the role of academia and the research community in assisting the food and drug administration to...

COMMENTARY

The role of academia and the research community inassisting the food and drug administration to ensureU.S. drug safetyy

Vincent Lo Re III MD, MSCE*,z and Brian L. Strom MD, MPH§

Department of Biostatistics and Epidemiology, Center for Clinical Epidemiology and Biostatistics,University of Pennsylvania School of Medicine, Philadelphia, PA, USA

SUMMARY

Purpose Academia can play a prominent role in the drug safety arena, unique from that of industry, and a clearerarticulation of how it could positively influence the current system is needed. We sought to examine ways that academiacould expand its role in U.S. drug safety.Methods An ad hoc meeting of academic experts in drug safety and risk management was convened at the Institute ofMedicine (IOM) in Washington, D.C.Results Academia should develop a stronger partnership with the Food and Drug Administration (FDA) to increaseresearch on regulatory issues and public health questions and facilitate the prioritization of critical issues on drug safety. Sucha collaboration could also facilitate the development of a network of academic centers of excellence in pharmacoepide-miology to address drug safety and risk management questions from a public health standpoint in a timely fashion. Thedevelopment and testing of methodologic innovations on drug safety should also be encouraged.Conclusions Greater partnership between academia and the FDA could facilitate the prioritization of important issues ondrug safety, allow more research questions on drug safety to be answered in a timely fashion, promote the development ofnetworks for answering these questions, and help generate additional research ideas, ultimately providing enormous benefitto the public health. Copyright # 2007 John Wiley & Sons, Ltd.

key words—FDA; academia; CERTs; cooperative agreements; conflicts of interest

Received 28 July 2006; Revised 12 February 2007; Accepted 19 February 2007

INTRODUCTION

Health care providers and patients increasingly expectthat new therapies will have benefits that aredocumented and risks that are accurately detected,quantified, and communicated.1,2 Recent labelingchanges,3 ‘black box’ warnings,4 and withdrawalsfrom the market of high-profile drugs,5 all due tomajor safety issues, have generated concern thatthe current drug regulatory system is inadequate forthe protection of the public health. Because of theselimitations, proposals have been requested to for-mulate reforms that the U.S. Food and DrugAdministration (FDA) could implement to protectthe public health better.

pharmacoepidemiology and drug safety 2007; 16: 818–825Published online 16 April 2007 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/pds.1398

*Correspondence to: Dr V. L. Re III, Department of Biostatisticsand Epidemiology, Center for Clinical Epidemiology and Biosta-tistics, University of Pennsylvania School of Medicine, 711 Block-ley Hall, 423 Guardian Drive, Philadelphia, PA 19104-6021, USA.E-mail: [email protected] conflict of interest was declared.zVincent Lo Re III is an Instructor in Epidemiology, Department ofBiostatistics and Epidemiology, Center for Clinical Epidemiologyand Biostatistics, Center for Education and Research on Therapeu-tics at University of Pennsylvania School of Medicine.xBrian L. Strom is George S. Pepper Professor of Public Health andPreventive Medicine; Professor of Biostatistics and Epidemiology;Professor of Medicine; Professor of Pharmacology; Chair, Departmentof Biostatistics and Epidemiology; Director, Center for Clinical Epide-miology and Biostatistics; Director, Penn Center for Education andResearch on Therapeutics; Associate Vice Dean, University of Penn-sylvania School of Medicine; Associate Vice President for StrategicIntegration, University of Pennsylvania Health System.

Copyright # 2007 John Wiley & Sons, Ltd.

Very few proposals to date have specificallyaddressed ways that academia and the researchcommunity could help to improve the existing drugsafety and risk management system. Academia has aparticular role to play in the drug safety and riskmanagement arena, which is unique from that of bothindustry and government, and a clearer articulation ofhow academia could positively influence the currentsystem is potentially useful. In addition, FDA’s role infostering such a partnership with academia to improvedrug safety and risk management requires a furtherdelineation if such a partnership is to succeed.

To address these issues, we review opportunities foracademia to expand its role in U.S. drug safety. In thiscontext, we discuss the potential involvement of theCenters for Education and Research on Therapeutics(CERTs) to help improve drug safety and riskmanagement. Second, we discuss how academia canimprove the capacity of the FDA in drug regulation.Third, we provide ways for academia to fostermethodologic innovations to advance the drug safetyfield. Finally, we address academia’s special role inconflicts of interest and issues of independence.

EXPANDING ACADEMIA’S ROLE IN THEDRUG REGULATORY SYSTEM

Role of academia on FDA advisory committees

Advisory committees, which are made up of outsideexperts, provide FDA with independent opinions ondraft or existing policies and review specific drugs inspecialized medical areas (e.g., cancer and cardio-vascular disease). This advice contributes substan-tially to the quality of the agency’s regulatorydecision-making and provides credibility to theproduct review process. However, a number ofshortcomings associated with advisory committeescurrently limit their effectiveness.

First, the specific types of drug safety and riskmanagement issues that are chosen by FDA to bringbefore its advisory committees have not been welldelineated. These decisions are made by FDA itself,without guidelines transparent to the rest of thecommunity. There are no evident criteria regardingissues as worthy of this special attention. To addressthis issue, academia could play a more prominent rolein creating a context of education and partnership onthese committees. Academia should collaborate moreclosely with the FDA to evaluate and prioritizeimportant issues and needed studies on drug safetythat require national attention. An oversight advisorycommittee that includes academics involved in the

drug safety research could also be created andconvened, even as often as once per month, to assistin specific decisions on the questions regarding drugsafety that should be brought before advisorycommittees.6

Second, FDA advisory committees are governed bystringent rules that prevent academics from discussingthe content of the meeting in advance or duringmeetings in an informal fashion between sessions,preventing advisory committee members from initiat-ing discussions outside the formal committee struc-ture. These rules therefore prevent important oppor-tunities for education and dissemination ofinformation that would enhance the quality of thecommittees’ decisions. As a result, committeemembers may be unprepared to decide adequatelyon important issues of drug safety that they are askedto review. The formation of working groups ofacademics within advisory committees could alleviatethis problem, but the feasibility of these is unclearunder the current government rules.Third, there has been a considerable recent criticism

of potential conflicts of interest among the advisorycommittee members.7 Furthermore, there are manyother types of conflicts other than just financial (e.g.,intellectual and professional) that need to beconsidered, which often may be more severe. If anyof these conflicts are relevant, then academics servingon these committees should declare all the potentialconflicts and either recuse themselves or continue toparticipate with full disclosure, depending upon thewill of the committee.Fourth, wiser and more extensive use of the Drug

Safety and Risk Management Advisory Committeeshould be sought. This committee was created in 2004to advise the Commissioner of Food and Drugs on riskmanagement, risk communication, and quantitativeevaluation of spontaneous reports for drugs for humanuse and for any other product for which the FDA has aregulatory responsibility.8 The committee also has hada role in evaluating the information on the safety,efficacy, and abuse potential of drugs or othersubstances and recommends actions to be taken bythe Department of Health and Human Services withregard to the marketing, investigation, and control ofsuch drugs or other substances.8 The committeerepresents an opportunity for academics in the drugsafety and risk management arena, particularlypharmacoepidemiologists and biostatisticians, toprovide a valuable input to the FDA. In this capacity,the Drug Safety and Risk Management AdvisoryCommittee members have given expertise in pharma-coepidemiology, risk communication, and pharmacy

Copyright # 2007 John Wiley & Sons, Ltd. Pharmacoepidemiology and Drug Safety, 2007; 16: 818–825DOI: 10.1002/pds

commentary 819

to these discussions, providing a balanced view to thatof clinicians. In particular, to enhance discussions ofdrug safety and risk management on current FDAadvisory committees, the FDA should continue andincrease the seeding of other advisory committeeswith members of the Drug Safety and Risk Manage-ment Advisory Committee. Alternatively, the FDAcould turn to individual members of the Drug Safetyand Risk Management Advisory Committee forconsultative experience. As such, academics withpharmacoepidemiologic expertise could be broughtinto the drug development process earlier, potentiallyas consultants to the FDA. Finally, the FDA shouldcontinue to convene the committee jointly with otherspecialty committees when crucial issues of drugsafety emerge, since their expertise could help theFDA in these difficult circumstances. Given thenumber of such requests already, the size of the DrugSafety and Risk Management Advisory Committeemay need to be increased so that the workload for eachcommittee member is not prohibitive.

Revisions in the structure of FDA cooperativeagreements

Academia and the research community provide theFDA and industry with an enormous amount of dataabout the safety of drugs, and FDA cooperativeagreements represent an ongoing way for the researchcommunity to deliver such information. Historically,cooperative agreements between the FDA andacademia functioned like grant mechanisms, withsome additional input by the funding agency. Principalinvestigators at academic medical centers had thefreedom to formulate research questions on drugsafety themselves, and these agreements representedopportunities for FDA to obtain pharmacoepidemio-logic expertise from academia that helped to advancean overall knowledge on drug safety and riskmanagement. However, over time, the structure ofcooperative agreements has shifted closer to function-ing like contract mechanisms, used primarily to obtainand analyze data rather than expertise. Indeed, theyhave now changed formally to contracts. Cooperativeagreements have thus become task-oriented mechan-isms that are no longer cultivating a collaborativerelationship between FDA and academia, to thedetriment of both. The structure of cooperativeagreements should therefore be revised to allow theirfocus to shift back to functioning as grant mechanismsrather than merely data purchasing mechanisms. Thismechanism could fund a range of studies that a publichealth agency needs, particularly drug utilization,

surveillance for adverse effects, and specific hypothe-sis-oriented research. Academic centers could partnerwith the FDA to help set the research agenda for thesespecific areas, identify the appropriate data sourcesthat could answer the questions, and design andimplement the studies. Such modifications to FDAcooperative agreements could allow the FDA to utilizeacademic expertise more effectively and promote agreater spirit of partnership between the FDA andacademia.

The FDA should also fund the extramural drugsafety research at a markedly increased level. Thisincreased funding could be used to allow centers ofexcellence in epidemiology, outcomes of research, andhealth economics to expand and flourish. The FDA’sdrug safety cooperative agreement program was quitesmall in the 1980s, at which time it sought toinvestigate the safety of all the drugs with which wasconsidered a modest funding level of approximately$1.2 million in 1985. Since that time, resourceshave varied, despite the inflation in the interim, andfunding for FDA drug safety cooperative agreementsreached a low of $900 000 in 2000 (personalcommunication, G. Dal Pan, FDA, March 25,2007). For FY07, funding for FDA drug safetycontracts (the replacements for the cooperativeagreements) totals only $950 000. Since this is tocover all the drugs, it is especially meager whencontrasted with the more than $800 million spent indeveloping a single new drug.9,10 The Institute ofMedicine (IOM) estimates that at least 10 such drugsafety signals per year could be evaluated extra-murally, at an annual cost of $10–60 million.11

Improving the timeliness of academia’s input

One understandable concern with the role ofacademics in providing data needed for regulatorydecisions is the history of a lack of timeliness inacademics’ work. In particular, academics do notroutinely conduct research under time constraints forfear that this might compromise the quality of theirwork. In addition, as non-profit entities, academiccenters cannot support a staff in reserve to be ready fora project that required a rapid turnaround time, were itrequested.

These shortcomings could be addressed by atwo-part suggestion. First, as noted above, FDA(and/or other partners) could support the developmentand maintenance of centers of excellence in drugsafety and risk management, potentially the existingCERTs, described in the next section. These centerswould then have the infrastructure, particularly a


820 v. lo re iii and b. l. strom

critical mass of faculty and staff, biostatistical support,active pharmacoepidemiology training programs, andaccess to pharmacoepidemiologic databases, to beable to respond in a timely fashion when questions ofdrug safety emerge. Second, specific projects could beplanned in two phases. The first phase would providepreliminary results, sufficiently definitive to support apublic health decision. The second phase would thenencompass more comprehensive, secondary dataanalyses, with the time for final manuscript editing,peer review, and publication. To support the needs ofthe academic investigators, funding commitmentswould need to be maintained for both phases so as toensure the completion through the second phase. Thenet effect will be a higher cost but an earlierpreliminary answer.

THE POTENTIAL ROLE OF THE CERTs

The FDA was reviewed by Congress in 1997, withextensive input from academia, the pharmaceuticalindustry, and FDA itself in deliberations leading to theFDA Modernization Act.12 In it, the FDA’s chargeshifted from solely being a regulatory agency tobecoming a full service public health agency that wouldadvance the health of the people through improvedtherapeutics. One of the specific interventions recom-mended was that there should be a program in researchon therapeutics to improve the nation’s practice in themonitoring and use ofmedications, including drug safetyand risk management.12,13 To that end, the CERTs werecreated in 1999 to increase an awareness of the benefitsand risk of drugs, biologics, and devices and conductresearch to improve therapeutics.13,14 Thus, these centerscould, in part, help to build the capacity that FDA couldturn to in order to allow critically important questionsabout drug safety to be answered in a timely fashion.This would, in turn, foster a vision of the FDA partneringwith academia to act as a public health agency ratherthan merely a regulatory institution. Indeed, in thismodel, academics would conduct drug safety researchwith interpretation by and partnership with FDAscientists.

The CERTs seek to improve the health care andpatient safety. The role of these centers is to: (a)develop and nurture public–private partnerships tofacilitate a research on therapeutics; (b) support andencourage research on therapeutics that is likely to gettranslated into policy or clinical practice; (c) developeducational modules and dissemination strategies toincrease an awareness of the benefits and risks ofpharmaceuticals; and (d) create a national informationresource on the safe and effective use of thera-

peutics.13 Funding to support the CERTs was providedby Agency for Healthcare Research and Quality(AHRQ), a research agency committed to improvingtreatments and healthcare delivery in general with thecollaboration from and a dual supervision under theFDA. AHRQ succeeded in funding four centers in thefirst wave and three additional ones in the secondwave. These seven centers are led by a coordinatingcenter at Duke University, which coordinates theeducation and research on therapeutics. Thus, eachCERT is mutually supportive of one another. As of thisdate, four new CERTs were just funded, as well.Although the creation of the CERTs represents an

auspicious beginning to the advancement of thenation’s knowledge on therapeutics, each CERT’sbudget is currently extremely small given its charge($800 000 per year, including indirect costs), and istherefore far from sufficient for its stated goals. As aresult, each CERT must spend a substantial amount oftime raising the additional money for its successfuloperation, and the actual research output from theCERTs, while substantial, is less than what it couldbe. The CERTs’ program had a total annual budgetof only $9 million in fiscal year 2006 (L. Bosco, MD,written communication, AHRQ, March 27, 2007). Incontrast, the $70 billion expended annually ondrugs as a part of Medicare Part D15 will dramaticallydwarf the amount spent by the CERTs on ensuringthat they are used correctly, as will the $33.2 billionspent by the pharmaceutical industry in 2003 onresearch and development,16 and the $15.7 billionexpenditure on the promotion of drugs in 2000.17

The FDA and/or AHRQ will therefore needmarkedly increased resources to implement such amodel.As noted above, there is a need for a group of

centers of excellence in pharmacoepidemiology,which can be accessed by FDA, industry, etc., toanswer specific projects. The existing CERTs couldserve this role although other centers could be selectedin addition or instead. However, to be able to addressbetter drug safety and risk management questionsfrom a public health standpoint in a timely fashion,additional funding would be needed to maintain theexisting CERTs and to allow an increase in theirnumber (to a total of perhaps 20). This additionalfunding could allow more research on drug safety tobe performed with a more rapid turnaround time. Itcould also allow the development of the educationalcurricula needed to train the next generation ofpharmacoepidemiologists and biostatisticians to beable to perform this research. Additionally, the CERTscould and should partner with academic institutions


commentary 821

that do not have such centers so that much of academiawould be a part of the CERTs’ network. Since thecurrent R01 grant application process does not allowresearch questions on drug safety to be answered in atimely fashion, expansion of the role of the CERTs,supplemented by a project-specific task order process,as noted above, would allow the nation to performspecific research on drug safety more quickly as suchissues arose. Thus, FDA, industry, and other agenciesshould provide an infusion of funding to the CERTs, oranother similar network, to ensure that academia canoptimally provide data on drug safety.

ACADEMIA’S ROLE IN IMPROVING THECAPACITY OF THE FDA

Personnel in place and training of personnel

At present, there are extremely few academic centerscapable of providing appropriate training in pharma-coepidemiology, and insufficient numbers able toprovide training in biostatistics. This has led to aconsiderable dearth within the nation of adequatelytrained personnel in drug safety and risk management.Without such personnel, the U.S. will not be able toproperly ensure drug safety. To address this lack oftraining, academia must take a lead by increasing andpromoting education in pharmacoepidemiology andbiostatistics, so that there will be a steady pipeline ofwell-trained personnel to manage drug safety in theU.S. Such training could be done within the academiccenters of excellence to be established, as noted above.To improve the training in drug safety and risk

management, career development awards for individ-uals who wish to enter the field of pharmacoepide-miology should be funded. Historically, these awardshave only been available from industry-relatedfoundations, but these are clearly not sufficient tomeet the needs of the field, and are no longer available.The federal government has never offered suchawards, except to those whose interests matched theinterests of categorical institutes of the NIH, allowingthem to apply to those institutes for Mentored CareerDevelopment Awards, like the K01, K08, or K23Awards. At the current time, the only federal fundingavailable for training in pharmacoepidemiology iswithin the CERTs’ program. The National Institute ofGeneral Medical Sciences (NIGMS) is about to awardits first pharmacoepidemiology training grant out of itsclinical pharmacology training program. However,while a very positive movement, it will only containtwo slots per year, and this is far too small to meet theneeds of FDA alone, no less the field in general. This

leaves the field of pharmacoepidemiology dependentlargely on donations from the industry and preventsthe opportunity to train more individuals to meet theenormous personnel needs.

While working to provide trainees the basiccompetencies in pharmacoepidemiology, trainingprograms should encourage the participation in publichealth organizations devoted to drug safety, such as theFDA. In addition, academic centers should maintain apartnership with their trained personnel to allow themto remain current in the field and promote theutilization of new techniques as they may arise. Tofoster further a culture of partnership between theFDA and academia, active membership in andattendance at meetings of professional societies(e.g., International Society for Pharmacoepidemiol-ogy and American Society for Clinical Pharmacologyand Therapeutics) by officials of the FDA should beincreased. The promotion of such a strategic alliancewill help to maintain training and currency in the fieldand promote a greater collaboration between membersof academia and the FDA. It will also ensure that theresearch of academics is informed by the real-worldissues being confronted by the FDA.

Finally, in considering other educational require-ments, there is also a critical need to improve thepharmacoepidemiologic training to the prescribingcommunity in general. Health care providers currentlyreceive very little, if any, instruction or training onpharmacoepidemiology, risk-benefit concepts formedications and devices, and pharmacovigilance.Incorporation of this training more formally into theexisting curricula will help health care providers tounderstand better the risks of using new drugs and theissues surrounding uncertainty of drug effects. Sucheducation will also enable a proper interpretation ofemerging data in the medical literature. This is a majoreducational task that needs to be undertaken at all theschools of medicine, pharmacy, dentistry, etc., and thedesignated centers of excellence in pharmacoepide-miology can certainly serve as leaders in this effort.

Data resources

Data have become a valuable commodity, and largeelectronic databases of claims or computerizedmedical records are attractive options in pharmacoe-pidemiology, as they permit relatively quick andinexpensive studies. However, their proper userequires substantial scientific and programmingexpertise. As such, the FDA should engage academicexpertise to develop and make available databases thatwill allow pharmacoepidemiologic research to deter-



mine drug safety more definitively, so as to reduce thelevel of uncertainty surrounding drug adverse effects.It is not sufficient to simply purchase access to thedata. In particular, the development of a closerpartnership between the FDA and academia mayallow a better use of such rich data sources.

As a particular example, there are presently no plansin place to establish research databases containingMedicare Part D data, an enormous public resourcethat could be used for numerous pharmacoepidemio-logic studies of drug adverse effects. What better wayis there to examine longitudinally the drug safety andlong-term drug adverse effects in the elderly than byevaluating the experience of the large number ofpatients who actively use these medications? Plans toutilize such a database need to be established with anappropriate input from the academic research sector.The data must then be made accessible to the generalresearch community as soon as practical. To dootherwise is a waste of an enormous potential resourceat a great cost to the public’s health.

In addition, the FDA is in control of a huge amountof pre-marketing data, across all marketed products,including potentially valuable information fromplacebo-treated patients. These data could be enor-mously useful in determining the comparative effectsof drugs. They could also be used for studies that aimto determine the pre-marketing signals that arepredictive of subsequently discovered post-marketingsafety problems.18 Academic investigators couldassist in performing such studies, but FDA needs topermit an access to them for this to be feasible.

Academia is also in control of an increasing amountof primary health data, given the computerization ofacademic health care systems. The proposed FDA/academic partnership described throughout this reportwould also allow academic centers to provide FDAwith an access to their data and their patients.

Given academia’s role in facilitating the interpret-ation of drug safety data, better data standards need tobe established. Development of a national electronicmedical record systemmay eventually contribute to animproved knowledge on drug safety, and the analysisof population data across care sites may help to discernadverse effects of drug treatments. In addition, newoutcome definitions are continually being used withvarious data sources, and this inconsistency makes itdifficult to compare among studies. One couldentertain a standardization of outcomes. However, itis often from the lack of standardization and theresulting lack of consistency in the results that greatinsight ensues. At the least, to address this issue, itwould be beneficial to have transparency of clinical

outcomes and adverse events based on similar types ofdata sources (i.e., investigators should freely share thecodes that they used).

Academia should promote a model of partnership

To promote a greater research collaboration betweenFDA and academia, a model could emerge wherebythe FDA could utilize networks of centers ofexcellence in pharmacoepidemiology, such as theCERTs described above, that possessed sufficient coresupport and infrastructure to be able to answerquestions of public health importance in a methodo-logically rigorous fashion. Cooperative agreementscould be created to allow investigators withinacademic medical centers to pursue research thatthey deem to be of public health importance, incollaboration with FDA scientists. In addition,requests for work on specific drug safety or riskmanagement questions that required a rapid timelineor that were of regulatory concern could be issued bythe FDA to this network to ensure that these issueswere addressed expeditiously. This is similar to theapproach now used by AHRQ in its DevelopingEvidence to Inform Decisions About Effectiveness(DEcIDE) Network, but DEcIDE does not providecore support to its awardees.

PROMOTION OF METHODOLOGICINNOVATIONS

Methodologic innovations about drug safety can helpadvance the drug safety field. Examples of recentmethodologic innovations include adaptive clinicaltrial designs, pharmacogenomics, and use of datamining techniques to screen structured data sources togenerate hypotheses about drug safety issues. How-ever, the development of such innovations is sloweddown by a lack of funding for such activities. Onecannot expect the industry alone to fund suchactivities, which are not product-specific. Yet, NIHspends little on methods’ development. The centers ofexcellence described above, however, could representopportunities to do such work.In addition, new methodologic innovations are

being implemented only slowly in the field, and thedrug safety enterprise has heretofore provided fewrewards for such innovations. The FDA is slow toaccept new approaches, and industry seeks to performonly what it knows the FDA will accept. Academia,FDA, and rest of the drug safety research enterprisemust partner to design new approaches to thedevelopment and testing of methodologic innovations.


commentary 823

Then, once demonstrated to the FDA’s satisfaction, itshould provide a feedback to sponsors that it wouldwelcome data that emerged from such newapproaches.There also need to be major innovations in the

existing risk management programs.19 These pro-grams are designed to optimize the use of newlyapproved drugs, improving the risk/benefit balance ofmedications that otherwise might be marginal.20 Atpresent, evaluations of risk management programs areoften only shared with the FDA. Evaluations of theseinterventions need to be made public so that theadvantages and disadvantages of each program can bedetermined and methods of improvement formulatedin a public forum. Without this, there is great risk ofsimply repeated implementation of ineffectiveapproaches to risk management.

ACADEMIA’S ROLE IN CONFLICTS OFINTEREST AND INDEPENDENCE

Conflict of interest has become an acutely sensitiveissue during the process of evaluating drug safety,particularly since members of academia can serve asboth researchers and consultants, funded by industryor government or both.7 Censorship can and doesoccur, at least as often with a government sponsor aswith an industry sponsor. Since a greater partnershipbetween the drug safety enterprise and academia isrecommended in this report, the need for academicresearchers to maintain their independence, from bothindustry and government, must be addressed.This report encourages not just funding of

academics, but true partnership with academics,including formal collaboration. However, when aca-demics collaborate with members of either industry orgovernment, the sponsoring agency frequentlyrequires review and approval, and such a reviewcan contribute substantively to a manuscript. How-ever, reviews by sponsoring agencies can also be aserious impediment to collaboration if they lead toinfractions on academic freedom and censorship ofpublication. In particular, if results are found not to beconsistent with the views of the study sponsor, reviewof a manuscript has been known to be deliberatelydelayed to avoid publication (i.e., pocket veto) or hasrequired excessive revisions by the sponsor that caninfluence the study’s conclusions. To avoid thesesituations and foster a spirit of partnership betweenacademia and the drug enterprise, use of suchcapricious rules should be condemned. Timelinesfor manuscript review should be reasonable, and onlypersonnel who are knowledgeable in the area of

research should act as reviewers. Furthermore, if anacademic cannot reach a suitable agreement during asponsoring agency’s approval mechanism, then thecollaborators who represent the sponsoring agencyshould be removed from authorship and a note of thismade in the Acknowledgements section of themanuscript. Routine insertion of language withincontracts to avoid such outcomes should also beconsidered as a deterrent.

CONCLUSIONS

As described above, academia could positivelyinfluence the current U.S. drug safety and riskmanagement system. We have proposed a model ofgreater partnership between academia and the FDA.This partnership could allow an expansion of theexisting network of centers of excellence in pharma-coepidemiology to allow more research on regulatoryissues and public health questions. Such a collabor-ation could also facilitate the prioritization ofimportant issues on drug safety, allow more researchquestions on drug safety to be answered in a timelyfashion, promote the development of networks foranswering these questions, and help generateadditional research ideas, ultimately providing enor-mous benefit to the public health.

ACKNOWLEDGEMENTS

The recommendations in this article were created atthe request of the Institute of Medicine (IOM) Com-mittee on the Assessment of the U.S. Drug SafetySystem. They were informed, in part, by an ad hoc

KEY POINTS

� Greater partnership between academia and theFDA should be fostered to allow more researchon regulatory issues and public health questionsand facilitate the prioritization of critical issueson drug safety.

� A network of academic centers of excellence inpharmacoepidemiology is needed, and thesemust have core support for their own activitiesand must be provided with supplemental supportrapidly when questions of importance arise.

� The development and testing of methodologicinnovations on drug safety should be encour-aged, and the FDA should provide guidance ontheir use when convinced of their validity.



meeting group that convened on 7 March 2006 at theIOM headquarters in Washington, D.C. The membersof this group were convened by Brian Strom, MD,MPH, and included Robert Califf, MD; WilliamDuMouchel, PhD; Marie Griffin, MD, MPH; JoshuaMetlay, MD, PhD; Allen Mitchell, MD; Richard Platt,MD, MSc; Sebastian Schneeweiss, MD, ScD; JosephSelby, MD, MPH; and Hugh Tilson, MD. Dr P.H.David Korn, MD and Robert D. Gibbons, PhD werepresent and represented the IOM Committee on theAssessment of the U.S. Drug Safety System. KathleenStratton, PhD; Rose Marie Martinez, ScD; and AmyGrossman were present as representatives of the IOM.Finally, Vincent Lo Re III, MD, MSCE also attendedthe meeting and drafted the document in collaborationwith Brian Strom, initially as a report to IOM. Theyassume responsibility for the final paper.

REFERENCES

1. Fontanarosa PB, Rennie D, DeAngelis CD. Postmarketingsurveillance: lack of vigilance, lack of trust. JAMA 2004;292: 2647–2650.

2. Ray WA, Stein CM. Reform of drug regulation: beyond anindependent drug-safety board. N Engl J Med 2006; 354: 194–201.

3. US General Accounting Office. FDA Drug Review: post-approval risks 1976–85. Washington DC: US General Account-ing Office; 1990: 24. Publication GAO/PEMD-90-15.

4. Lasser KE, Allen PD, Woolhandler SJ, Himmelstein DU, WolfeSM,Bor DH. Timing of new black boxwarnings andwithdrawalsfor prescription medications. JAMA 2002; 287: 2215–2220.

5. Topol EJ. Failing the public health: rofecoxib, Merck, and theFDA. N Engl J Med 2004; 351: 1707–1709.

6. Wood AJ, Stein CM, Woosley R. Making medicines safer: theneed for an independent drug safety board. N Engl J Med 1998;339: 1851–1854.

7. Steinbrook R. Financial conflicts of interest and the Food andDrug Administration’s advisory committees. N Engl J Med2005; 353: 116–118.

8. Food and Drug Administration. Committee Charter: DrugSafety and Risk Management Drugs Advisory Committee.2004 [Accessed 13 March 2006, at http://www.fda.gov/cder/audiences/acspage/DSaRMcharter.htm]; 2004.

9. Adams CP, Brantner VV. Estimating the cost of new drugdevelopment: is it really $802 million? Health Aff (Millwood)2006; 25: 420–428.

10. DiMasi JA, Hansen RW, Grabowski HG. The price of inno-vation: new estimates of drug development costs. J Health Econ2003; 22: 151–185.

11. Institute of Medicine—Committee on the Assessment of the U.S. Drug Safety System. ‘‘The Future of Drug Safety: Promotingand Protecting the Health of the Public.’’ 2006 [Accessed 5February 2007, at http://www.iom.edu/CMS/3793/26341.aspx].

12. Golodner LF. The US Food and Drug Administration Modern-ization Act of 1997: impact on consumers. Clin Ther 1998;20(Suppl C): C20–C25.

13. Califf RM. The need for a national infrastructure to improve therational use of therapeutics. PD Saf 2002; 11: 319–327.

14. Centers for Education&Research on Therapeutics. Home Page.[Accessed on 22 January 2007 at http://www.certs.hhs.gov/index.html].

15. Datamonitor. Medicare Part D: its impact on US drug pricingissues and role as a driver towards European style pricingmechanisms. December 2005. [Accessed 14 March 2006,at http://www.medadnews.com/special_reports/sample.cfm?reportID¼208].

16. Pharmaceutical Research and Manufacturers of America.Pharmaceutical Industry Profile 2004. Washington DC:Pharmaceutical Research and Manufacturers of America;2004.

17. Rosenthal MB, Berndt ER, Donohue JM, Frank RG, EpsteinAM. Promotion of prescription drugs to consumers. N Engl JMed 2002; 346: 498–505.

18. Strom BL. Potential for conflict of interest in the evaluation ofsuspected adverse drug reactions: a counterpoint. JAMA 2004;292: 2643–2646.

19. Centers for Education and Research on Therapeutics (CERTs).Risk assessment of drugs, biologics and therapeuticdevices: present and future issues. PD Saf 2003; 12: 653–662.

20. Kramer JM. Managing the risks of therapeutic products: pro-ceedings of a workshop. PD Saf 2005; 14: 619–628.


commentary 825

the role of academia and the research community in assisting the food and drug administration to...

Documents