mixing direct data and proxy data appearscerned that "industry" would use informa- to increase the chance that the data will showtion about the breeds of the dogs and ZIP an association between 2,4-D and cancer.code locations to track down and harass the Even with the mixing of data, however, no dog owners. The 2,4-D saga excess of cancer incidence was associated When the NCI finally released the data, sci- shows the impor- with 2,4-D in the Iowa-Minnesota study. entists at Michigan State University reana- tance of citizens' Although collecting and analyzing the lyzed the data. Their reanalysis revealed that 2,4-D data gathered from the telephone sur- the NCI-supported scientists had not clearly having access to vey delayed the publication of the NCI studyseparated 2,4-D exposures from other chemi- data so that they by two years, there was no mention of the cal exposures. When exposures were adequate- can check on the study when it appeared.43 The NCI scientists ly documented, the association between 2,4-D also did not report the telephone survey data and cancer in dogs disappeared.48. work of govern- during their testimony to a special EPA com- ment scientists. mittee that was considering regulation ofThe Lessons of 2,4-D 2,4-D.44 To the considerable embarrassment The 2,4-D saga shows the importance of of the NCI scientists who had ignored thecitizens' having access to data so that they data, industry scientists, who had obtainedcan check on the work of government scien- the data through FOIA requests, reportedtists. In the 2,4-D studies, much of the the data to the EPA committee.45 research was done by NCI scientistswho are The NCI authors of the Iowa-Minnesotagovernment employeesand that research study explained that they had decided not towas subject to FOIA requests. The errors that publish the results of the telephone surveywere revealed only after the data had been because of weaknesses in the study.46 Surely made available to other parties indicate quite the weaknesses would have been evident dur- clearly that mistakes can be made. ing the planning and development of the Manufacturers of 2,4-D have been harmed telephone survey, but the NCI scientistsby the NCI reports. Every report of risk plunged ahead, deciding against publication attracts attention and, probably, reduces sales, after the results were in. and every report of risk causes unwarranted worry on the part of users and consumers. The Dog Study However, reports that refute the cancer risk go In 1991 the NCI was finally able to pub- largely unnoticed. "Bad news sells." lish a report that seemed to substantiate the The NCI scientists have benefited from Kansas findings of an association betweentheir work. Their institute is in the midst of 2,4-D exposure and cancer. This time, thewhat is expected to be 20-year, $100 million NCI claimed to have found an association study of farmers' health. The failure to find 1' any risk from 2,4-D after well over a decade's study had eliminated from the study, without study has been translated into justification any scientific basis, 52 percent of the women for looking for other risks or renewing thewho had reported such infections and 25 per- search for risks from 2,4-D, despite the factcent of the women who had had no such that cancer rates among farmers are lowerinfections. Moreover, the researchers who than among the general population. reexamined the data suggested that adverse Unfortunately, the flawed NCI studies ofpublicity about IUDs at the time of the 2,4-D continue to have repercussions. TheWomen's Health Study could have affected small number of scientists whose reputa-patients' recall about their medical history tions depend on their reports of associationsand care and that doctors may have been more between 2,4-D exposure and cancer risk con- inclined to diagnose pelvic infections in tinue to refer to the NCI studies as supportwomen who used the Dalkon Shield. for their own results and conclusions. ForSumming up theirreexamination,the instance, Lennart Hardell, associate profes-authors of a 1991 report in the Journal of sor of oncology at the Orebro Medical CenterClinical Epidemiology said the study that had in Sweden, whose early reports showing sig- linked the Dalkon Shield to pelvic infections nificant increase in incidence of cancer from"showed almost a complete disregard for epi- a few days' exposure to 2,4-D have never been demiological principles in its design, conduct, replicated, and cannot be repeated, even by analysis and interpretation of results."52 him," cites without reservation the NCI However, the reanalysis was two years too studies as evidence for a 2,4-D association late. In 1989, A. H. Robins was purchased by with cancer.50 American Home Products, Inc., in what was then termed a "steal deal."53 The Dalkon Shield was never reintroduced to the market. The National Institutes of Health and the Dalkon Shield The FDA and Fen-Phen The Dalkon Shield. was an intrauterine birth control device (IUD) that was marketed In August 1999, a Texas jury awarded $23 by A. H. Robins Company beginning in 1971. million to a former user of the diet drug fen- In 1974 the U.S. Food and Drug Administra- phen. The verdict went against American tion asked the company to remove theHome Products, Inc., which made fenflu- The failure to Dalkon Shield from the market because oframine, or the "fen" portion of fen-phen. The find any risk reports of health problems. Later that year, verdict cost American Home Products' stock- the FDA said A. H. Robins could resume sales holders $8 billion in share value in one after- from 2,4-D after if a registry of users was kept. The company noon and cast a pall over the future value of well over a declined to keep a registry. the stock. decade's study In 1976 the National Institutes of Health Unscrupulous personal injury lawyers commenced the Women's Health Study to operating in an unbridled tort system has been translat- investigate the possible association betweendeserve much of the blame for that verdict; ed into justifica- IUDs and pelvic infections. The study washowever, a great deal of fault lies, as it did published in 1981 and reported that IUDs, inwith similar verdicts in cases of silicone tion for looking general, increased the risk of pelvic infections breast implants, with the Food and Drug for other risks or by 60 percent.51 The study was cited in litiga- Administration. renewing the tion that eventually forced A. H. Robins into The controversy over fen-phen began in bankruptcy proceedings. July 1997, when the New England Journal of search for risks In 1991 researchers reexamined the origi- Medicine rushed to publish a Mayo Clinic from 2,4-D. nal data from the Women's Health Study. report detailing how 24 fen-phen users had They found that the authors of the original experienced heart-valve damage.54 The report

12 10 was hardly an exhaustive scientific study.vented? Perhaps. After the FDA browbeat Only five of the claimed cases of heart-valveAmerican Home Products into withdrawing damage had actually been verified by surgical fenfluramine from the market, the agency examination. The heart-valve damage report-refused to provide to the company the origi- ed in the remaining cases had been assumednal data gathered in response to its letters to from echocardiograms, which may not be physicians,rendering American Home sufficiently reliable. The study did not com-Products virtually defenseless. pare rates of heart-valve problems in fen- phen users with rates in nonusers. The actual number of cases of heart-valve damage was The Shelby Amendment very small considering that millions of people had used or were using fen-phen at the Although members of the public had time of the study.55 access to agency-collected data, until passage The FDA then sent in 1997 letters to physi- of the Shelby Amendment, they did not have cians requesting information about heart-access to data collected by agency grantees. valve disease among fen-phen users. A month The amendment came about because of later, the FDA had collected 92 reports of dis- Congress's and others' inability to obtain the ease among 291 patients tested. Within days data that underlay EPA's then-proposed After the FDA and despite the availability of less drasticstringent new airqualityregulations, browbeat optionsthe FDA called for a halt in fen-phenannounced by EPA administrator Carol American Home use and persuaded American Home Products Browner in November 1996, for fine particu- to "voluntarily" pull fenfluramine and anoth- late matter (PM) and ozone. Products into er diet drug, Redux, from the market. The proposed regulations would be very withdrawing fen- There was no scientific evidence againstexpensive. EPA estimated that "partial attain- fen-phen, such as a controlled comparison ofment" of the required PM standard would fluramine from heart-valve damage incidence among users cost $6.3 billion per year and that partial the market, the and nonusers of the drug. If there had beenattainment of the ozone standard would cost agency refused to real and significant problems with fen-phen, $2.5 billion per year.57 Pointing out that "par- one would think they would have shown uptial attainment" has no legal or analytical provide to the much earlier and in many more individuals,basis, independent researchers estimated that company the since the drug had been used for years by mil-the costs of full attainment would be original data lions of people. The FDA's hasty action between $69 billion and $144 billion per year opened the litigation floodgates, just as thefor the PM standard and about $2.8 billion gathered in agency's 1992 moratorium on silicone breast for the ozone standard.58 EPA justified the response to its implants had done.56 As of November 1999,regulations by claiming they would prevent only a few months after the decision of the 15,000 premature deaths per year. Given that letters to physi- Texas jury, more than 5,000 lawsuits hadthe estimation of the health benefits of the cians, rendering been filed across the nation by former users regulations relied largely on a single contro- American Home of fen-phen. versial study, known as the "Pope study,"59 Three years after the FDA's action, though, it that had never been verified or replicated, Products virtually remains unclear whether fen-phen causes anysome scientists questioned whether the pro- defenseless. harm. The American Heart Association says it's posed rules, when implemented, would pro- too soon to tell whether fen-phen caused signif- duce the health benefits claimed by EPA.6° icant damage to anyone and whether the effects wore off when use of fen-phen stopped.Questions Surrounding the Pope Data Furthermore, in the Texas case, the plaintiff's It is particularly remarkable that no one own cardiologist testified that her heart prob- except the original researchers has ever seen lems predated her use of fen-phen. the data that underlie EPA's air quality regu- Could this series of events have been pre- lations. EPA, without ever having verified the

11 13 It is particularly data from the Pope study,61 nevertheless general publicare not allowed to see the data. remarkable that heaped billions of dollars in regulatory costs on American industry, cities, and consumers.Congressional Action no one except the The data on health, behavior, and exposure Members of Congress who requested to original used in the Pope study were collected by thesee the data were irked by EPA's refusal to researchers has American Cancer Society through a telephone provide access to those data and were buffet- survey of 1 million people. The survey wased by complaints about the new regulations ever seen the data conducted by 77,000 volunteers who, using a from organizations as diverse as the National that underlie questionnaire prepared by the ACS, inter-Conference of Black Mayors and the viewed neighbors, family, and friends aboutAmerican Farm Bureau. In 1997 members of EPA's air quality their lifestyles and their health practices. Congress introduced legislation requiring regulations. None of the information collected by the the federal government to make data from volunteers was validated or verified by anfederally funded research available to the independent party. The interviewers were notpublic. That measure was defeated. well trained, and there was little quality con- Less than a year later, in early 1998, trol. How dependable are the data? Imagine Congress passed legislation that would have asking a friend, neighbor, or relative aboutrequired the Office of Management and her smoking and drinking habits. How reli-Budget to study the implications of public able will the answer be? access to data from federally funded research. In the study, Pope and his coauthors com- For unrelated reasons, President Clinton bined the health-related information from vetoed that legislation. the ACS survey with air pollution data. That In October 1998, Sen. Richard Shelby study became the basis for EPA's November(R-Ala.) introduced language to the appro- 1996 air quality regulations. priations bill that provided funding for When industrialorganizationsandOMB. The language, which remained in the Congress asked to see the data used in thelegislation that was passed by Congress and Pope study, EPA and the investigators deniedsigned by the president, required OMB to the requests. Mary Nichols, then-head of the revise its rules EPA's Office of Air and Radiation, put it this

way: "We do not believe ...there is a useful to require Federal awarding agencies purpose for EPA to obtain the underlying to ensure that all data produced data [since the studies were published in under an award will be made avail-

peer-reviewed journals]. ...Securing more able to the public through proce- detail about this information is not necessary dures established under the Freedom as part of EPA's public health standard-set- of Information Act.63 ting process."62 In July 1997, EPA promulgat- ed the new air quality regulations. Ironically, the specific data that were Pope eventually agreed to release the data requested by Congress are, according to EPA, to a committee of the Health Effects Institute, not subject to the provisions of the amend- a group funded jointly by EPA and the auto-ment. The authors of this paper, under the mobile manufacturing industry. HEI is sup-auspices of the Citizens for the Integrity of posed to report its analysis of the data in Science, submitted an FOIA request to EPA 2000three years after the regulations wentfor the Pope study data." EPA refused, stat- into effect. But the agreement between theing: "We are not providing the health survey authors and HEI forbids members of the HEIdata you seek, because these data are not in committee to release the data to anyone else.the Agency's possession.... Since the records The parties most affected by EPA's regula-were not produced under an EPA award, the tionsstate and local governments, manufac- Public Law cited as authority for your request turers, transportation companies, and theis also not applicable."65 EPA's interpretation

12 is probably correct, because OMB's Circularrequest can be charged for all the expenses The agency insist- A-110 (discussed below), which incorporatednecessary to obtain the data. ed that the data the Shelby Amendment into rules for federal In the summer of 1999, a number of leg- grantees, says that "the Federal Governmentislative efforts were made to repeal or to delay are accurate and has the right to obtain, reproduce, publish, orimplementation of the Shelby Amendment. that the analysis otherwise use the data first produced underNone was successful. is well done. an award [from the federal government]."66 The health data used in the Pope study wereArguments against Data Access That may be true; first collected using ACS, not federal, funds. During the debate over access to the Pope however, since study data, EPA insisted there was nothing to OMB's Circular A-110 be gained from the public's having access to EPA officials On February 4, 1999, OMB proposedthe data. The agency insisted that the data deny ever having OMB Circular A-110, a revision to federalare accurate and that the analysis is well examined the grant policy to implement the Shelbydone. Amendment. OMB received more than 9,000 That may be true; however, since EPA offi- data, we wonder comments on the proposal, with 55 percentcials deny ever having examined the data,69 how they could of the respondents supporting the proposedwe wonder how they could possibly know. revisions. In August 1999, OMB issued a clar-Science is rooted in skepticism, including possibly know. ification to the revision, which garneredskepticism about one's own work. The physi- 3,000 comments. cist Richard Feynman, one of the century's OMB's final revision to Circular A-110,great scientists, cautioned about fallibility. In issued on October 8, 1999,67 provides that ascience, he wrote, "the first principle is that private citizen can make an FOIA request to a you must not fool yourself and you're the eas- federal agency that has supported research iest person to fool." It's a stretch to think that for any data produced with federal fundingthat first principle applied to Feynman but and that has resulted in a published report. Anot to the authors of the Pope study. report is "published" when it appears in a sci- Before OMB issued its regulations imple- entific or technical journal or when "amenting the Shelby Amendment, some scien- Federal agency publicly and officially citestific organizations, including the National the research findings in support of an agencyAcademy of Sciences,7° objected to imple- action that has the force and effect of law."68 mentation of the new law. The NAS argued This language excludes research resultsthat the law, by making scientists comply that are not published in the scientific or tech- with data requests, would open up to public nical literature or translated into policy and scrutiny notebooks from every federally sup- rules. The result is that data from much feder- ported laboratory and hobble scientific ally supported research are still not accessible. research. OMB's language for Circular A-110 In particular, data that are used in formulating put some of those concerns aside by limiting government risk assessments or in promulgat- access to data that are used in published ing "guidelines" cannot be obtained under thereports or cited for the development of rules provisions of Circular A-110. and regulations. The wording of the circular excluded from George Thurston, an EPA contractor FOIA requests "preliminary analyses, draftswhose study results have also been used by of scientific papers, plans for future research, the agency to justify new regulations, said, "If peer reviews, or communications with col- past history is any indication, vested interests leagues." In addition, it provides protectionwill misuse [the Shelby Amendment] to dis- for trade secrets and similar materials and forcredit valid research results they don't like personal medical information "which wouldand to harass the researchers doing the constitute a clearly unwarranted invasion ofwork."71 As the history of opposition to the personal privacy." The party making thePM regulations demonstrates, "vested inter- 15 13 Protecting the pri- ests" include many government organiza-health of the 1,200 Air Force personnel who vacy of individu- tions and industrial concerns. The largest sprayed 90 percent of the Agent Orange used class of "vested interests" consists of taxpay-during the Vietnam War. At five-year inter- als who partici- ers who paid for the research and who will vals, each of the 1,200 people who participat- pate in studies is eventually bear the costs of the regulations.ed in spraying Agent Orange and a compara- often difficult, Surely, "vested" or not, those organizationsble number of Air Force personnel who flew and individuals should be able to see the data similar aircraft in Southeast Asia during the but it's not on which regulations are based. Vietnam War but did not spray Agent impossible. Thurston and other EPA-supported scien-Orange undergo a thorough physical and tists supposedly put their motives above thosepsychological examination. of industry-employed scientists, who are pic- The Air Force's response to requests for tured as being willing to do anything fordata from that study illuminates how such money or other incentives. Angell sees manyrequests can and should be handled. In 1990 reports about health matters, which is the type Air Force scientists were contacted by veter- of "scientific" information that interests mostans' groups about releasing the data from the people. She is quite clear in stating that gov-study. After a discussion with the Depart- ernment-supported scientists are subject to ment of Health and Human Services commit- pressures and enticements, just as other scien-tee that oversaw the Air Force study, the deci- tists may be. In answer to a question aboutsion was made that the data would be made government reports, Angell said, "We're naive available to anyone who requested them. if we think government isn't subject to politi- The Air Force and the advisory commit- cal considerations."72 tee were very concerned with protecting the privacy of the study participants. Accord- Proprietary Data and Privacy Concerns ingly, the National Center for Health At least two issues that arise with dataStatistics "scrubbed" the data to remove accessproprietary interest in data and pro- personal identifiers. tection of the privacy of study participants Releasing personal data about health is deserve serious thought and action. not a trivial matter, and the Agent Orange Scientists can work for years to develop study by the Air Force demonstrates that databases, and then they can spend yearsconfidentiality of study participants can be "mining" those data and producing publica- preserved. Furthermore, FOIA rules and tions. Nevertheless, when the data are used to Circular A-110 provide for the protection of impose regulations on organizations and the identity of study participants. individuals, those organizations and individuals have proprietary interests in the data. The scientists' proprietary rights to their data Conclusion are certainly a consideration, and provisions are made in FOIA to protect them. However, The Shelby Amendment makes govern- those rights should not override the right ofment regulators who depend on scientific the public to have access to those data. data to justify their actions more responsible Protecting the privacy of individuals whoto the public. The amendment may, as participate in studies is often difficult, butSenator Shelby acknowledged, introduce it's not impossible. For example, the Agentsome unexpected problems as OMB imple- Orange study by the U.S. Air Force, one of thements the law; however, as he also stated, most politically sensitive epidemiologic stud-those problems can be dealt with as they arise. ies done by the government, shows that the Two potential problems are already privacy of individuals can be preserved. The apparent: (1) the new data access rules do not study, which began in 1982 and will end with apply to grants issued prior to the date on a reexamination in 2002, investigated thewhich the OMB regulations were issued and 16 14 (2) the rules apply only to actions that have methodology can serve only to strengthen The concern that the force and effect of law. The result of thethe scientific foundations of public policy. flawed scientific first problem is that the Pope study data, for example, can be used ad infinitum by EPA research may be without ever being made available to the Notes used to justify public. We recommend that Congress fix that loophole by directing OMB to imple- 1. Karl Popper, The Logic ofScientific Discovery (New misguided and ment the data access law so that data from all York: Harper & Row and Basic Books, 1959). costly regulations significant scientific studies funded and 2. Paul Davies, The Mind of God: The Scientific Basis is well grounded relied on by a federal agency, regardless of thefor 0 Rational World (New York: Simon and date of the grant, can be made available on Schuster, 1992), p. 28. in fact. Without request. With regard to the second problem, 3.Marcia Angell, Science on Trial: The Clash of independent the legislative history of the data access law Medical Evidence and the Law in the Breast Implant review of scientif- included federal policy as well as rules. Case (New York: W. W. Norton, 1996). However, the current OMB implementation ic data and 4. Marcia Angell, "Scared Sick," Speech delivered exempts federal policy from the law. Theat the meeting of the Independent Women's methodological effect is that agency policies, most notably Forum, Washington, February 17, 1999. for health and environmental risk assess- practices, policy ments, which can have significant societal 5.Jonathan Rauch, Kindly Inquisitors (Chicago: mistakes are University of Chicago Press, 1993). impact, are not covered by the data access inevitable. rule. We recommend that Congress direct 6. See Paul Brodeur, Outrageous Misconduct (New OMB to revise its implementation of the law York: Pantheon, 1985). to apply to any significant federal regulatory 7. See Robert N. Proctor, The Nazi War on Cancer action, including policy. (Princeton, NJ.: Princeton University Press, 1999). The concernthatflawedscientific researchimmune from independent review 8. Richard Doll, "Mortality from Lung Cancer in may be used to justify misguided and costlyAsbestos Workers," British Journal of Industrial regulations is well grounded in fact. Without Medicine 12, no. 1 (1955): 81-92. independent review of scientific data and 9.Dennis Cauchon, "Risk of Cancer in USA Is methodological practices, policy mistakes areBarely Measurable," USA Today, February 11, inevitable. 1999, p. 1. Although legislative efforts to stall or 10. Morton Corn, "Asbestos and Disease: An block the Shelby Amendment failed in theIndustrial Hygienist's Perspective," American summer of 1999, they may be resurrected in Industrial Hygiene Association Journal 47 (1986): the future. One argument that is certain to 515-23. be used against the amendment is that it's "a 11.Brooke T. Mossman etal.,"Asbestos: backdoor way"73 to interfere with regulatoryScientific Developments and Implications for programs. Imbedded in such arguments is Public Policy," Science 247, no. 2490 (January 19, the idea that the government cannot be ques- 1990): 294-300. tioned. The government and its scientists 12. Ibid. appear to be as prone to mistakes as anyone else. Requiring the government "to show its 13. Michael Gough, "Uncle Sam Flunks Asbestos work" opens up the regulatory process. in Schools," Issues in Science and Technology 4, no. 3 (1988): 81-85. Moreover, it ensures that federal regulations are based on sound science and reduces 14. Brooke T. Mossman et al., Asbestos in Public and doubts about the need for federal interven- Commercial Buildings (Cambridge, Mass.: Health tion. Whether one supports or opposes regu- Effects Institute, 1991), pp. 1-8, and passim. latory action, we should all acknowledge that 15. Mossman et al., "Asbestos: Scientific Develop- independent review of scientific data and ments and Implications for Public Policy."

15 16. Environmental Protection Agency, Office of Carcinogens and Anticarcinogens in the Human Diet Pesticides and Toxic Substances, Managing (Washington: National Academy Press, 1996). Asbestos in Place: A Building Owner's Guide to Operations and Maintenance Programs for Asbestos- 33. Quoted in Kaiser, "Tulane Inquiry Clears Lead Containing Materials (Washington; Environmental Researcher." Protection Agency, 1990). 34. Ibid. 17. Cassandra Moore,.Haunted Housing (Washing- ton: Cato Institute, 1997), pp. 159-208. 35. National Research Council, Hormonally Active Agents in the Environment (Washington: National 18. See Mossman et al., "Asbestos: Scientific Academy Press, 1999). Developments and Implications for Public Policy." 36. Gregg Easterbrook, "Science Fiction," New Republic, August 30, 1999, pp. 18-22. 19. Cauchon. 37. Quoted in ibid., p. 22. 20. Olin Jennirigs, a consultant who has tracked asbestos-related spending, quoted in Cauchon. 38. Sheila K. Hoar et al., "Agricultural Herbicide Use and Risk of Lymphoma and Soft-Tissue 21. Theo Colborn, Dianne Dumanoski, and John Sarcoma,"Journal of the American Medical Association Peterson Myers, Our Stolen Future (New York: 256, no. 9 (September 5, 1986): 1141-47. Dutton, 1996), p. 19. 39. Sheila K. Hoar et al., "Correction," Journal of the 22. Ibid. American Medical Association 256, no. 9 (November 21, 1986): 3351. 23. S. F. Arnold et al., "Synergistic Activation of Estrogen Receptor with Combinations of Envi- 40. Sheila H. Zahm et al., "A Case-Control Study ronmental Chemicals," Science 272, no. 5267 of Non-Hodgkin's Lymphoma and the Herbicide (June 7, 1996): 1489 (retracted publication). 2,4-Dichlorophenoxyacetic Acid(2,4-D)in EasternNebraska,"Epidemiology1,no. 24. K. Ramamoorthy et al., "Potency of Combined (September 1990): 349-56. Estrogenic Pesticides," Letter to the editor, Science 275, no. 5298 (January 17, 1997): 405-6. 41. Remarks of Ken Cantor, in "Modest Link between 2,4-D Exposure, Cancer Found in Iowa, 25. John Ashby et al., "Synergy between Synthetic Minnesota Study, AuthorSays,"Chemical Oestrogens?" Nature 385, no. 6616 (1997): 494. Regulation Reporter, May 1, 1992, pp. 173-75. 26. John A. McLachlan et al., "Response" to42. Geary W. Olsen and Kenneth M. Bodmer, Ramamoorthy et al., Science 275, no. 5298 "The Effect of the Type of Respondent on Risk (January 17, 1997): 406-7. Estimates of Pesticide Exposure in a Non- Hodgkin's Lymphoma Case-Control Study," 27. J.A. McLachlan, "Synergistic Effect ofJournal of Agromedicine 3 (1996): 37-50. Environmental Estrogens: Report Withdrawn," Science 277, no. 5325 (July 25, 1997): 459-63. 43. Kenneth P. Cantor et al., "Pesticides and Other Agricultural Risk Factors for Non-Hodgkin's 28. John C. LaRosa, chancellor, Tulane University Lymphoma among Men in Iowa and Minnesota," Medical Center, "Tulane Investigation Complet- Cancer Research 52, no. 9 (1992): 2447-55. ed," Letter to the editor, Science 284, no. 5422 (June 18, 1999): 1932. 44. EPA has concluded that the evidence for 2,4-D causing cancer is so unconvincing that it does not 29. Quoted in Jocelyn Kaiser, "Tulane Inquiry warrant classifying the herbicide as even a "possible Clears Lead Researcher," Science 284, no. 5422 carcinogen." See Environmental Protection Agency, (June 18, 1999): 1905. Health Effects Division, "Memorandum: Carcino- genicity Peer Review (4th) of 2,4-Dichlorophenoxy- 30. Daniel Byrd, "Goodbye Pesticides?" Regulation acetic Acid (2,4-D)," Memorandum from J. Rowland 20, no. 4 (Fall 1997): 57-62. and E. Rinde, Health Effects Division, to J. Miller and W. Waldrop, Special Review and Reregistration 31.Kaiser,"TulaneInquiryClearsLead Division, January 29, 1997. Researcher." 45. Geary Olsen, D.V.M., Ph.D., Dow Chemical 32. See, in general, National Research Council, Company, Letter to Special Joint Committee of

18 16 the EPA Science Advisory Board and the FIFRA part 1 (March 1995): 669-74. Scientific Advisory Panel, Washington, March 23, 1993. 60.See Michael Fumento, Polluted Science (Washington: AEI Press, 1997), pp. 29-30. 46. Kenneth P. Cantor et al., Letter to the editor, Cancer Research 53, no. 10 (supplement) (1993): 61. In response to a request for those data from 2421. Citizens for the Integrity of Science, EPA main- tained that "the health study data you seek are 47. H. M. Hayes et al., "Case-Control Study ofcontained in a data base that is proprietary with Canine Malignant Lymphoma: Positive Association the American Cancer Society (ACS). The EPA has with Dog Owners' Use of 2,4-Dichlophenoxyacetic never had access to this database." L. N. Wegman, Acid Herbicides," Journal of the National Cancerdirector, Air Quality Strategies and Standards Institute 83, no. 17 (1991): 1226-31. Division, Environmental Protection Agency, Letter to Steven J. Milloy, June 9, 1999. 48. J. B. Kaneene and R. A. Miller, "Re-analysis of 2,4-D Use and the Occurrence of Canine 62. Quoted in John Merline, "Industry Can't Get Maliginant Lymphoma," Veterinary and Human Numbers on the New Pollution Rule," Investor's Toxicology 41, no. 3 (1999): 164-70. Business Daily, February 21, 1997, p. Al.

49. Lennart Hardell and Mikael Eriksson, "A 63. P.L. 105-277. Case-Control Study of Non-Hodgkin Lymphoma and Exposure to Pesticides," Cancer 85, no. 6 64. Steven J. Milloy and Michael Gough, directors, (1999): 1353-60. Citizens for the Integrity of Science, Letter to JeraleneB.Green, national FOIA officer, 50. Ibid., p. 1354. Environmental Protection Agency, May 11, 1999.

51. R T. Burkman, "Intrauterine Device Use and 65. Wegman. the Risk of Pelvic Inflammatory Disease," Obstetrics and Gynecology 57, no. 3 (1981): 269-76. 66. Office of Management and Budget, "OMB CircularA-110, 'UniformAdministrative 52. R. A. Kronmal et al., "The Intrauterine Device Requirements for Grants and Agreements with and Pelvic Inflammatory Disease: The Women's Institutions of Higher Education, Hospitals, and Health Study Reanalyzed," Journal of Clinical Other Non-Profit Organizations,"' Federal Register Epidemiology 44, no. 2 (1991): 109-22. 64, no. 195 (October 8, 1999): 54926-30.

53. Malcolm Gladwell, "American Home's 'Steal 67. Ibid., p. 54930. Deal' for A. H. Robins," Washington Post, December 15, 1989, p. F1. 68. Ibid., p. 54928.

54. David Brown, "2 Weight Loss Drugs Linked to 69. Wegman. Heart Valve Ills," Washington Post, July 9, 1997, p. Al. 70. Bruce Alberts, president, National Academy of 55. H. M. Connolly et al., "Valvular Heart Disease Sciences, Letter to Jacob Lew, director, Office of Associated with Fenfluramine-Phentermine,"Management and Budget, January 26, 1999. New England Journal of Medicine 337, no. 9 (August 1997): 581-88. 71. George Thurston, professor of environmental medicine, New York University School of Medicine, 56. See Angell, Science on Trial. quoted in Jocelyn Kaiser, "New Law Could Open Up Lab Books," Science 282, no. 5391 (November 6, 57. Anne E. Smith et al., "Costs and Benefits of 1998): 1023b (in "News of the Week"). Ozone and PM 2.5 NAAQS," Reason Public Policy Institute Policy Study no. 226, June 1997. 72. Angell, "Scared Sick." 58. Ibid., p. 2. 73. Kevin Casey, senior director of federal and staterelations,OfficeofGovernment, 59. C. Arden Pope et al., "Particulate Matter as a Community, andPublicAffairs,Harvard Predictor of Mortality in a Prospective Study of University, quoted in Philip J. Hilts, "A Law United States Adults," American Journal ofOpening Research Data Sets Off Debate," New Respiratory and Critical Care Medicine 151, no. 3, York Times, July 31, 1999.

17 OTHER STUDIES IN THE POLICY ANALYSIS SERIES

365. Critiquing Sprawl's Critics by Peter Gordon and Harry W. Richardson (January 24, 2000)

364. The Greatest Century That Ever Was: 25 Miraculous Trends of the Past 100 Years by Stephen Moore and Julian L. Simon (December 15, 1999)

363. How and Why to Privatize Federal Lands by Terry L. Anderson, Vernon L. Smith, and Emily Simmons (December 9, 1999)

362. Milking the Sacred Cow: A Case for Eliminating the Federal Dairy Program by Kevin McNew (December 1, 1999)

361. Smart Growth at the Federal Trough: EPA's Financing of theAnt-Sprawl Movement by Peter Samuel and Randal O'Toole (November 24, 1999)

360. Imperial Overeach: Washington's Dubious Strategy to Overthrow Saddam Hussein by David Isenberg (November 17, 1999)

359. The Rise of Worker Capitalism by Richard Nadler (November 1, 1999)

358.Executive Orders and National Emergencies: How Presidents Have Come to "Run the Country" by Usurping Legislative Power by William J. Olson and Alan Woll (October 28, 1999)

357. Faulty Justifications and Ominous Prospects: NATO's "Victory" in Kosovo by Christopher Layne (October 25, 1999)

356. Energy Efficiency: No Silver Bullet for Global Warming by Jerry Taylor (October 20, 1999)

355. Strengthening the Biological Weapons Convention: Illusory Benefits and Nasty Side Effects by Eric R. Taylor (October 18, 1999)

354. The Community Reinvestment Act: Looking for Discrimination That Isn't There by George J. Benston (October 6, 1999)

353. Grave Robbers: The Moral Case against the Death Tax by Edward J. McCaffery (October 4, 1999)

352. Microsoft Redux: Anatomy of a Baseless Lawsuit by Robert A. Levy (September 30, 1999) 351. Fixing What Ain't Broke: The Renewed Call for Conscription by Doug Bandow (August 31, 1999)

350. Corporate Welfare for Weapons Makers: The Hidden Costs of Spending on Defense and Foreign Aid by William Hartung (August 12, 1999)

349. Gambling America: Balancing the Risks of Gambling and Its Regulation by Guy Calvert (June 18, 1999)

348. Replacing Potemkin Capitalism: Russia's Need for a Free-Market Financial System by Kurt Schuler and George A. Selgin (June 7, 1999)

347. Restoring the Boundary: Tort Law and the Right to Contract by Michael I. Krauss (June 3, 1999)

346. Speed Doesn't Kill: The Repeal of the 55-MPH Speed Limit by Stephen Moore (May 31, 1999)

345. Blunder in the Balkans: The Clinton Administration's Bungled War against Serbia by Christopher Layne (May 20, 1999)

344. Old Wine in New Bottles: The Pentagon's East Asia Security Strategy Report by Doug Bandow (May 18, 1999)

343.The State Spending Spree of the 1990s by Dean Stansel and Stephen Moore (May 13, 1999)

342. Is Readiness Overrated? Implications for a Tiered Readiness Force Structure by James L. George (April 29, 1999)

341.The Increasing Sustainability of Conventional Energy by Robert L. Bradley Jr. (April 22, 1999)

340.In Praise and Criticism of Mexico's Pension Reform by L. Jacobo Rodriguez (April 14, 1999)

339. Sports Pork: The Costly Relationship between Major League Sports and Government by Raymond J. Keating (April 5, 1999)

338. U.S. Assistance for Market Reforms: Foreign Aid Failures in Russia and the Former Soviet Bloc by Janine R. Wedel (March 22, 1999)

337. National Missile Defense: Examining the Options by Charles V. Perla and Barbara Conry (March 16, 1999)

0i 336. Internet Gambling: Popular, Inexorable, and (Eventually) Legal by Tom W. Bell (March 8, 1999)

335. Cars, Cholera, and Cows: The Management of Risk and Uncertainty by John Adams (March 4, 1999)

334. Hard Choices: Fighter Procurement in the Next Century by Williamson Murray (February 26, 1999)

333. Universal Preschool Is No Golden Ticket: Why Government Should Not Enter the Preschool Business by Darcy Ann Olsen (February 9, 1999)

332. Tilting at Windmills: Post-Cold War Military Threats to U.S. Security by Ivan Eland (February 8, 1999)

331. Time to Trash Government Intervention in Garbage Service by Peter M. Van Doren (January 21, 1999)

330. The Comprehensive Test Ban Treaty: The Costs Outweigh the Benefits by Kathleen C. Bailey (January 15, 1999)

329. Long Hot Year: Latest Science Debunks Global Warming Hysteria by Patrick J. Michaels (December 31, 1998)

328. What Term Limits Do that Ordinary Voting Cannot by Einer Elhauge (December 16, 1998)

327. Rethinking the Dayton Agreement: Bosnia Three Years Later by Gary T. Dempsey (December 14, 1998)

326. Reviving the Privileges or Immunities Clause to Redress the Balance Among States, Indivdivals, and the Federal Government by Kimberly C. Shankman and Roger Pilon (November 23, 1998)

325. Encryption Policy for the 21st Century: A Future without Government- Prescribed Key Recovery by Solveig Singleton (November 19, 1998)

324. Dismal Science Fictions: Network Effects, Microsoft, and Antitrust Speculation by Stan Liebowitz and Stephen E. Margolis (October 27, 1998)

323. The Government's War on Mergers: The Fatal Conceit of Antitrust Policy by William F. Shughart II (October 22, 1998) 22 AMI1110...21.7°

A " . - ' - . - . . es . - % e

- - . ' . - - - . -'

0 - ' . .. ' S. . - 1 1 1 . . . . , , 0 . . - :le 414 o . A II U.S. Department of Education Office of Educational Research and Improvement (OERI) ERICI National Library of Education (NLE) Educational Resources Information Center (ERIC)

NOTICE

REPRODUCTION BASIS

riThis document is covered by a signed "Reproduction Release (Blanket) form (on file within the ERIC system), encompassingall or classes of documents from its source organization and, therefore, does not require a "Specific Document" Release form.

This document is Federally-funded,or carries its own permission to reproduce, or is otherwise in the public domain and, therefore,may be reproduced by ERIC without a signed Reproduction Release. form (either "Specific Document"or "Blanket").,

EFF -089 (9/97)