Tobacco Documents Online

. w.... ..~.... .- . .... ..,. w. <cn ,vn:niN~u SECTION EDITOR: FRANCFSJ. STORttS, MD Independent Investigators and For-Profit Companies Guidelines for Biomedical Scientists Considering Funding by Industry Mary-Margaret Chren, MD A s physician-scientists, we balance many obligations in our professional and personal lives. Schooled in the tradition of patient care, teaching, and research, we learn to juggle and choose among competing responsibilities, competing duties, competing accountabilities. Yet we value :adependence in our aca- detnic lives.' We strive to be independatt investigators, by which we mean free to choose our own scientific goals and to set our own course toward those goals, free from reliance on others for scientific cre- ativity. Much of our maturation as scien- tists is directed toward asking questions, designing studies, orchestrating experi- ments, analyzing data, and publishing re- sults as independently as possible. We do not often think of competingacoounnbili- ties as restricting the independence of an investigator. Yet, in the sense that our lib- erty to do our science to pursue truth is restricted by those to whom we are an- swerable, whether consciously or uncon- sciously, our independence is jeopar- dized. In this essay, I propose ideal guide- lines for physician-scientists.vho are con- sidering funding by for-profit companies. Much hasbeen written aboutpot,ential and real conflicts of intrsest for academic sci- entists in these interactions?'u and I will not disctus the reasonfngtlutt jtistifies such concern. Instead, beginning with the premise that such interactions may cre- ate situations that threaten our indcpen- dence, I will present a stance with respect to these companies that will serve as a foun- dation for specific guidelines. I will oh- Front the Depa.tmcnt of Dennatotogy, Cteveland (Ohio) Vetaans Affairs Medical Center and University Hospitair of Ctcvdand, and Case Westem Reserve Universlty School of Medicine, Clevdand ARCH DEStMATO1NdL 130, APA 1944 432 serve that our accountabilitie.s, ie, the ob- ligations.ve feel to concepts that motivate us or to entities that support us, vary in their importance. Using eaasting models in business and publicly funded research, I.vill suggest two changes Crom conven- tion: first, that our accountabilitics to for- profit companies should always be de- fined by contracts, and second, that an independent not-for-profit instltutional in- termedixry should be responsible for all interactions beaeeen physician-scientists and companies, to ml nim z. the oppor- wnities for noncontractual relationships between the two. My aim is to clarify ac- countabilities, with the goal of facilitat- ing our quest to become independent in- vestigators. Physiciatts who are not primarily scholars often perform studies paid for by companies, and scientists occasionally carry out discreet tasks on contract for in- dustry. While the conceptual framework I will develop applies most clearly to acn- d:emic scientists performing independent investiga,tinn, rnany of the advantages of the resulting guidelines for ensuring the interests of both physician and company apply also in these different, but related, situations. ACCOUNTAB1LiTIES: ESSENTIAL OR CONTEtACTUAL i rcl tn be filll co W} tio dis PC tta, tan a tl ~ To whom are we answerable in our pro- V 0 fcssi.onal lives as physician-scienrisrs? I pro- 5 ~ NOTICE THS MATERA!- MAY BE TED BY COPYR/GHT LAW W r.9 PROTEC (TrLE 17 lf.S. CODE) N 4* 4k Q 1

f-rom: Jcan ak rax i o: Loreen wica7pin 4 r s )s posc that we distinguish lxtween accountabilities that are essential, and those that are contractual (FlIwm 1). By essential aceourtt- abilities. I mean crucial, unavoid- able responsibilities foreach of us by our very existence as physician- scientists, obligations that endure and cannot be simply discharged once and for all. For example, we feel an essential accountability to seek truth and avoid bias in our science, to give excellent care to our pa- ticnts while becoming expert teach- ers and performing exemplary rcsearch. Another essential account- ability ariscs from our personal am- bitions for academic success- pressure to achieve exciting results is compelling since our profes- sional futures depend, in part, on publici,y through publicatioa1° Fi- nally, we feel compelling account- ability to our patrons-the men- tors, advisers, and other benefactors who have supported us intellectu- ally or financially. Satisfying these various accountabilities by recog- n izi ng and valu ing their goals is c hal- lenging, precisely because what is re- quired of us is often not clear, cannot he written in a set of rules, and re- quires continuous effort through- out our professional lives. ONTRACTUAL ac- countabilities, on the other hand, are those in which our responsibilities are related to the situation in a con- tract; theyare ckarty limited, and can be satisfied once and for all by ful- filling the contract. For example. a contractual accountability exists when a scientist receives funding, from whatever source, to perform a discreet task, even if the task is the performance of experiments. Con- tractual accountabilities arc impor- tant and justice requires their dis- charge, yet they are less compelling than essential accountabilities in which obligations are more perva- sive and less easily satisfied. uaie: 'i'u"iya uii". WHEN ACCOIJNTABI[1TIES CONFLICT Serving multiple accountabilities may be challenging enough from a time management standpoint, but more demanding is the fact that the attendant obligations may conflict. For example,, our personal dcsire to succeed academically by publish- ing exddna results may suvggl e wi th our conviction that further experi- ments are needed. Our commit- ment to our patients may conflict with otu need to devote energy and time to teaching or research. Ac- countability of a researcher to the agency that funds his or her re- search may also generate tensions be- cause research sponsors have cer- tain goals for the research they support, and the researcher may feel an implicit obligation to work to- ward these goals. For federally funded agencies, these goals may in- clude congressional and public pri- orities; for private not-for-profit foundations, the goals are deter- mined by boards that ensure that re- sources are used to further the mis- sion of the organization. For-profit companies, while they may share same altrttistic goals with other fund- ingagencies, also have responsibili- ties to ensure corporate iongeviryand serve shareholders by maximizing profits. A real concern is that medi- cal scientists supported by for- profit companies may be influ- enced to feel accountable to those companies and their goals, in ways that conflict with other essential re- sponsibilities, and that are not eas- ily specified and fulfilled by contract- A PROPOSED SOLUTION In my stratification of accountabili- ties, conflicts of interest are more confusing if they involve essential accountabilities because of their power and lack of definition. Simi- larly, structuring obligations, where possibie, Into contractual account- abilities may clarify or diffuse some conflicts of interest, because ARCH DERMAIOWOL 130, APR 1944 433 responsibilities and compensatloII would then be explidt and open for scrutiny, unacceptable obligations would be clarified, and appropriate disclosure would be cssicr.4 Given concerns about conflicts of interest if physician-scientists feei account- able to obligations of for-profit com- p.nies.2"24 I suggest that, where pos- sible, we restrict our research relationships with for-profit com- panies to contractual accountabili- ties. Interestingly, the use of con- tracts to define the litnits of each party's obligations is standard prac- tice in corporate-corporate interactions. Besides composing the con- tnct, how do we ensure that ourac- countability to comp.nies is re- stricted to thrt contract? The perception of an essential account- ability can be craited by the estab- lishment of a rdationsltip bet*xeen two parties-like that between a tnentor and student, a physitian and patient, or a patron and benefi- dary, the alliance between the two engenders a sense of respoasibility with its attendant duties. In the course of designing, planning, and executing a research project with a for-profit company, physician- scientists have many opportunities to develop significant personal re- lationships with the cotnpany ard its staff. Given the concerns I have out- lined about essential-accountabtli- ties of scientists to companies, it makes sense to limit these relation- ships as much as possible. like oth- ers who have suggested that corpo- rate research support be in the form of gnnts to universities rather than to individuals,16 i propose thatallne- gotiations between scientist and company be performed by an ind,e- pendent third party, to minim.ize op- portunities for the development of noncontractual relationships with the scientist. Thus, I have constructeda oon- ceptual framework for guidelines for physician-scientists whose work is funded by industry, to preserve thea independence as investigators

., , ,.. ax i u. "o,._:_n m..eyw tirrs 1. To wAoru aa pfty%kAarrac*tltrta.nrtwraW !n ttwb ptaAmsioad /isw7 Acdrporineian of axainhfit7la of pM'rhw"CArrthif. (ftwe 2). In this scheme, the in- dependent investigator is perceived as serving many crucial and coat- petirtg accountabilities. I have sug- gested the advantages of preserving his or her independence as much as possible by diminishing the poten- tial for conflict-diluting for-profit motives from his or her essential ac- countabilities by tttaking all fund- ing relationships with companies purely contractual, and lessening the opportunities for noncontractual re- lationships by having all interac- tions between companies and sci- entists through an independent institutional intermediary. The guidelines built on this framework have two keystones that derive directly from existing models in business and publicly funded research agencies- contracts are standard practice in corporate relationships, and the iateratediary of the National Insti- tutes of Health (NIH) functions both to establish a uniform com- petition for funds and to ensure scientific quality by peer review. GUIDELINES FOR PHY S iC IAN-SC [E N T I S TS FUNDED BY FOR-PROFIT COMPANIES All obligations of the scientist and the company should be defined in a detailed Contract (l4bf.). The Fipn 2. A caeatplual ham.r»rk !ar pu+daYnas to prsswtv ft ind.pwidana ot phyrlcirn-sdWlsts +Mtosa wwt /t (~ by Jndus?y. written Contnct between funding agency and researcher, to be com- posed by an independent institu- tional Intermediary (see below), should be signed by both parties be- fore any money is given. Although not a panacea for conflicts of inter- est,i0 explicit delineation of the re- spective obligations of the scientist and the funding agency probably minimizes more powerful tacit ob- ligations that arise from relation- ships °-* The Contract should be a standard form used for all compa- nies,24 and should emphasize the in- dependence of the investigator." Specific topics to be covered in- clude discreet statements about the scientist's obligation to promote health, to patients in the study (if ap- plicable),4 or to other scientists (to communicate findings). For ex- araple, the Contract should state that the investigator is obliged to ensure that patients in a study are told about the source and amount of funding of studies in which they partici- pateu and receiving results of the study."Similariy, the obligations to the company should be outlined, in- cluding reference to its specific goals for the study. Many of the guide- lines that follow can be incorpo- rated in the written Contract. A not-for-proHt ir~stitutional In- termediary should be responsible for all communication between the scientist and the company. This ar- AACH DERMATOWOL 130. APR 1994 434 rangement ensures that, as much as possible, the sciattfst is free from bal- ancing ututecessary competing ob- ligations that may arise from devel- oping a relationship with a campeny or its staff. Although it may be per- ceived as a negative or restrictive code, its primary purpose is posi- tive, to preserve the independence of the investigator. Modeled on the NIH (which essentially functions as an Intermediary between Con- gress and the scientist), the Inter- mediary would arrange all initiil communications with the com- pany about its desire to fund re- search based at an academic site, would publish a request for appli- cations approved by the company, and would organize scientific study sections that would judge submis- sions and award the support based on academic merit. All funding would be from the company to the Intermediary; Zand through the In- termediary to the scientist, and all negotiations would be handled through the intermediary, which should be independent of both the company and the scientist's aca- demic site. The Interatedisuy would also write the standard Conitact dc- scribed above and would transmit periodic communications from the scientist about the progress of the work. Investigators would be free from receiving gratuities and spe- cial favors Erom research sponsors, SI ir

I ... ~..,~c.......o,,..., situ;itions recogni2ed to risk impart- ing hias in research.' The mandate of the Interrnedi- an• would come from the profes- sian: its composition-largely ad- ministrative personnel who would orchestrate communication about studies, cotnpetitive applications for thcm, and dispersal of funding from companies to scientists-would be drtermined by representatives from each specialty. The costs of the in- termediary would be paid by pooled funds from industries submitting re- quests for applications. Scientific freedom of the in- dependent Investigator should be preserved. The scientist will have complete freedom to determine study hypotheses, design, inclu- sion ofcomparative entities, and data interpretation. Many aspects of sci- cntific analyses can be structured tconsciously or subconsciously) to produce favorable results.'6 These as- pects of the scientific enterprise and the challenges they present should, as in publicly funded research, be the responsibility of the scientist, whose independence from competing ac- countabilities has been preserved as much as possible. The scientist retains owner- ship of data and publication rights. Data generated in a study would be- long to the investigator,u and the sci- entistwill have publication rights to these da ta?` including the timing and site of publication. This stipulation includes the publication of 'nega- tive" results. The single restriction of this aspect is the right of the com- pany to file for patent rights'"-` as- suming there is no compelling rea- son to publicize results immediately, a point that could be determined by the Intermediary at the time of the Contract or at any time during the research. The Intermediary would also be responsible forensuring that any proprietary information of the company is handled in a just man- ner to ensure the company's rights. The scientist controls the use of his or her name in promotions. Scientists have obligations to share research information in an undis- toned way to promote the public good, which has implications for the promotion of products' In general, the participation of the scientist in promotional activities for products that may have resulted from the re- search should be discouraged, and the scientist controls the use of his or her name and the data in all pro- motions or other publications by the company, except those which quote the medical literature. Funding for a project should not be contingent on results. In gen- eral, support should cover the completion of a study or series of studies so that funding cannot be withdrawn if it appears that certain hypotheses will be disproved./6 Reimbursement for a study should equal the direct and indi- rect costs of the work. Payment for a studyshould approximate the costs of the work, and detailed budgets should be submitted to the Interme- diary for review as part of the grant application. Research that gener- ates excessive profits for investiga- tors, their departments, or their medical centers is worrisome.10•22 since, theoretically, this incentive could exert undue pressure on the scientist to generate provocative re- sults, or to propose enrollment of pa- tients in trials when it is not in their best interests.2i Certain relationships be- tween scientists and for-profit com- panies during the time of the re- search should be lintited. A major ARCH OERMATOLNOL 13o.APR 1094 435 goal of these guidelines is ensuring the independence of investigat.ors by limiting their relationships (and at- tendant obligations) with for-profit companies who may support their research. While the ability of a sci- entist to use proprietary informa- tion to guide profitable purchases or sale of securities ("insider trad- ing") is prohibited by law, the Con- tract should be more explicit about other relationships between re- searchers and companies. Liauted model arrangements have already been drafted's From the tune the re- search is funded until the results are disseminated to the public, investi- gators and their immediate families should not buy, sell, or hold stock or stock options in the company funding the research or companies making related or competing prod- ucts. In addition, other relation- ships with these companies should not occur during the study period, including participating in educa- tional symposia sponsored by the companies, or serving as paid con- sultants to" or on the board of the cotnpanies.1 Specifie rules about disclo- sure of relationships with for- profit companies should be imple- m.ented. Broad disclosure of financial relationships with companies has generally been regarded as a key fea- ture of codes to prevent and resolve conflicts of interest,' although its function in this regard is limited.10 Furthermore, some believe disclo- sure undermines scientific prin-

ciples of free exchange of informa- tion because it biases listeners and readers about the work" Nonethe- less, most observers believe that the relationship between the scientist and for-profit companies is infor- mation readers and others judging the research need to know,lsand that peer review alone may not be sen- sitive to alterations in study design or int:erpretation that may have been under many influences_t•'' I pro- pose that detailed rules about dis- closure be included in the Con- tract, detailing what, when, and how disclosure should occur_ Suchh dis- closure should include sources of fi- nancial support for research, speak- ing, consulting (including serving as expert witnesses), and personal eq- uity or debt that are related to bio- technology and health services (eg, stocks and stock options). It is rea- sonable to insist on disclosure to the Intermediary (which will notify com- panies funding relevant research). and to the medical center at which the scientist is based, in all publica- tions, when speaking tn educa- tional symposia, or research fo- rums, and when applying for other research funding, reviewing ar- ticles, and consulting. Disclosure should be in writing and, if the site is an oral presentation, ac the begin- ning of a lecture or presentation at a symposium. ADVANTAGES AND DRAWBACKS OF THE GUIDELINES lnsisting on contracts and an inter- mediary not only will help ensure the independence of each investigator butcan have advantages for the qual- ity of science, for patients, and for companies. Because research pro- posals will be judged based on their scientific merit, the present free-for- all in which company money for studies is actively solicited by inves- tigators and departments (and may, in fact, go for thebcst lobbying rather than the best science) would be im- proved. Patients would benefit be- cause more explicit disclosure to them.vouldbe possible of the exact terms of agreement between doc- ton and companies aboutstudies in which they particip.te.'= Finally, companies would benefit in rnany ways. Their legitimatedaires to col- labo rate w ith aaidernic sdentists will be satlsfied more clearly by selec- tion of researchers by peer re- view.1° Furthermore, companies, which are accustomed to compar- ing the costs of services for which they contract, would benefit from as- surance that proposed budgets had been reviewed and compared by the Intermediary in its study sections. D [SADVANIAGES Of the guidelines will undoubtedly be raised by many. Scientists who have not regarded their indepen- dence as threatened bydirect indus- trial support rtutyviewcontracts and an independent third party as para- doxically resttictive. Concerns about conflicts of interest raised by oth- ers?'2' however, and the stratifica- tion of accountabilities I have pre- sented support efforts such as these to protect investigatots indepen- dence. Furthermore, scientists who are accustomed to having their work peer-reviewed in NIH or founda- tion scientiflc studysectionswillwel- come the chance to compete for funds based on the quality of their science rather than less explicit cri- teria. Academic departments nwy re- gret that budgets will be limited to direct and indirect costs, or they may fear an overall decrease in corpo- rate money for academic research if compa nies believe thetr interests will be diluted. Like scientists, how- ever, departments too can be un- duly pressured by financial incen- tives in research and need safeguards against conflicts of interest. More- over, institutional efforts presently spent on soliciting company re- search funds might be used instead to shape the direction and priori- ARCH OERMATOLNOL 130, APR 1944 436 ties of publicly supported agencies. Itxt,ustsy may lament the added cost of funding the InterraedUry: a cea- tral agency to standardize the re- search funding prncess and restric- tion of study budgets to direct and indirect costs only, however, will probably offset this added cost. More important, companies may fear that their own interests will be lost in what they may see as another bu- reaucratic layer in the sdenti$c en- terprise: If a company's pritaarygods for funding a study involve market- ing and promotion as well as re- searchan.ddeveioptnetu, then the ar- rangement proposed in these guidelines may be unattractive. But for legitimate company desires for collabora,tion with academic sden- tists to capitalize on respective streagths and generate exemplary re- search, the guidelines more closely match standard procedures for fruit- ful corporate collaboration: con- tncts, eompetitivz "bidding" (or peer review), and justified costs. CONCLUSION The research enterprise hinges on trust,t' and codes of behavior for sci- entists risk critidsm for ignoring this trust. As with guidelines written for otherpurposcs, some will undoubt- edly find these overly restrictfve," or even scientifically unaccept- able.r' My hope is that we will gain a fresh perspective about what it means to be an independent inves- tigator; perhaps we will be more in- clined to apply existing models for ensuring fairness in corporate inter- actions and publicly funded re- search to academic research sup- ported by companies, Implementing these guide- lines will require a major restruc- turing of the way In which our pro- fession regards and interacts with industry. Regardless of whether or when more specific codes are insti- tuted by the profession or legisia- tors, two introspective points may help us in the meantime as indi- vidual investigators faced with the I so cc di Ph rai ge ca sci to- tia ca bil dc ing stii fer it i ce>t by ter PM 3hu am gai a lf tiol Pai mu wh 902 ess+ Acc 199 b3'! of A i}is: stith

f possibitity of having our research 1' funded by for-profit companies. We should realize that our per- sonal goals for the research and the company's goal for the research are different. For-profit companies and physiciyn-scientists often collabo- rate and achieve mutually advanta- geous goals, but this fortuitous out- came is serendipitous. While the scientist ultirnately may be striving tc+ answer an obscure seientific ques- tion andfor publish the results, a company always has the responsi- bility not to waste its revenue in en- deavors with little chance of mak ing money. That laudable results czn still be achieved despite these dif- ferences does not lessen the fact that it is prudent to keep them in mind. When faced with any eon- cetn that our sdentific freedom to hypothesize, e.xperitltent, and in- terpret is being affected by a com- pany supporting our reseuch, we should ask ourselves, "for w6otn am I working?" This self-interro- gation about accountability serves as a litmus test for worrisome situa- tions in which further input ofan im- partial third party may be useful. We must be sensitive to situations in which we feel accountable to the goals of a co mpany rather than more essential, important aims. Accepted for publication September 9, 2993. This work was supportcd, in part, by grants from the Nationai Instftute ofAnhritis, Muscutoskeietal, and Skin Diseases (P30AFt39750), National In- stttutes of Health, Bethesda, Md, and the US Public Health Service, and the Skin Diseases Research Center, Uni- versity Hospitats of Clevekurd (Ohio). Dr Chren is a recipient of a eareer development award fro>s+ the L.eaderx' Society of the Dcrm.atoiogy Founda- ttox, Evaruton, 111. I thank C. Seth l.andefeld, MD, and Franca j. Storrs, MD, for ideas and advice for this work. Reprint requests to 10701 East Bivd, Clevelarxd, OH 44106 (Dr Chrcn). 1. Enpkr Ri. Cwr.ii JW, Friedman PJ. Mteer PS, P.4rr: RAt. Yrsreprwsn~ sM r.epawib{I- try in med}ql rH.erch. N frnyf J MUd. 1N7; 317:1343-13M. Z Anwlrm F.dwtdon tor txinicti A.rrch Guida- linu for Avflfdtn4 Calflkt of hwwteet. CNn wes. 1990;3~35-240. 3. Abodstfon o1 Msrrfun fMdical Corepes_ SWdr una Mraurrp M fwWO, coancm aaxn- mRrnwnf xrd Cat}pc>Y d Mtxut in Aaaerth. Wathkpton, tlC: iM A»ock9on at Am.rfcin Medicat coli"; 1990. 4. BeeuCtMmp it, Cook RR. Fayurne!!w WE at at. Edkal ouid.ikus for epidetAbiapkts. JCIIn EAfdemhL 1991;N(suppl 1):151S-16QS. S. lludt 1H. IRdustry-tund.d dmnatolapie re- s.ecch vrRAfn .csd.mie in tisr Unlbd Sws. J lnnrt OrnrolsL 1s792 9E2ES 2EE. 6. BtumrRhi t), 61udr IA, Louit K8, BtolO 1AA, Mlke 0. Urdverdty-lr4utry (esurch nklfartehips in 61awodopy: YrpkrYonro fa iM unts.nity. Sd- ma.19S6=A61-19Q6. 7. alummihmt D. Andemic-kW~uttry ntsConrtapa in the Ith ac{enue: *aent sonfeCuM+cu, w)d muvpm.rd_ JAA/A 10922ES3344-Si49. E. Chnn tAM. lmd.hk! CS, MSurralr TN. Doctora, druQ compartlts. and pitK. JAAd4. 1999;2Ci2: 344s-34s1. 9. Ctinm MM. Induttry•tundod dtmutoiopic rw- sarcti witAia scademhln th. UMtad St>Zes: ru- cU utd tthial contldatffartt. J lmrtt Qumf- tol. 1sOZAaSS4. 10. CouncY on Sdandtk Athlrs end Council on Ethi- u! and Judiciti Athtra„ Artwrian Aledkal As- ARCH D@RMATOINOL 130, APR 1944 437 tecitKoa. CarrfYcts of SPYrW fa medfnl arr.uf iNustry rMrroh nt.tlawhtpe. JAMAI 1990,• 2iS2711tS-PT33. 11. t:uf1lcrr !J. NtN, Mq; tiN CRACA baocn. SaF- ayoe.1 Mt,M:1034-1 t>,6. 12, t1w+A.on ft Sartzo a funarp and aa*a,m of dkSial trfi. J®an Alp77J iMd.19ba;1:165- tsat. 13. OpertRrNlt of 1iWr1 end Hurtre Servicw. Food anr onq A4n"Strrefai. flraR po{ky tsrimerR on krdwtryaYPOaet+d seNrtfNe aad .diea- tia+e{ amvkf.a ftdry i#rp~ir.19QZ57:F6412- StN14. 14. tlordY L ENW uid p*aelorr! 4s"at b Vu ctr/qkg tuicWa.t epidanlaioqy. J cdh fR!- ewaat 1lv1;41(wppl l)31S-1sS. 15. liwy l, CimpW L. 9rry i31, st Ri. tkttStfd-af- kftMt 9ridellus ior a mhbwRu cRnlal tri+l of t>ubliem Ow ooronery utwy bypsaqnn sutgry. NfiiplJ A62 111l9;32o:W9•A61. 7t1. HYk- At, fis.nON1 JM. liwy MV, tlfoom ®S, spar H, Klaotien 1. SefrraRZ ,15. kmtd- 1np Ws hr the eorrie ard r.portln@ of eat- .it*Wv.Mft t..nreit :pansa.d hy ptum+i- c.utlnf Wm"r>ia. N 6pl J AMG. 1991=4: 13d2-156i. 17. KaWr Wl om promotlon ted tcl.r,etk a- chaaW: 1!N rob of ff1* c9nfcsl imr.etlartar. N Fitlf J MYd 1lfi1 a25:201-20.L 1a. A.lonIn AS. Qriktp aith conriicts ol irMrxt N 6qfJAfrd.19d4310:11dZ-Si 19. RGCr~n Mrl Piy~kluri' eonriie2e d 1nlrrt+! tlr NmNKlans ot ftdoww. M Ertp1 J M.d: 199G; 321:148i-1406. 20. RotMrrti IGt. ThN ktrKs of cNsrsfi s9onwt- ship. J C9r EpJGrmlot 1991;44(wppl 10):25S- 265. 21. Rottxtrn KL CaMkt d itMuist tfM Rew Mc- CutlqYrfl in sc.na. JAlW.19M;2bY27112- 27&4. 22. Shl+rrrr US, 99m RG. induetry r>rirtlburwmurt tor etrlwks4 04thah tnta clineYl trirh: sept wd ahicet luuK. Mn lMom AMd. 1901;115:14i- 151. 23. Staffyr P0. @t1lCRl NauN Inudvinp tanllicb d YAN1H14r vp1d%n110IoQiC N7YNi4gi1Df7: a rtport of thf oorrxNCw ar erAai QuidtNora of tlM So- ciev tor Eyi4mlofoplo ttntwrch. J cltn Eplcr miul. 11Q1;44(euppl 1)23S41S. 24. VurR1 AG, KuYkfl OS, Gufd+Yrxt for kiduttry- qsoshond t.rartifi at unitiencltla. Soncc 1986; 227:38E-38L 25. Wiauds Al. ww .dRaW poky on connids d IntnrnsL N EApJ J Ahd 1990;323:15d@-1589.