Ensuring Independence and Objectivity 140 years, the National Academy of Sciences (NAS) has played a prominent role in setting national policies that are based on sci-ence. A congressional

EnsuringIndependence

andObjectivity

at The National Academies


andObjectivity


andObjectivity

atThe National Academies

CENTER FOR SCIENCE IN THE PUBLIC INTEREST

Washington, DC

Copyright © 2006 by Center for Science in the Public Interest

The Center for Science in the Public Interest (CSPI) founded in 1971, is a nonprofit health-advocacyorganization. CSPI conducts innovative research and advocacy programs in the areas of nutrition, foodsafety, and alcoholic beverages and provides consumers with current information about their own healthand well-being. CSPI is supported by almost one million subscribers in the United States and Canada toits Nutrition Action Healthletter and by foundation grants.

1875 Connecticut Avenue, NW, #300Washington, DC 20009

ph: 202-332-9110 fax: 202-265-4954www.cspinet.org

Acknowledgements

The Integrity in Science Project of the Center for Science in the Public Interest gratefully acknowledges the financial support of the Beldon Foundation, the Richard and Rhoda GoldmanFund, the Rockefeller Family Fund, the Saperstein Family Fund, and the Arkay Foundation.

This report was co-written by Integrity in Science project director Merrill Goozner and program manager Corrie Maudlin. The report was reviewed by CSPI executive director Michael F. Jacobson, Ph.D. Design by Cheryl Halterman.

Prefaceout of great respect for the National Academies. Each year,

Congress and government agencies call upon the NAS to provide the best scientificadvice possible on important and controversial topics. NAS reports invariably earn highmarks from the scientific community, and this study, which did not evaluate the qualityof any particular NAS report, makes no effort to question that consensus view.

Rather, because commercial pressures on the conduct of science have grown expo-nentially in recent decades, we focused on the process of nominating committees, as wellas the composition of committees, both of which strongly influences the conclusions andrecommendations of reports. We focused on the committee selection process because ofthe importance of protecting NAS committees from those pressures. Only through pre-venting the appearance of bias, imbalance, and conflicts of interest can the NAS ensurethat the high quality and reputation of its reports will be maintained in the future.

Unfortunately, we found serious deficiencies in the NAS’s committee-selectionprocess that could jeopardize the quality of future NAS reports. The NAS has allowednumerous scientists (and others) with blatant conflicts of interest to sit on committees.Compounding that problem, those conflicts of interest usually are not disclosed to thepublic. Whether complete avoidance of conflicts of interest on committees would haveimproved the committees’ recommendations is impossible to know. At the very least,though, improving the process of committee formation and excluding individuals withconflicts or balancing them with individuals having sharply contrasting views, wouldprovide added insurance that Academy recommendations will continue to be both top-notch and merit respect throughout the world.

Michael Jacobson, Ph.D.Executive DirectorCenter for Science in the Public Interest

This report is written

TABLE OF CONTENTS

Executive Summary ....................................................................................................................i

I. The Evolving Role of the National Academies ........................................................................1

II. Conflicts of Interest on NAS Committees ..............................................................................2

III. Policy Comparison: NAS and Selected Federal Agencies ......................................................5

A. NAS................................................................................................................................5

B. Environmental Protection Agency ..................................................................................7

C. Food and Drug Administration ......................................................................................7

IV. Analysis of NAS Committees..................................................................................................9

A. Eliminating Direct Conflicts of Interest ..........................................................................9

Sidebar: Hiding Ties to Industry Is a Risky Business ..................................................12

B. Balancing Ideological Biases..........................................................................................11

C. Disclosing Conflicts of Interest and Biases....................................................................15

V. Recommendations ................................................................................................................18

Appendix A – The Federal Advisory Committee Act ................................................................20

Appendix B – The 21 Committees ............................................................................................22

Notes ........................................................................................................................................23

BACKGROUND:now com-

prised of the National Academy of Sciences,the Institute of Medicine, the NationalAcademy of Engineering and the NationalResearch Council (hereafter referred to col-lectively as NAS) – plays a prominent role insetting national policies. Each year, the NASselects over 8,000 scientists to serve on morethan 500 committees that produce more than200 scientific analyses, many of them crucialto the nation’s health and environment.

This report focuses only on the selectionprocess for and composition of NAS panels.In 1997, NAS committees funded by the fed-eral government came under the jurisdictionof the Federal Advisory Committee Act(FACA). Congress required the agency toavoid using scientists with direct conflicts ofinterest on study panels (unless their expert-ise is deemed essential and not available else-where) and to seek balance on points of view.This report found serious breaches of both ofthose rules.

STUDY FINDINGS:1. A Failure to Eliminate

Direct Conflicts of InterestFor the purposes of this study, CSPI

defined a direct conflict of interest as a finan-cial tie within the last five years to a compa-ny or industry that is relevant to the commit-tee topic. A five-year window is comparableto policies at leading academic journals thatrequire conflicts of interest disclosure in pub-lished articles. Among the 320 professionalson 21 NAS committees investigated by CSPIover the past three years, at least 56 (18 per-cent) had direct conflicts of interest.

Finding: Nearly one out of every five sci-entists appointed to an NAS panel hasdirect financial ties to companies or indus-try groups with a direct stake in the out-come of that study. This consistent patternof appointing scientists with conflicts ofinterest clearly violates the spirit of theFederal Advisory Committee Act amend-ments that apply to NAS.

2. A Failure to Achieve BalanceIndependent background checks by

CSPI investigators found at least 66 scientistswhose employment, significant and long-term financial relationships, published writ-ings, think-tank membership, or courtroomtestimony demonstrated pro-industry posi-tions. On the other hand, only nine of the320 scientists worked for or were closelyidentified through their writings or publishedstudies with environmental or public interestgroups.

Finding: NAS did a poor job of balancingpoints of view on a majority of the studypanels examined. The NAS does notappear to consider information aboutpotential bias or conflicts of interest priorto nominating individuals to a committee.As a result, about half the panels exam-ined had scientists with identifiable biaseswho were not offset by scientists withalternative points of view.

3. A Failure to DiscloseInformation about Conflictsof Interest

Among the 320 scientists and other pro-fessionals examined in this study, nearly afifth had some ties to industries that might beaffected by the study. Yet the NAS revealed

EXECUTIVE SUMMARY

i

Executive Summary

The National Academies –

those ties to the public (via its website) inonly 46 percent of those cases.

Finding: The NAS provides brief biogra-phies of nominees to its committees onthe agency’s website. Such biographiescould assist people who were consideringcommenting on a committee’s composi-tion. However, those biographies are woe-fully inadequate because, in a majority ofcases, they fail to provide crucial dataregarding conflicts of interest and pointsof view.

RECOMMENDATIONSTo achieve its stated goal of “ensuring

independent, objective advice” for federalagencies, Congress, and the general public,the NAS should take the following steps:

• The NAS should expand its definitionof conflicts of interest to include anyfinancial ties within the past five yearswith companies that might be affectedby the committee’s work, either direct-ly or indirectly.

• If it allows a scientist with a conflict ofinterest to serve on a committee, theNAS should publish prior to the firstmeeting of the committee a waiverdocumenting the nature of the conflictand the reason for the waiver.

• The NAS should expand its definitionof balance on committees to includebias and point of view, in addition toareas of expertise. The definition ofpoint of view should be expanded toinclude scientists with extensive ties toindustry as having a “pro-industry”bias. Balance should be defined asnever having more than three suchscientists on any committee and

always balancing pro-industry scien-tists with at least an equal number ofpublic health-oriented scientists.

• To encourage compliance, the NASshould adopt sanctions, such as athree-year ban from serving on anyNAS panel, when a committee mem-ber fails to disclose appropriate infor-mation on the NAS disclosure forms.

• The NAS should make a much greatereffort to expand the pool of candidatesfor committees by reaching out to thebroader scientific community and notrelying on existing boards, committeesand routine contacts. It should pub-licly announce through its website andother outreach channels that it isforming a new study committee; thatannouncement should occur at leastthree months before the first meetingof the committee and include an invi-tation to the public to submit nomina-tions for membership on the commit-tee.

• Twenty days before a panel’s first meet-ing (or at the same time that it poststhe tentative roster for the committee),the NAS should publicly disclose allconflicts of interest, biases, and pointsof view of the scientists chosen forthat panel. Only after consideration ofpublic comments should NAS staffmake that roster final.

• The NAS should maintain a listservthat notifies interested individualswhen a committee is being formedand when it has chosen a tentative listof nominees. The latter notice shouldinclude a link to the proposed list ofnominees.

ENSURING INDEPENDENCE AND OBJECTIVITY AT THE NAS

ii

1

140 years, the National Academy ofSciences (NAS) has played a prominent role insetting national policies that are based on sci-ence. A congressional charter signed byPresident Abraham Lincoln on March 3, 1863,established the original NAS as a non-profitcorporation made up of distinguished academ-ics willing to provide free technical and scien-tific advice to the government. In its firstdecades of existence, the largely voluntaryNAS created committees on a wide variety ofsubjects with the twin goals of furthering sci-ence and technology and promoting the gen-eral welfare.

Over the course of the 20th century, theNAS expanded its role in science and technol-ogy policy-making. The need for sophisticatedtechnological expertise during World War Iled to the formation of a full-time NationalResearch Council (NRC) to serve as the princi-pal operating agency for the NAS. The grow-ing industrialization of society in the post-World War II era led to the formation in 1964of the National Academy of Engineering,which also came under the NRC umbrella.The growing importance of health care in theU.S. economy led to the formation of theInstitute of Medicine (1970), the last expan-sion of the National Academies’ structure. (Forthe remainder of this report, NAS refers to thefour national academies.1)

Today, reports from the NAS influencenational policies concerning virtually everyarea of modern life, from space and nano-technology to global warming and nutrition.Each year, the NAS selects more than 8,000academics and other experts to serve on over500 committees, which collectively producemore than 200 reports. While none of thecommittee members are paid, the budget ofthe 2,000-person organization reached $218million in 2004, largely for staffing and sup-

porting the various committees. Over 80 per-cent of that budget came from governmentagencies.2

The NAS can initiate the study processon a particular subject by using its own funds,generated mostly from sales of reports andcontributions. But this accounts for less than20 percent of its budget. Most requests comedirectly from federal agencies, which oftenlook to NAS for outside scientific support fornew policy initiatives. Sometimes two or morefederal agencies will ask the NAS to serve asan outside arbiter of an inter-agency dispute.A recent example was an NAS panel evaluat-ing the environmental effects of the rocket fuelcomponent perchlorate on local communities,which pitted the Defense Department andgiant contractors like Lockheed Martin againstthe Environmental Protection Agency (EPA) ina dispute over the pollution level that wouldrequire extensive clean-up operations.

Requests for new study committees oftenoriginate on Capitol Hill through a legislatorresponding to requests from interest groups.As a result, the NAS often finds itself involvedin the most contentious issues of the day andsubject to significant political pressure. Forinstance, an industry group opposed to animminent action by a regulatory agency mayask a friendly legislator to obtain funding foran NAS study reviewing the scientific reason-ing behind the agency’s decision. Public inter-est groups may lobby legislators to authorizemoney for an NAS study that might help buildsupport for new public health policies.Legislators seeking controversial laws or regu-lations favoring a special interest or con-stituent may seek an NAS study hoping it willjustify their proposals. In most such cases, themake-up of the NAS committees, which couldhave a major impact on their recommenda-tions, comes under intense scrutiny.

THE EVOLVING ROLE OF THE NATIONAL ACADEMIES

I. The Evolving Roleof the National Academies

For over


2

about conflicts of interest onNAS panels evaluating health and environ-mental policies is not new. In 1980 the NASFood and Nutrition Board issued a reportstating that Americans did not have to reducetheir intake of cholesterol and saturated fatexcept to achieve and maintain a normalbody weight, despite rising concern aboutthe association between blood cholesterollevels and heart disease.3 After the report’srelease, the media revealed that three mem-bers of the committee were food companyofficials while two others had served as con-sultants to egg producers.4 The media con-demned the report for being radically at oddswith prevailing scientific opinion, and, inresponse, NAS overhauled the Food andNutrition Board. The threat that outside pres-sure – whether from industry or health andenvironment advocates – might somedayinfluence NAS reports had become real. Yetthe brouhaha appeared to have little lastingimpact on internal NAS policies that guardagainst commercial or public interest advoca-cy influence over its reports.

Even before that report raised concernabout NAS procedures, the NAS began open-ing its committees to greater public scrutiny.That was partially in response to the 1972passage of the Federal Advisory CommitteeAct (FACA). That law applied open-meetingand conflict-of-interest rules to the thousandsof scientists and other professionals thatserved on the nearly 1,000 federal advisorypanels that make up “the fifth branch” ofgovernment. While the architects of the lawexpressly excluded NAS from its provisions,5

the quasi-governmental agency adopted sev-eral of the law’s provisions, including open-

ing its data-gathering meetings to the publicand publishing advance notice of meetings ina monthly newsletter.6 But fearing a loss ofindependence to the government agenciesthat provided funds, NAS officials vigorouslyobjected to coming under the FACA umbrel-la, a stance that did not change in responseto the controversial Food and Nutritionreport.

However, a lawsuit filed in 1989 byPublic Citizen called the NAS’ cherishedindependence into question. Public Citizenwanted American Bar Association meetingsbrought under FACA because of the JusticeDepartment’s habit of asking the ABA forjudicial recommendations. In denying theclaim, the Supreme Court drew a distinctionbetween independent groups like the ABA,which were clearly outside government, andthose whose mission and charter includedthe need to respond to government requests.The high court offered the NAS as a “paradig-matic example” of the latter because it was“in receipt of public funds” and was anorganization “created or permeated by theFederal government.”7

Legal ambivalence about NAS’ inde-pendence officially came to an end in the1990s. In one of the periodic protests againstspecial interest-influenced NAS committees,the Animal Legal Defense Fund and twoother animal rights groups sued the NationalResearch Council over a care guide forresearch animals prepared for the U.S.Department of Health and Human Services.The groups claimed that the NRC violatedFACA when it denied the public access to thecommittee’s deliberations. In January 1997,the U.S. Appeals Court in the District of

II. Conflicts of Intereston NAS Committees

Public concern

3

CONFLICTS OF INTEREST ON NAS COMMITTEES

Columbia reversed a lower court decisionand declared that NAS was subject to FACA.

NAS immediately protested that itsindependence would be severely compro-mised by government officials and lobbyinggroups if it were subjected to FACA rules.The science community rose to NAS’ defense.An editorial in Science, one of the nation’smost respected scientific journals, claimedthat federal agencies that provided fundswould inappropriately influence NAS studiesand that committee deliberations would besubjected to cumbersome regulations.8

Responding to those pleas, Congress inDecember of 1997 passed a new section ofFACA covering NAS whenever one of itscommittees received federal funds. In thosecases, the new rules required NAS to:

• Make names and brief biographies ofcommittee members public;

• Post notice of open meetings;

• Make available written materials pre-sented to the committee;

• Open data-gathering meetings to thepublic;

• Post summaries of meetings that arenot data-gathering meetings;

• Make copies of the final committeereport available to the public;

• Make available the names of the prin-cipal non-Academies reviewers of thedraft report;

• Follow conflict-of-interest and balancerequirements that were similar tothose established by FACA for advisory committees.9

The NAS was also directed to offer thepublic a reasonable amount of time to com-ment on the scientists selected to serve oncommittees, since the committees now had tobe “fairly balanced” and not include any indi-vidual with a conflict of interest “relevant tothe functions to be performed” unless theconflict was “publicly disclosed” and deemed“unavoidable” (see Appendix). The NASresponded by requiring panelists at a com-mittee’s initial meeting to disclose all poten-

tial conflicts of interest and biases to fellowcommittee members. In theory, this wouldresult in exclusion of panelists whose con-flicts were deemed too extensive by othercommittee members. In practice, this closedprocess may not result in full disclosure, letalone dropping conflicted members, especial-ly when it takes place at a first meeting of thecommittee where the participants may barelyknow one another. It is likely that the poten-tial effects of biases and conflicts of interestonly manifest themselves much later in theprocess, when a committee’s recommenda-tions are being drafted, for instance.

A General Accounting Office (now theGovernment Accountability Office or GAO)review of NAS’ compliance with the new lawless than a year after its passage found only afew minor violations of its provisions, mostlyhaving to do with incomplete or inaccuratepostings.10 But the outside auditors remainedconcerned. The 1998 GAO report noted thatfully 24 percent of all NAS panel members inthat year hailed from industry, typically fromcompanies or trade associations with aninterest in the committees’ topics and recom-mendations. The report raised serious con-cerns about undue industry influence overthe NAS study process.

In the late 1990s, protests about indus-try influence over federal advisory commit-tees intensified, largely due to frustrationwith government agencies rolling back ordelaying environmental, health and safetyregulations. Little of this anger was aimed atthe NAS. A 2001 GAO report, issued inresponse to a request from Rep. HenryWaxman (D-CA), called on the EPA toimprove its policies and procedures for advi-sory committees to ensure independence andbalance. The auditing agency reviewed thework of four committees established by theEPA’s Scientific Advisory Board in 1999 andfound numerous conflicts of interest involv-ing industry-funded scientists. The reportcompared EPA’s policies unfavorably to therelatively new NAS disclosure requirements.However, the GAO noted that the NationalAcademies’ definition of conflicts of interestwas less stringent than the rules imposed on


4

federal agencies by the Office of GovernmentEthics.11

During the George W. Bush administra-tion’s first term, protests over the make-up offederal advisory committees focused onalleged political interference in their delibera-tions. Editorials appeared in leading scienceand medical journals attacking the adminis-tration for ideologically packing committeesand subjecting potential members to politicallitmus tests. For instance, in 2003 the Centerfor Science in the Public Interest (CSPI) and21 academics sent a letter to the Office ofGovernment Ethics protesting the absence ofenforceable conflict of interest rules at theEPA, the Food and Drug Administration(FDA), the Interior Department, and otherfederal agencies. Democratic legislators askedthe GAO to once again review the issues ofconflict of interest and bias on advisory committees.

In April 2004 the GAO called for athorough overhaul of the rules.12 Numerousagencies had evaded conflict-of-interest rulesentirely by calling committee appointees con-stituency “representatives” instead of “specialgovernment employees,” the usual designa-tion for science-oriented appointees. OnlyEPA collected and evaluated informationabout advisors’ biases to ensure that advisorycommittees were properly balanced in termsof points of view. The GAO recommendedthat other agencies adopt similar guidelines.

The heightened scrutiny of federal agen-cies was matched by growing concern aboutthe make-up of NAS panels. In October2003, the Natural Resources DefenseCouncil, the Environmental Working Group,and CSPI wrote a joint letter protesting themake-up of a panel appointed to evaluate thehealth implications of perchlorate ingestion.The panel included Dr. Gilbert S. Omenn, amember of the board of directors of Rohmand Haas, which owned technology for per-chlorate removal, and Dr. Richard Bull, a for-

mer EPA scientist at Pacific NorthwestLaboratories, whose consulting contractsincluded perchlorate users like theDepartment of Defense and its contractors.13

The NAS ignored that protest as well as aMay 2004 protest letter from CaliforniaSenators Dianne Feinstein and Barbara Boxer.A month later a group representing perchlo-rate users heralded a separate new reportcommissioned by the Defense Department’scontractors that claimed trace amounts ofperchlorate were harmless. NAS panel mem-ber Bull had been one of its authors. Anembarrassed NAS, which had previously saidBull’s work for the industry had ended,forced Bull to resign from the panel.14

While that mini-scandal was unfolding,NAS moved to bring its policies more in linewith federal agencies. In 2003, NAS formallyunveiled a new policy governing committeeformation, disclosure, and how it definedbalance and conflicts of interest. The policystated that all committee nominees’ biogra-phies will be disclosed on NAS’ website 20days before the first meeting, giving the pub-lic time to comment on the roster. It requiredstaff to screen prospective members, whowere required to complete a “BackgroundInformation and Confidential Conflict ofInterest Disclosure,” which included employ-ment history, conflicts of interest, stock hold-ings of more than $10,000 in a potentiallyaffected corporation, research funding from aparty that would be affected by the commit-tee’s decision, and relevant patents. Thosedisclosures could lead to a person’s exclusionfrom the committee if he or she was “under-stood to be completely committed to a par-ticular point of view and unwilling, or per-ceived to be unwilling, to consider other per-spectives.”15 These new rules have, over thepast two years, led consumer groups to issuenumerous challenges to individuals chosen toserve on NAS panels.16

5

POLICY COMPARISON

the NAS creates and man-ages its study committees under special rulesincorporated in 1997 into the FederalAdvisory Committee Act. While the rules aresimilar to those at such government federalagencies as the EPA and FDA, there are someimportant differences. A close examination ofthose differences highlights areas where theNAS could improve its process (see box).

A. National AcademiesAccording to the 1997 amendments to

FACA, the NAS must provide public notice20 days prior to a committee’s first meetingof the names and brief biographies of indi-viduals nominated to committees; providethe public an opportunity to comment on theappointments before they are made or, if theNAS determines such prior comment is not

practicable, in the period immediately follow-ing the appointments; provide public noticeof committee meetings that will be open tothe public; and make meetings and writtenmaterial of the selected committees availableto the public, unless NAS deems the natureof the meeting or written material is of a pri-vate nature.17

In addition, the NAS must ensure thatno individual appointed to serveon the committee has a conflict ofinterest that is relevant to thefunctions to be performed, unlesssuch conflict is promptly andpublicly disclosed and NAS deter-mines that the conflict is unavoid-able. The committee membershipmust also be fairly balancedregarding the functions to be per-formed, and individuals that NASappoints or intends to appointmust inform the agency of any rel-evant conflicts of interest.

The NAS follows a four-stageprocess when conducting a study.The National Research Committee(NRC) Governing Board reviewsand approves the study scopewritten by NAS staff and thestudy’s sponsors (i.e., the group

that is requesting the study). The scope con-sists of a set of questions that form the “state-ment of task.”

NAS staff has the primary responsibilityfor choosing the members of individual com-mittees. It proposes a list of potential com-mittee members, which is then reviewed andapproved by the President of the NAS, who isalso the Chair of the NRC. NAS then disclos-

III Policy Comparison:NAS and Selected Federal Agencies

Disclosure: how three agencies compare

* FDA Advisory Committees are made up of standing members and special gov-ernment employees added for each meeting; the public is invited to submitnominations for the standing committees when they are periodically re-formed

** In several cases, the agency has posted conflict of interest information afterpublic protests.

*** The Federal Insecticide, Fungicide and Rodenticide (FIFRA) AdvisoryCommittee is an exception.

NAS EPA FDA

Public notice of committee formation N Y N*

Public can submit nominations N Y N*

Public invited to comment on committee roster prior to first meeting Y Y N

Conflicts of interest include past and present financial relationships N Y Y

Conflict-of-interest data made public N** N Y

Point-of-view data collected Y N*** N

Point-of-view data made public N N N

As noted above,

es the biographies of all selected committeemembers on its website under “CurrentProjects” and allows 20 days for public com-ment. According to NAS, the staff tries toselect a variety of experts who offer comple-mentary experiences and perspectives tobuild a “reasonably balanced” committee.Balance, in most cases, refers to having therange of professional expertise that is neededfor the committee to carry out its assignedtask. Provisional committee members areasked to disclose their conflicts of interestand biases at the first closed meeting of thecommittee.18

The NAS staffers do not conduct anindependent background search on potentialcommittee members for potential conflicts ofinterest and biases. They rely on self-disclo-sure. NAS officials state that there are toomany committee members selected each yearto allow independent assessment by NASstaff.19 If, after posting the committee rosteron the website, the public provides pertinentinformation, that information may beappended to the original information dis-closed on their confidential disclosure form.When an NAS staffer determines a scientistwith a conflict of interest has expertise thatcan not be found elsewhere, that conflict isdeemed “unavoidable” and publicly dis-closed.20

Potential sources of bias are not well-defined compared to direct conflicts of inter-est (e.g., a full or part-time employee of acorporation with a financial interest in theissue). Potential biases generally do not dis-qualify an individual from participating. Theyare taken into account as committee compo-sition is considered. An NAS official may aska person to resign from the committee if thepotential member is “completely committedto a particular point of view.”21

The NAS staff obtains information aboutconflicts of interest from “BackgroundInformation and Confidential Conflict ofInterest Disclosure” forms filled out byprospective members. The forms state thereason for disclosure: “The individual, thecommittee, and the institution should not beplaced in a situation where others could rea-

sonably question, and perhaps discount ordismiss, the work of the committee simplybecause of the existence of conflicting inter-est.” It is important to note that the NASdefines conflicts of interest as applying “onlyto currentinterests.” Itis not a con-flict for theindividual ifan interesthas expiredand “cannotreasonablyaffect currentbehavior.”22

In otherwords, anindustry-funded research project or consulting con-tract with companies directly affected by thestudy would not exclude a scientist from par-ticipating as long as that work was concludedprior to the start-up of the committee.

Financial conflicts of interest are definedas a relationship between the individual and(1) the sponsor or program reviewed in thestudy or (2) any person or organizationdirectly affected by the outcome of the study.Examples may include a stock holding ofmore than $10,000 in a potentially affectedcorporation, funding for research from aparty that would be affected, or a relevantpatent. The financial interests of a prospec-tive member’s spouse, employer, clients, orbusiness research partners are also consid-ered.23

The committee may use public meet-ings, information submitted by outside par-ties, scientific literature, and committeemembers’ work to produce its report. Thecommittee meetings where reports are delib-erated and written are closed. The public hasaccess to summaries of those meetings, butinterim and final drafts of the report remainconfidential. “Committee members continueto be screened for conflict of interestthroughout” the deliberation process.24

Independent experts are chosen toreview and provide anonymous comments to


6

The NAS staffers do notconduct an independentbackground search onpotential committeemembers for potentialconflicts of interest andbiases. They rely on self-disclosure.

POLICY COMPARISON

7

the committee once the report is produced.Reviewers are not checked for individualconflicts of interest because, the NAS states,they don’t have the final say in the report;they assess the report for scientific support,effective organization, and objectivity, whichthe committee considers in its final decision.25

Once the committee responds to the review-er’s comments, a new set of internal reviewmonitors ensures that the review criteria havebeen met. Top NAS officials approve the finalreport before sending it to the sponsor andreleasing it to the public.26

B. EnvironmentalProtection Agency

The EPA, like other federal agencies,must follow the Federal Advisory CommitteeAct of 1972. The EPA’s Scientific AdvisoryBoard (SAB) creates dozens of “science advi-sory panels” that throughout the year consid-er various issues facing the agency. Panelreports have a direct impact on regulationsand guidelines issued by the agency. Anextensive internal and external peer reviewprocess scrutinizes reports before they areissued.

While most of EPA’s regulations for itsSAB panels are similar to NAS’, there are sev-eral significant differences. Notices that theSAB is forming an advisory panel are printedin the Federal Register and sent out to inter-ested members of the public through anemail listserve. The public can nominate peo-ple to serve on SAB committees. The SABstaff creates a short list of nominees after itreviews candidates’ background informationincluding financial ties and “other informa-tion related to conflicts of interest or appear-ance of lack of impartiality.”27 Each candidateselected for an SAB advisory committee shortlist must then submit a ConfidentialFinancial Disclosure Form. Prospective boardmembers must list employment (compensat-ed and non-compensated), research supportand funding, consulting, expert testimony,and assets and liabilities over $10,000 ownedby themselves and their spouses during thetwo years prior to the date of filing.

Like many federal agencies, the EPArelies on a standard Office of GovernmentEthics form (Form 450) to collect financialconflicts of interest information. While simi-lar to the NAS form in the financial disclo-sures it requires, Form 450, according to theGAO’s 2004 review of the agency’s advisorycommittee policies, “does not request otherinformation relevant to assessing points ofview, such as previous public statements orpositions on the matter being reviewed,including statements in articles, testimony, orspeeches; positions taken in various legalforums, particularly in providing expert legaltestimony; research conducted on the matter;interests of their employers or clients in thematter; and sources of funding for researchor other activities.” One EPA advisory com-mittee – the Federal Insecticide, Fungicide,and Rodenticide (FIFRA) Scientific AdvisoryPanel – does collect such information inde-pendently and considers it in selecting mem-bers for its peer review panels.28

C. Food and DrugAdministration

The FDA maintains more than 30 advi-sory committees to advise the agency aboutscientific and regulatory issues and newproducts applications involving drugs,devices, biologics, veterinary medicines, andfood. Most committees have a permanentroster that is supplemented, as needed, byadditional members chosen because of theirexpertise in the topic of the meeting. Meetingnotices appear in the Federal Register at least60 days in advance of the meeting, andinclude the date, place, time, and agenda forthe meeting. However, the full roster of thecommittee is not released until the day beforeor the day of the meeting.

Like the EPA, the FDA asks for informa-tion about conflicts of interest from the rela-tively recent past that may be related to thecommittee topic. The “Confidential FinancialDisclosure Report” asks those selected if theyhave had “expert witness” or “speaking andwriting” work related to the topic in the pastyear. The form also asks if “you, your spouse,

minor child, general partner, organization inwhich you serve as an officer, director,trustee, general partner or employee” haveany past involvement with the meeting/taskissues.29 In addition, each member must dis-close information related to “financial hold-ings and professional activities related to theproduct(s) or its competitor(s) that are sub-jects of the meeting.”30

The FDA makes liberal use of clauses inFACA that allow scientists with conflicts ofinterest to serve on advisory panels. Theagency issues waivers in situations whereofficials consider the conflicts to be minimalor the scientist’s expertise is considered cru-cial to the committee’s deliberations. Virtuallyall FDA advisory panels include some scien-tists who receive conflict-of-interest waivers.There have even been cases where a majorityof the committee received waivers. When acommittee is discussing a scientific or regula-

tory issue that applies to all firms in theaffected industry, the FDA will grant a “gen-eral” waiver to anyone who has any financialties to any company in that industry.

After a controversy erupted over con-flicts of interest on a panel evaluating a classof painkiller drugs known as COX-2inhibitors,31 Congress in October 2005passed an amendment to the FDA appropria-tions bill that required the agency to postcopies of all waivers on the FDA’s website atleast 15 days before the meeting date. Thewaivers include the nature of the conflict ofinterest and a range of its financial value, butnot the names of companies. However, fornew product advisory committee meetings,the relationship requiring a waiver is desig-nated as being the sponsor or a competitor.In the former case, that is tantamount to dis-closing the name of the company since thesponsor is known.


8

ANALYSIS OF NAS COMMITTEES

9

in recent years theNAS has come under increased scrutiny forconflicts of interest and biases on some of itscommittees. Given the range of practices atfederal agencies that face similar issues andsome obvious differences with NAS practices,CSPI analyzed NAS practices in light of bestpractices at other agencies. Over the studyperiod, CSPI researchers analyzed 21 NASpanels with a total of 320 scientists. Thosecommittees were not randomly selected.Most of the committees addressed regulatoryissues that affect health or the environment.

The committees were analyzed for:

• Direct conflicts of interest – Did theNAS allow scientists or other profes-sionals with direct conflicts of interestto serve on the committees? If so, didthe NAS tell the public why thoseindividuals were critical to the com-mittee’s function?

• Ideological bias – Did the NAS identifycommittee members with apparentbiases about the subject matter athand? Did the NAS inform the publicabout those members’ biases? Did theNAS seek to balance those views withmembers who could serve as an ideo-logical counterweight?

• Disclosure – Did the NAS provide thepublic with information about panelnominees’ potential and direct con-flicts of interest and their biases dur-ing the committee formation processand after the first general meeting ofthe committee?

A. Eliminating Direct Conflictsof Interest

Scientists’ commercial conflicts of inter-est are influencing the nature and quality ofresearch at the nation’s leading universities.32

This has triggered fears that the private com-mercial interests are having an inordinateinfluence over the policies of the govern-ment’s most influential science-based agen-cies, including the EPA, the FDA and theNational Institutes of Health.

These concerns have led to a vigorousdebate over the definition of conflicts ofinterest at a science-based agency. NIH, forinstance, defines conflicts of interest in sci-ence as “situations in which financial or otherpersonal considerations may compromise, orhave the appearance of compromising, aninvestigator’s professional judgment in con-ducting or reporting research.”33 A recent callfor ending conflicts of interest at academicmedical centers stated that “conflicts of inter-est occur when physicians have motives orare in situations for which reasonableobservers could conclude that the moralrequirements of the physician’s roles are orwill be compromised.”34

Note that both definitions go beyondthose rare circumstances where research out-comes have been crudely or fraudulentlyinfluenced by persons with financial ties to abusiness with an interest in the outcome of aparticular research project. They specificallyallude to situations where the conflicts mayinfluence professional judgment, even thoughmost researchers vehemently deny that suchpossibilities exist. The ambivalent postureassumed by many researchers when confront-

IV. Analysisof NAS Committees

As previously noted,

ed by the idea their financial ties might com-promise their professional obligation to main-tain strict objectivity was insightfully por-trayed in a recent Hastings Center essay bybioethicist Carl Elliott:

The degree of dissembling and rational-ization here might be funny if thestakes were not so high. “I take themoney but it doesn’t influence me.” “Itake the money from many differentsources in order to keep my objectivity.”“I take the money but I make sure thatno more than forty percent of our cen-ter’s funding comes from corporatesources.” “I take the money but Ialways disclose.” “I take the money butI say what I want.” Or my favorite: “Itake the money but I use it to advocatefor social justice.” The rationalizationsalways begin with the phrase: “I takethe money.” No one will just say no.35

Both definitions also include public per-ception, which involves the appearance ofconflict of interest. Self-interested parties,including Congress, government agencies,and corporate lobbying groups, are increas-ingly turning to the NAS to define the scien-tific state of play on controversial topics,whether it is global warming, stem cellresearch, or a specific toxic chemical. For theNAS to maintain its credibility in this role, itmust be vigilant in rooting out even theappearance of conflicts of interest among itscommittee members.

Conflicts of interest stem from manydifferent types of financial arrangements.Many are not readily apparent from individu-als’ positions at universities or university-affiliated centers or institutions. University-based researchers can receive personal remu-neration in the form of industrial support ofresearch contracts, grants, consulting con-tracts, speaking fees, honoraria, and serviceon corporate advisory boards. Those arrange-ments may be ongoing or intermittent, andthey may not be disclosed on a scientist’sresumé or curriculum vitae. While a skilledresearcher may be able to find evidence for

these ties (gleaned from disclosures in aca-demic articles or at scientific meetings, forinstance), they are generally hidden frompublic view.

For the public to maintain its faith inthe integrity of the NAS process, it is criticalthat all financial ties that even appear to be aconflict of interest get disclosed on the NASwebsite during the panel selection process.The FACA amendments of 1997 are quitespecific with regard to the NAS’ need to barscientists with conflicts of interest from itscommittees: “The Academy shall make itsbest efforts to ensure that no individualappointed to serve on the committee has aconflict of interest that is relevant to thefunctions to be performed, unless such con-flict is promptly and publicly disclosed andthe Academy determines that the conflict isunavoidable.” (See Appendix)

For the purposes of this study, CSPI dif-ferentiated between two types of financialconflicts of interest (COI) for each member: apotential COI and an actual or direct COI.Potential COIs are defined as any financialties to any company or industry group withinthe last five years, even if that private entityhad no relationship to the committee’s topic.The extent of these financial ties was used todetermine if that scientist was biased towardan industry or held a corporate point of view(see the next section on Balancing IdeologicalBiases). Direct COIs are financial ties to acompany or industry within the last fiveyears that are relevant to the committeetopic, that is, where the person is an employ-ee or has financial ties to an industry tradegroup, research institute, a company in theindustry, or a direct competitor of a companyin the industry with a direct stake in the out-come of the committee’s deliberations. Whilethat is more stringent than the current defini-tion employed at NAS, it is in line with poli-cies established at many leading academicjournals that require conflicts of interest dis-closures to accompany published articles.

Among the 320 scientists on 21 com-mittees investigated by CSPI, 136 (43 per-cent) had some ties to industry or potentialCOIs in the five years prior to the commit-


10


11

tee’s formation (see Figure 1). At least 56 ofthose scientists (18 percent) had direct con-flicts of interest, that is, financial ties to com-panies, industry trade groups, or researchinstitutes with a stake in the outcome of thestudy. These figures may be underestimatedsince CSPI did not have access to members’disclosure forms, but relied on searchingpublicly available and reputable sources ofinformation to document these financial ties.

Example: In March 2005, the IOMselected Richard Hynes of the MassachusettsInstitute of Technology to chair a committeedeveloping “Guidelines for HumanEmbryonic Stem Cell Research.” Hynes is amember of the scientific resource board forGenentech, the nation’s second leadingbiotechnology firm with extensive researchcapacity in stem cells and stem cell-relatedfields. According to NAS staff, Hynes’ tie toGenentech was not considered a conflict ofinterest because the topic involved bioethicsquestions that would have “no differential onindustry.”36

Example: In April 2005, the IOMappointed David C. Bellinger of the HarvardSchool of Public Health to a committee on“Nutrient Relationships in Seafood: Selectionsto Balance Benefits and Risks,” which isweighing the relative risks methyl mercuryposes to some consumers against the health

benefits of fish consumption. A few monthsafter the panel convened, the AmericanJournal of Preventive Medicine published astudy co-authored by Bellinger and fundedby the United States Tuna Foundation andthe National Food Processors AssociationResearch Foundation.37 NAS staff said thatBellinger did not have a conflict of interestbecause his work on the study and its sub-mission to the journal had been completed

prior to formation of the commit-tee. (See sidebar on next page.)

Examples: Some commit-tees were heavily populated withscientists with ties to the affectedindustries. The “State Practicesin Setting Mobile SourceEmissions Standards” panelincluded at least 4 of 11 mem-bers with direct financial ties tooil or vehicle industries. Of the11 scientists evaluating the“Department of Energy’s CarbonSequestration Program,”launched in August 2005, 10had ties to carbon-emittingindustries. The 16-member“Food Marketing and the Diets

of Children and Youth” committee, whichissued its report in 2006, included at least 7individuals with direct ties to the food, mediaand advertising/marketing industries, all ofwhich had a direct stake in the study’s con-tents.

Finding: Nearly one out of every five sci-entists appointed to an NAS panel haddirect financial ties to companies or indus-try groups with a direct stake in the out-come of that study. This consistent patternof appointing scientists with conflicts ofinterest clearly violates the spirit of theFederal Advisory Committee Act amend-ments that apply to NAS and in somecases may have violated the law itself.

B. Balancing Ideological BiasesFACA requires that federal agencies bal-

ance advisory committees both in terms offunctions to be performed and points of view.

Independent

57%

Potential Conflicts

of Interest

25%

Direct Conflicts of Interest

18%

Figure 1: Conflicts of interest of members of 21 study committees (n=320)

The language in Section 15, which applies toNAS, is less prescriptive. It states only that“the committee membership is fairly bal-anced as determined by the Academies to beappropriate for the functions to be per-formed,” omitting the phrase about “points ofview.” (See Appendix.) However, the 2004GAO report that reviewed federal agency

compliance with FACA included a long dis-cussion about balance requiring detailedinquiry into scientists’ points of view, and,based on interviews with NAS officials,praised the agency for its efforts to go beyondthe definition of balance required by thestatute. “The National Academies state thatallegations of conflict of interest or lack of


12

Hiding Ties to Industry is a Risky Business

Eating mercury-laden fish can be risky for children and pregnant women. Yet fish isan important source of omega-3 polyunsaturated fatty acids, which are good for cardio-vascular health.

Where should a consumer turn to for advice?The Environmental Protection Agency and Food and Drug Administration warn

pregnant women and children to avoid eating large predatory fish (which have the high-est mercury content) and limit other fish-intake to twice a week.38 But the AgricultureDepartment’s food pyramid lists swordfish, mackerel, and tuna as good sources of pro-tein even though they have high levels of mercury. The food pyramid contains no warn-ings.39 That isn’t surprising. The department is heavily involved in promoting fish con-sumption through five regional aquaculture centers.40

It is not alone. The National Oceanographic and Atmospheric Administration(NOAA) promotes off-shore fish farming.41 In late 2004 NOAA commissioned theInstitute of Medicine to study the relative risks and benefits of seafood consumption.“People tell our administrator the [EPA/FDA] advisory is confusing,” said SpencerGarrett, director of NOAA’s National Seafood Inspection Laboratory in Pascagoula,Mississippi.

Given the high stakes nature of the debate where the public health community andthe fishing industry have widely divergent views, the IOM should have ensured that thescientists and other professionals on its committee would be free from conflicts of inter-est. Yet when the panel was unveiled in January 2005, it included a scientist who hadjust completed a study of the same topic – a risk-benefit analysis of fish consumption –that was paid for by the National Fisheries Institute, an industry trade group, amongothers.42 David C. Bellinger of Children’s Hospital in Boston and his colleagues at theHarvard Center for Risk Analysis concluded that pregnant women who reduced fishconsumption instead of substituting low-mercury species for high-mercury specieswould be doing more harm than good for their developing fetuses. The study appearedin the academic literature, complete with disclosure of its funding sources, months afterthe committee’s work got underway.

Bellinger’s involvement in a fishing industry-funded study represented a conflict ofinterest that violated any reasonable interpretation of the NAS/IOM’s conflict of interestrules. The NAS/IOM even failed to disclose his industry ties on its website or tell con-sumer groups who met with IOM staff in February 2005.43

When later questioned about Bellinger’s conflict of interest, NAS executive directorWilliam Colglazier defended his inclusion. He stated Bellinger’s study had been complet-ed and submitted for publication prior to his inclusion on the committee.44


13

balance and objectivity (emphasis added) canundermine the conclusions of fully compe-tent committees.”45

While direct conflicts of interest are rel-atively straightforward to understand becausethey involve financial relationships, bias orpoint of view is more difficult to track. Onthe one hand, a scientist’s employment, orga-nizational affiliations, and published writingscan indicate a perspective on issues. Thechief scientist for a major corporation, a sci-entist who writes op-ed columns with a cer-tain perspective, a staff scientist at an envi-ronmental group – all can and should beidentified as having a particular point of viewfor the purpose of balancing a committee.But simply because a scientist receivesresearch grants or speaking fees from a com-pany or industry research institute doesn’tmean that the scientist is biased in favor ofindustry. While studies have shown thatresearch results tend to reflect the points ofview and interests of funders, speculation asto the cause of this relationship remains pre-cisely that – speculation.46

On the other hand, scientists with along history of working for industry or pub-lic interest groups can be assumed to bebiased for purposes of balancing committeerosters. “A university professor who is also anofficial of an environmental advocacy organi-zation may reasonably be viewed by a spon-soring agency and others as representing anenvironmental rather than an unaligned ‘aca-demic’ perspective,” the GAO said in its 2004review. “Similarly, a university professor whois also an official of a toxicology institute thatreceives funding from chemical companies orwho provided expert legal testimony for acorporation may reasonably be viewed by asponsoring agency and others as providingan industry perspective.” The GAO analysisclearly suggests that a committee with severalindustry consultants, unless they were bal-anced by roughly the same number of healthor environmental advocates, would violateFACA.47

The issue of bias is more clearly definedfor scientists who are directly employed byorganizations with an avowed point of view.

When the chief scientist of General Motorssits on an advisory panel, there can be littledoubt that the positions he takes will reflectthe interests of his employer. In a similarvein, a scientist employed by the NaturalResources Defense Council is safely assumedto be pro-environment.

The gray area occurs when scientistshave ties with industry groups or advocacyorganizations that are so extensive that theycan be said to reflect that particular point ofview even though they are not directlyemployed by those groups. As noted above, acommittee dominated by such scientistswould be considered unbalanced. But shouldagencies establishing committees take stepsto counterbalance the presence of even oneor two such persons on a committee? Thatrequires independent investigation into a sci-entist’s point of view. The GAO report sug-gested such bias may be inferred from pub-lished statements in press articles, public tes-timony and speeches, court appearances, andaffiliations with ideologically-oriented thinktanks or research institutes, in addition toextensive relationships with industry.

Identifying points of view of potentialappointees is not aimed at barring those sci-entists from apanel,although thatcould be oneapproach tobalancing acommittee.Rather, FACAmerelyrequires thatsuch biases beoffset by scien-tists withopposing or alternative points of view.Depending on the committee’s task, it is pos-sible that the ideal panel would include anumber of scientists with clashing points ofview in order for the full committee toreceive a complete airing of the issue at hand.

Some science policy experts do not sup-port collecting information on possible biasesfor the purposes of establishing balance.

When the chief scientistof General Motors sitson an advisory panel,there can be little doubtthat the positions hetakes will reflect theinterests of his employer.

Responding to complaints about politicalinfluence over the advisory committeeprocess, a 2004 NAS panel chaired by formerCongressman John Porter (R-IL) called foreliminating such questions from advisorycommittee appointment forms. “Scientists,engineers, and health professionals should beappointed to federal advisory committeesbased on their expertise and integrity,” thereport said. “They should not be asked forinformation that would have no bearing onthe scientific or technical expertise theywould provide during committee discussions– such as political party affiliation, votingrecord, or personal positions on particular issues(emphasis added).”48 That clashes directlywith the spirit of FACA and the GAO viewthat collecting such information“would better ensure that thecommittees are, and are per-ceived as being, fairly balanced interms of points of view – and thatno one interest or viewpointdominates.”49

Ironically, the NAS in 2003began soliciting such informationabout potential panelists’ viewson its confidential disclosureforms. Candidates are asked: “Listyour relevant articles, testimony,speeches, etc., by date, title, andpublication (if any) in which theyappeared, or provide relevantrepresentative examples if numer-ous. Provide a brief description ofrelevant positions of any organi-zations or groups with which youare closely identified or associat-ed.”50

It is impossible to know how many ofthe 320 scientists on the 21 panels underreview disclosed information about theirpoints of view since the completed formsremain sealed and the discussions about biasand conflicts of interest at the first meeting ofpanels take place behind closed doors. NASofficials told CSPI that the staff makes noattempt to investigate possible sources of biasindependent of the disclosures in the forms.The CSPI investigation found few instances

where the NAS identified a point of view orbias in the panel member’s brief biographyposted on the Internet.

On the other hand, CSPI’s limited back-ground check of the 320 scientists examinedfound at least 66 scientists (21 percent)whose employment, significant and long-term financial relationships, published writ-ings, think tank membership, or courtroomtestimony demonstrated a pro-industry bias(see Figure 2). In contrast, only 9 of the sci-entists worked for or were closely alignedwith environmental, health or public interestgroups. Moreover, at least 9 of the 21 panelsexamined had a significant number of scien-tists with extensive financial ties to industrywho were not balanced by scientists with sig-

nificant ties to health, environmental, orother advocacy groups. Such committeesshould, under the GAO definition of bias thatincludes not only a demonstrated point ofview but extensive ties to either industry orpublic interest groups, be categorized asimbalanced and in violation of FACA.

Example: Nine of the 19 members ofthe NAS committee assessing the toxicity ofdioxin and dioxin-like compounds had


14

Figure 2: Committee Balance for 21 NAS Panels (number of panel members)

70

60

50

40

30

20

10

0Lean to Industry Lean to Public Interest Groups

66

9


15

extensive ties to industry, mostly in the formof research contracts. Four of them conduct-ed research for manufacturers with a stake inhow regulators handle dioxin, which is atoxic by-product of the chemical, pesticide,and pulp and paper bleaching industries. Nota single scientist on the panel was directlyaffiliated with an environmental organizationor closely aligned with environmentalistgroups.

Example: In 2004 Congress asked NASto evaluate the EPA’s recently adopted“Changes in New Source Review Programsfor Stationary Sources of Air Pollutants.” TheNAS appointed Princeton University profes-sor William Happer, a prominent globalwarming skeptic whose writings haveappeared in Hoover Institution publications,as chairman. The committee also includedGeorge Mason University professor Brian F.Mannix, whose writings have appeared inpublications issued by the Mercatus Institute,an industry-oriented think tank that raisesfunds from fossil fuel industries. Mannix in2001 wrote that the “EPA’s New SourceReview program is a substantial deterrent toinvestment in new oil refinery and powergeneration capacity.” No scientists on thepanel had a pro-environmentalist bias.Following protests against their presence onthe panel, Happer and Mannix resigned.

Example: The NAS committee studying“Toxicity Testing and Assessment ofEnvironmental Agents,” impaneled in March2004, included 12 scientists biased towardindustry but only one biased toward environ-mental protection. The group included GailCharnley, a private consultant to DowChemical who is affiliated with the AnnapolisCenter, an anti-regulatory think tank basedin Maryland; Nancy Kerkvliet of OregonState University, also connected to theAnnapolis Center; and Abby A. Li, a formertoxicologist for Monsanto who recently hadjoined Exponent, a private consulting firmwhose clients are mainly corporations. Theonly environmentalist on the panel was Dr.Gina Solomon, a staff scientist for the NaturalResources Defense Council.

Example: In June 2004, the NAS’ Boardon Environmental Studies and Toxicologylaunched a 20-month review of state pollu-tion control programs. The request was initi-ated by Sen. Kit Bond (R-MO), who wasseeking an exemption for a major employerin his state.51 The panel included GaryMarchant, an environmental law professor atArizona State University who once represent-ed major automobile manufacturer tradeassociations in regulatory proceedings. Thepanel did not include any lawyers from envi-ronmental groups or state agencies. More-over, the committee included four scientistswith close ties to industry and none with tiesto environmental groups.

Example: The August 2005 panel evalu-ating the feasibility of the Department ofEnergy’s Carbon Sequestration Programincluded 10 of 11 members with recent orcurrent financial ties to the petroleum, energyand chemical industries. The only committeemember without financial ties to carbon-based industries was a university oceanogra-pher from Rhode Island.

Finding: NAS did a poor job of balancingpoints of view on a majority of the 21study panels examined. Even where anattempt was made to create balance, theresults are almost always tilted towardindustry dominance. The NAS makes noindependent attempt to research nomi-nees’ potential biases. It relies on scien-tists’ self-disclosure, plus whatever infor-mation is provided by the public.

C. Disclosing Conflicts ofInterest and Biases

Disclosure is one means of managingconflicts of interest and bias in science andscientific publishing. Disclosure providesusers of the data or readers of the publica-tions with information that will help themevaluate the quality of the research and theconclusions of the researchers. Initially con-troversial, it is now considered routine atmany leading science and medical journals.Most government agencies that employ scien-

tists as advisors require disclosure, as domedical societies at most of their meetings.However, as the controversy of requiring dis-closure declined, a new debate emerged: doesdisclosure go far enough? “[Disclosure] servesto highlight the potential for bias, but cannotand does not eliminate the conflicts,” the edi-tor of the Journal of the American MedicalAssociation wrote in 2001.52

Nor does disclosure eliminate the out-right manipulations that sometimes occur.Disclosure does not prevent authors with tiesto commercial entities or with their owncommercial interests from delaying publica-tion, keeping conflicting information confi-dential, or skewing their interpretations to fittheir sponsors’ needs.53

Some authors have failed to disclose rel-evant financial interests. The May 2005 stemcell research breakthrough announced inScience by Hwang Woo-Suk, Gerald Schattenand others, which was later retracted asfraudulent, failed to disclose the two scien-tists’ patent filings.54 After a CSPI study iden-tified a number of articles that failed toreport relevant conflicts of interest in fourprominent journals,55 one journal,Environmental Health Perspectives, pledged toban scientists from publishing within theirpages for three years if fraudulent disclosurefilings are later discovered.56

These well-publicized failures to dis-close conflicts of interest (well publicized inthe science press at least) should have madeNAS officials sensitive to the importance ofdisclosing on its website the conflicts ofinterest of committee members. Disclosure isnot only a way of providing this informationto the public, which is crucial to maintainingthe public’s faith in the integrity of the NASrole in the science policy process, but aninternal tool for preventing or managing con-flicts of interest and bias on its committees,which the previous two sections showed ithas consistently failed to do.

The disclosure requirements in Section15 of FACA are quite specific, stating that noone can serve on a committee:

...who has a conflict of interest that is

relevant to the functions to be per-formed, unless such conflict ispromptly and publicly disclosed andthe Academy determines that the con-flict is unavoidable. (See Appendix.)

The NAS is also required by statute togive brief biographies of potential committeemembers to help the public comment ontheir qualifications. The public cannot intelli-gently comment on the conflicts of interestand balance aspects of committee nomineesunless their biographies include that informa-tion. Yet, according to NAS officials, the biog-raphical postings are limited to providing “anappropriate statement of the member’s quali-fications to serve on the committee, high-lighting the member’s expertise and technicalcredentials related to the committee’s chargeand his or her relevant current and past affili-ations.”57

CSPI’s independent, but necessarily lim-ited, research found that nearly one out ofevery five (18 percent) scientists reviewedhad some ties to industry. Yet the NAS toldthe public about those ties in fewer than halfthe cases. Moreover, not a single scientist wasidentified as having a particular bias or pointof view beyond a handful of cases wheretheir employment by public interest/environ-mental groups or private firms made theirbiases apparent.

Example: With one exception, every scientist on the “Panel on DOE’s CarbonSequestration Program” had a direct conflictof interest, either consulting for or beingemployed by oil and energy companies witha potential stake in the committee’s outcome.Only half of those ties were disclosed on theNAS website prior to the committee’s forma-tion, or subsequently revealed by NAS.

Example: The “State Practices in Setting Mobile Source Emissions Standards”Committee, which was reviewing controver-sial standards, included four scientists withdirect ties to oil and vehicle companies. Justtwo of those ties were revealed by the NAS tothe public prior to formation of the committee.


16


17

Example: The “Review of the WIC[Women, Infants and Children] FoodPackages” committee included five scientistswith ties to the food industry or food indus-try-related research institutes. None of thoseties was revealed to the public prior to beingappointed.

Example: Excluding the reporting ofdirect employment by corporations or advo-cacy organizations, none of the biographieslisted for the 21 committees reviewed byCSPI included a statement suggesting that aprospective committee member had a non-financial bias about the subject matter. CSPI’s

independent investigation found at least 3scientists with biases that had been revealedin other forums. (See the examples of Happerand Mannix on page 15.)

Finding: The NAS provides brief biogra-phies of prospective members of its com-mittees on its website, and giving the pub-lic 20 days advance notice before the firstmeeting provides sufficient time to com-ment. However, those biographies are woe-fully inadequate for that purpose becausethey consistently fail to provide crucialdata regarding conflicts of interest andpoints of view.

of 21 NAS committeesimpaneled over the past three years foundthat the NAS:

• Failed to keep its legal obligation toappoint scientists without conflicts ofinterest to its study panels.

• Failed to identify numerous proposedpanel members with ideological bias-es; failed to balance them with panelmembers with opposite or alternativepoints of view; and allowed over-rep-resentation of industry-oriented scien-tists on many panels, thus failing tomeet the test of balance.

• Failed to provide the public withinformation about the conflicts ofinterest and biases of prospectivepanel members, even though the pub-lic is encouraged to comment on thoseand other issues during the committeeformation process.

To achieve its stated goal of “ensuringindependent, objective advice” for federalagencies, Congress, and the general public,the NAS should:

• Strengthen its definition of conflicts ofinterest to include any financial ties tocompanies, industry trade groups, orindustry-dominated research institutesthat might be affected by the commit-tee’s work. The NAS should considerany financial ties within the past fiveyears when evaluating conflicts ofinterests, not just current ties, which isits present policy.

• Explicitly state that it is NAS policy toexclude any scientists with conflicts ofinterest from committees unless theirexpertise is crucial to successful com-

pletion of the committee’s task. Insuch cases, the NAS should issue andpublish conflict of interest waiversdocumenting the nature of the conflictand the reason for the waiver.

• Change its definition of “balance” oncommittees to include bias or point ofview (in addition to the current defini-tion that committees will be balancedby including all the necessary types ofexpertise). The definition of point ofview should be expanded to includescientists whose extensive financial tiesto industry or public health/environ-mental groups (not just employmentby those entities) merit their beingconsidered as having that point of view.

• More closely evaluate scientists’ pointof view disclosures to determine ifcandidates need to be balanced withcandidates with opposite or alternativepoints of view. Balance should bedefined as having roughly equal num-bers from opposing or alternativepoints of view.

• Review compliance with the require-ment for disclosure of conflicts ofinterest and points of view/bias priorto a scientist being nominated for thecommittee. To encourage compliance,the NAS should ban any scientist whofails to disclose appropriate informa-tion from serving on any NAS panelfor three years.

• Inform scientists who are being con-sidered for panels that their conflictsof interest will be publicly disclosed.

• Publicly announce through its websiteand other outreach channels that it is


18

V. RecommendationsThe CSPI review

forming a new study committee. Thatannouncement should occur at leastthree months before the first meetingof the committee and include an invi-tation to the public to submit nomina-tions for membership on the committee.

• Post on its website, after a tentativeroster for the committee has been cho-sen, all financial conflicts of interest,points of view, and biases that are rele-vant to the topic being studied by thecommittee. Those disclosures shouldbe published at least 20 days beforethe first meeting of the committee to

allow the public sufficient time tocomment.

• Publish along with the final roster anassessment of public commentsregarding a committee’s nominees andoverall balance.

• Maintain a listserv that notifies inter-ested individuals when a committee isbeing formed and when it has chosena tentative list of nominees. The noticeshould include a link to the proposedlist of nominees.

RECOMMENDATIONS

19

TITLE 5 – APPENDIXFEDERAL ADVISORY COMMITTEE

ACTSec. 15. Requirements relating to the

National Academy of Sciences and theNational Academy of Public Administration -STATUTE-

(a) In General. - An agency may not useany advice or recommendation provided bythe National Academy of Sciences orNational Academy of Public Administrationthat was developed by use of a committeecreated by that academy under an agreementwith an agency, unless -

(1) the committee was not subject toany actual management or control by anagency or an officer of the FederalGovernment;

(2) in the case of a committee createdafter the date of the enactment of the FederalAdvisory Committee Act Amendments of1997, the membership of the committee wasappointed in accordance with the require-ments described in subsection (b)(1); and

(3) in developing the advice or recom-mendation, the academy complied with -

(A) subsection (b)(2) through (6), inthe case of any advice or recommendationprovided by the National Academy ofSciences; or

(B) subsection (b)(2) and (5), in thecase of any advice or recommendation pro-vided by the National Academy of PublicAdministration.

(b) Requirements. - The requirementsreferred to in subsection (a) are as follows:

(1) The Academy shall determine and

provide public notice of the names and briefbiographies of individuals that the Academyappoints or intends to appoint to serve onthe committee. The Academy shall determineand provide a reasonable opportunity for thepublic to comment on such appointmentsbefore they are made or, if the Academydetermines such prior comment is not practi-cable, in the period immediately followingthe appointments. The Academy shall makeits best efforts to ensure that

(A) no individual appointed to serveon the committee has a conflict of interestthat is relevant to the functions to be per-formed, unless such conflict is promptly andpublicly disclosed and the Academy deter-mines that the conflict is unavoidable,

(B) the committee membership isfairly balanced as determined by theAcademy to be appropriate for the functionsto be performed, and

(C) the final report of the Academywill be the result of the Academy’s independ-ent judgment. The Academy shall requirethat individuals that the Academy appointsor intends to appoint to serve on the com-mittee inform the Academy of the individual’sconflicts of interest that are relevant to thefunctions to be performed.

(2) The Academy shall determine andprovide public notice of committee meetingsthat will be open to the public.

(3) The Academy shall ensure thatmeetings of the committee to gather datafrom individuals who are not officials, agents,or employees of the Academy are open to thepublic, unless the Academy determines that a


20

Appendix AFEDERAL ADVISORY COMMITTEE ACT PROVISIONS

SPECIFIC TO NAS

meeting would disclose matters described insection 552(b) of title 5, United States Code.The Academy shall make available to thepublic, at reasonable charge if appropriate,written materials presented to the committeeby individuals who are not officials, agents,or employees of the Academy, unless theAcademy determines that making materialavailable would disclose matters described inthat section.

(4) The Academy shall make availableto the public as soon as practicable, at rea-sonable charge if appropriate, a brief summa-ry of any committee meeting that is not adata gathering meeting, unless the Academydetermines that the summary would disclosematters described in section 552(b) of title 5,United States Code. The summary shall iden-tify the committee members present, the top-ics discussed, materials made available to thecommittee, and such other matters that the

Academy determines should be included.(5) The Academy shall make available

to the public its final report, at reasonablecharge if appropriate, unless the Academydetermines that the report would disclosematters described in section 552(b) of title 5,United States Code. If the Academy deter-mines that the report would disclose mattersdescribed in that section, the Academy shallmake public an abbreviated version of thereport that does not disclose those matters.

(6) After publication of the finalreport, the Academy shall make publiclyavailable the names of the principal reviewerswho reviewed the report in draft form andwho are not officials, agents, or employees ofthe Academy.

(c) Regulations. - The Administrator ofGeneral Services may issue regulations imple-menting this section.58

APPENDIX

21


22

BEST-K-03-07-A; 06/01/04Division of Earth and Life StudiesBoard on Environmental Studies andToxicologyState Practices in Setting Mobile SourceEmissions Standards

BEES-J-04-01-D; 08/30/05Division on Engineering and PhysicalSciencesBoard on Energy and EnvironmentalSystemsPanel on DOE’s Carbon SequestrationProgram

BOGH-H-04-01-D; 01/21/05Institute of MedicineDivision of Behavioral and SocialSciences and EducationBoard on Global HealthSubcommittee on Care for the President’sEmergency Plan for AIDS Relief [PEP-FAR] Implementation Evaluation

HPDP-H-05-03-A; 06/24/05Institute of MedicineBoard on Health Promotion andDisease Prevention Sciences PolicyAsbestos: Selected Health Effects

HPDP-H-04-02-A; 06/15/04Institute of MedicineBoard on Health Promotion andDisease Prevention Sciences Policy Review of the National ImmunizationProgram’s (NIP’s) Research Proceduresand Data Sharing Program

HSPX-H-04-06-A; 01/24/05Institute of MedicineBoard on Health Sciences PolicyEthical Considerations for Revisions toDHHS Regulations for Protection ofPrisoners Participating as Subjects inResearch

HPDP-H-05-01-A; 04/06/05Institute of MedicineBoard on Population Health andPublic Health PracticeAssessment of the U.S. Drug SafetySystem

FNBX-H-04-03-A; 08/26/04Institute of MedicineFood and Nutrition BoardBoard on Children Youth and FamiliesFood Marketing and the Diets ofChildren and Youth

FNBX-H-03-01-A; 12/23/03Institute of MedicineFood and Nutrition BoardReview of the WIC Food Packages

FNBX-H-04-04-A; 11/01/04Institute of MedicineFood and Nutrition BoardNutrient Relationships in Seafood:Selections to Balance Benefits and Risks

FNBX-H-01-90-B; 03/23/05Institute of MedicineFood and Nutrition BoardCommittee on Military NutritionResearch

DELS-O-04-01-A; 03/23/05Division on Earth and Life StudiesReview of NIOSH Research Programs

BLSX-K-04-03-B; 09/14/04Division on Earth and Life StudiesInstitute Of MedicineBoard on Earth Sciences andResourcesGuidelines for Human Embryonic StemCell Research

BESR-U-04-06-A; 09/14/04Division on Earth and Life StudiesBoard on Earth Sciences andResourcesMine Placement of Coal CombustionWastes

BBXX-K-00-02-A; 03/14/03Division on Earth and Life StudiesInstitute of MedicineBoard on Life SciencesFood and Nutrition BoardAgriculture and Natural ResourcesBoardProcess to Identify Hazards and Assessthe Unintended Effects of GeneticallyEngineered Foods on Human Health

BEST-K-05-01-A; 07/13/05Division on Earth and Life StudiesBoard on Environmental Studies andToxicologyEnvironmental Impacts of Wind EnergyProjects

BEST-K-03-08-A; 09/09/04Division on Earth and Life StudiesBoard on Environmental Studies andToxicologyBoard on Agriculture and NaturalResourcesFood and Nutrition BoardReview EPA’s Exposure and HumanHealth Reassessment of TCDD andRelated Compounds

BEST-U-03-08-A; 03/18/04Division on Earth and Life StudiesBoard on Environmental Studies andToxicologyToxicity Testing and Assessment ofEnvironmental Agents

BEST-K-03-04-A; 02/25/04Division on Earth and Life StudiesBoard on Environmental Studies andToxicologyChanges in New Source ReviewPrograms for Stationary Sources of AirPollutants

BEST-K-03-02-A; 12/11/03Division on Earth and Life StudiesBoard on Environmental Studies andToxicologyEnvironmental Decision Making:Principles and Criteria for Models

BEST-K-03-06-A; 12/21/2004Division on Earth and Life StudiesBoard on Environmental Studies andToxicologyAssessing the Human Health Risks ofTrichloroethylene

Appendix BTHE 21 COMMITTEES

1 For more background on the history of the NationalAcademies, see http://www.nas.org.

2 Personal communication, William Colglazier, NASExecutive Officer, to CSPI, October 28, 2005. Seealso http://www.nasonline.org/site/PageServer?page-name=ABOUT_main_page; http://www.nationala-cademies.org/annualreport/rev04.htm.

3 Jane E. Brody, “Panel Reports Healthy AmericansNeed Not Cut Intake of Cholesterol,” New YorkTimes, May 28, 1980, p. 1.

4 Jane E. Brody, “Experts Assail Report Declaring Curbon Cholesterol Isn’t Needed,” New York Times, June 1, 1980, p. 1.

5 Animal Legal Defense Fund, Inc. v. Shalala, 104 F.3d424 (D.C. Cir. 1997).

6 Government Accounting Office, “Federal Research:The National Academy of Sciences and the FederalAdvisory Committee Act,” GAO/RCED-99-17,November 1998, p. 5.

7 Animal Legal Defense Fund, Inc. v. Shalala.8 M.R.C. Greenwood, “Raiders of the Last Bastion?”

Science, July 11, 1997, p. 163.9 GAO/RCED-99-17, p. 12.10 GAO/RCED-99-17, p. 2-3.11 Government Accounting Office, “EPA’s Science

Advisory Board Panels: Improved Policies andProcedures Needed to Ensure Independence andBalance,” GAO-01-536, p. 14-19.

12 Government Accountability Office, “FederalAdvisory Committees: Additional Guidance CouldHelp Agencies Better Ensure Independence andBalance,” GAO-04-328, April 16, 2004.

13 Letter from NRDC, EWG and CSPI to Bruce Alberts,President, National Academy of Sciences, Oct. 9,2003. On file at CSPI.

14 Peter Waldman, “Perchlorate-Panel MemberResigns,” Wall Street Journal, June 11, 2004, p. A3.

15 NAS, “Policy on Committee Composition andBalance and Conflicts of Interest,” May 2003.

16 See, for instance, “NAS Panel Rife With UndisclosedTies to Polluters,” Center for Science in the PublicInterest press release, Jan. 22, 2004(http://www.cspinet.org/integrity/press/200401221.html); and “Coal Ties Tar Science Panel, GroupsSay,” CSPI press release, Oct. 26, 2004(http://www.cspinet.org/integrity/press/200410261.html).

17 Federal Advisory Committee Act, Section 15, 1972.18 Personal communication, William Colglazier, NAS

Executive Officer, to CSPI, Oct. 28, 2005.19 Personal communication, Jim Jensen, NAS Director

of Congressional Relations, Oct. 28, 2005.20 NAS website;

http://www.nationalacademies.org/coi/index.html.21 The National Academies’ 2003 “Policy on Committee

Composition and Balance and Conflicts of Interest22 Personal communication, William Colglazier, NAS

Executive Officer, Oct. 28, 2005.23 NAS, “Background Information and Confidential

Conflict of Interest Disclosure” forms.24 NAS brochure, “The National Academies Study

Process: Ensuring Independent, Objective Advice.” 25 Personal communication, William Colglazier to

CSPI, Oct. 28, 2005.26 NAS brochure, “The National Academies Study

Process: Ensuring Independent, Objective Advice.”27 See “Confidential Financial Disclosure Form for

Special Government Employees Serving on FederalAdvisory Committees at the U.S. EnvironmentalProtection Agency,” at http://www.epa.gov/science1/pdf/epaform3110-48.pdf, accessed Jan. 31, 2006.

28 GAO-04-328, “Federal Advisory Committees –Additional Guidance Could Help Agencies BetterEnsure Independence and Balance,” April 16, 2004,p. 32-4.

29 See http://www.fda.gov/oc/advisory/conflictofinter-est/disclosure.pdf.

30 See http://www.fda.gov/oc/advisory/default.htm.31 Gardiner Harris and Alex Berenson, “Voters on Panel

Backing Pain Pills Had Industry Ties,” New YorkTimes, Feb. 25, 2005, p. 1.

32 See, for instance, Jennifer Washburn, University Inc.:The Corporate Corruption of Higher Education, BasicBooks, 2005; and David Willman, “Stealth Merger:Drug Companies and Government MedicalResearch,” Los Angeles Times, Dec. 7, 2003, p. 1.

33 Quoted in Derek Bok, Universities in the Marketplace,Princeton University Press, 2003, p. 66-7.

34 Troyen A. Brennan, et al., “”Health IndustryPractices That Create Conflicts of Interest,” Journalof the American Medical Association, Jan. 25, 2006, p.430.

35 Carl Elliott, “Pharma Goes To the Laundry: PublicRelations and the Business of Medical Education,”Hastings Center Report, Dec. 1, 2004.

36 Personal communication, Jim Jensen, NAS Directorof Congressional Relations, 10/28/05.

37 Marian Burros, “Advisories on Fish and the Pitfallsof Good Intent,” New York Times, Feb. 15, 2006, p. F5.

38 See http://www.cfsan.fda.gov/~dms/admehg3.html(accessed Jan. 19, 2006).

39 Michael Hawthorne, “What is a consumer to do?The USDA’s food pyramid recommends eating threefish that the EPA and the FDA warn are high inmercury,” Chicago Tribune, Jan. 19, 2006.

NOTES

23

Notes


24

40 See, for instance, the mission statement for theNorth Central Regional Aquaculture Center athttp://ag.ansc.purdue.edu/aquanic/ncrac/splan.htm(accessed Jan. 19, 2006).

41 See, for instance, http://www.nmfs.noaa.gov/media-center/ aquaculture/ (accessed Jan. 19, 2006).

42 Joshua T. Cohen, PhD, David C. Bellinger, PhD, etal., “A Quantitative Risk-Benefit Analysis of Changesin Population Fish Consumption,” American Journalof Preventive Medicine, Vol. 29, No. 4, p. 325-34.The journal reported the study was funded by theNational Food Processors Association ResearchFoundation and the Fisheries Scholarship Fund,whose membership rosters include numerous fishprocessors.

43 CSPI Integrity in Science director Merrill Gooznerwas present at the meeting.

44 CSPI interview with William Colglazier, Oct. 28,2005.

45 GAO-04-328, op. cit., p. 43-4.46 Bekelman JE et al, “Scope and impact of financial

conflicts of interest in biomedical research: a systematic review,” Journal of the American MedicalAssociation, Jan. 22, 2003, p. 454-63.

47 GAO-04-328, op. cit., p. 33.48 “Science and Technology in the National Interest:

Ensuring the Best Presidential and Federal AdvisoryPanel Science and Technology Appointments,” TheNational Academies Press, Nov. 14, 2004.

49 GAO-04-328, op cit., p. 40.50 The National Academies, BI/COI Form 1,

“Background Information and Confidential Conflictof Interest Disclosure for Studies Related toGovernment Regulation,” May 2003.

51 See, for instance, Sen. Kit Bond Press Release, “BondJoins Workers to Celebrate Groundbreaking onMissouri Briggs and Stratton Plant Expansion,”March 21, 2005 (available athttp://bond.senate.gov/atwork/recordtopic.cfm?id=234384).

52 DeAngelis et al., “Reporting Financial Conflicts ofInterest and Relationships between Investigators andResearch Sponsors,” Journal of the American MedicalAssociation, 2001; 286:89-91.

53 See Bekelman et al, op. cit.; and Curfman GD, et al,“Expression of Concern: Bombardier, et al.,‘Comparison of Upper Gastrointestinal Toxicity ofRofecoxib and Naproxen in Patients withRheumatoid Arthritis,’ N Engl J Med, 2000;343:1520-8,” N Engl J Med, 2005; 353:2813-4.

54 See CSPI letter to Donald Kennedy, editor of Science,Jan. 12, 2006, available at:http://cspinet.org/new/pdf/science_letter.pdf.

55 Center for Science in the Public Interest,“Unrevealed: Non-Disclosure of Conflicts of Interestin Four Leading Medical and Scientific Journals,”July 2004 (http://www.cspinet.org/new/pdf/unrevealed_final.pdf)

56 Goehl TJ, “Embracing Scutiny,” Environmental HealthPerspectives, October 2004, available online at:http://ehp.niehs.nih.gov/docs/2004/112-14/editorial.html.

57 Personal communication, William Colglazier toCSPI, Jan. 27, 2006.

58 Pub. L. 92-463, Sec. 15, as added Pub. L. 105-153,Sec. 2(b), Dec. 17, 1997, 111 Stat. 2689.

publisher of Nutrition Action Healthletter1875 Connecticut Avenue, N.W., Suite 300

Washington, D.C. 20009202-332-9110

www.cspinet.org

Documents

Ensuring Independence and Objectivity 140 years, the National Academy of Sciences (NAS) has played a prominent role in setting national policies that are based on sci-ence. A congressional