Tobacco Documents Online

, Communications Pu6lisher's note Unsolicited responses to articles that appear in JHPPL are wel- come and will be considered for publication on submission to the editor. Send items to Mark A. Peterson, Editor, 3G25 Forbes Quadrangle, University of Pittsburgh, PA 15260. JHPPL does not edit the communication that it publishes. To the editor: In a recent article, Barnes and Bero (Journal of Health Politics, Policy and Law, 21 [3]: 515-542) examined funding by the Center for Indoor Air Research (CIAR). The article appears to be little more than an exam- ple of bias in the search for bias. The ultimate and clearly expressed objective of the authors is to discredit researchers for receiving funding from tobacco-related interests and to silence all those who do not support the authors' own views. Their approach has public policy implications far beyond the Center for Indoor Air Research. Based upon my own experience with CIAR, I want to point out diffi- culties I have with the authors' underlying assumptions and analysis, and comment on the implications of the positions advocated by the authors. I believe I am qualified to do so because I have served on the CIAR Board of Directors since the CIAR was established. At present, I am chairman of the board. Although I believe that my comments reflect the feelings of all the board members and CIAR staff, I speak only for myself in this response. I am a scientist with R. J. Reynolds Tobacco Company. My supervisor (also a scientist) would be amused to hear that I am, as the authors repeatedly charge, a "tobacco industry executive:' Indeed, current and former members of CIAR's board of directors hold Ph.D. degrees in This article is for individual use only and may not be further reproduced or stored electronically without wntten permission from the copyright holder. Unsutbnrized reproduction may result in financial and otherpenalities. (c) DUKE UNIV PRESS

1280 Journal of Health Politics, Policy and Law several scientific disciplines, including chemistry, pharmacology, and toxicology. CIAR Funding Method In the 1980s, government agencies and private organizations recognized that the science relating to indoor air quality-including the environ- mental tobacco smoke (ETS) component of indoor air-was lacking and the data available often resulted from poorly considered protocols. Rec- ognizing the importance of the issue and the lack of serious funding in the science, CIAR was created to provide funding for fundamental and applied studies. As noted by the authors, CIAR remains the only private, nationally based organization funding such research. Barnes and Bero segregate research funded by CIAR into two cate- gories: (1) "peer-reviewed funding," in which CIAR's Science Advisory Board (SAB) makes recommendations, and (2) "special-reviewed fund- ing:" The designations "peer-reviewed" and "special-reviewed" relate to the authors' view of CIAR's research funding process,-not-to the standard of review given reports resulting from the research. The types of funding have not been secret, and notwithstanding Barnes and Bero's assertion, the existence of applied studies separate from those recommended by the SAB has been noted publicly-among other places, in CIAR's newslet- ter, CIAR Currents. Every project submitted to CIAR for research funding undergoes peer review. A list of peer reviewers, used for both SAB-recommended and applied study proposals (termed special-reviewed by the authors), is set out in CIAR's annual Request for Applications (RFA). Both SAB-recommended fundamental studies and applied studies are funded only after approval by the CIAR board of directors. As mentioned above, the board consists of Ph.D: level scientists from a number of sci- entific disciplines, so that the board is well qualified to make funding decisions. The only limiting factor in my vote as a member of the board, other than consideration of the merits of individual proposals, has been the amount of total funding available and ongoing commitments to fund previously approved multiyear research projects. The principal differ- ence between applied and SAB-recommended projects is that the latter, as the name implies, is recommended by CIAR's Science Advisory Board from among proposals submitted pursuant to the RFA. Many applied studies are of near-term interest and are applications oriented or

Communications 1281 intended to fill gaps in the science rather than being basic scientific research. CIAR enters into contracts with all researchers. Researchers are assured independence in pursuing their research proposals and are encouraged by contract to publish their findings fully and completely with only one stipulation-that CIAR funding be acknowledged in any publication. The Article's Assumptions and Bias The principal problem with the Barnes and Bero article is its approach to judging "quality:" Instead of judging quality on an article-by-article basis, Barnes and Bero believe they can judge quality by employing cer- tain "surrogates" for quality. This determination of "quality" turns upon several assumptions, assumptions supported for the most part by citation to other works (including unpublished findings) by the authors. Some surrogates, such as sources of funding acknowledged, appear to beg the question. One "surrogate" for quality used by Barnes and Bero is whether an article resulting from CIAR funding was classified (using the authors' definition) as "pro-industry." The definition of "pro-industry" employed by Barnes and Bero is itself suspect. The authors classify an article which finds "that the evidence [regarding alleged adverse health effects of ETS] is inconclusive" as "pro-industry" rather than neutral. This is little dif- ferent from an earlier version of the Barnes and Bero article in which the authors considered an article "pro-industry" if it discussed exposure to ETS "without mentioning the documented hazards of exposure at any point:'t Irrespective of the definition employed, the "pro-industry" criterion is curious. A study, according to the authors, is more likely of low quality whenever its findings could be classified as favorable to the funding source. This position is absurd. Research should go where the science leads. If it leads in a particular direction, then so be it. The validity and quality of research should stand or fall on its scientific merit and that is 1. The "earlier version" of the article refers to the manuscript submitted by the authors' coworker, Stanton Glantz, as part of his posthearing comments to the Occupational Safety and Health Administration. On the face of the OSHA submission, the authors state that a "version of this manuscript is to be published in the Journal of Health Politics, Policy and Law:' Both manuscript and article state: "Presented in part at the American Public Health Association meeting in Washington, DC" on 2 October 1994.

1282 Journal of Health Politics, Policy and Law how the results should be judged, in the public policy context as well as in the scientific context. Any evaluation of scientific research should focus on the methods and data, not on some "self-styled" surrogates of quality or the authors' perception that the outcome is "favorable to the sponsor." Included in the authors' references was a commentary by Kenneth J. Rothman entitled "Conflict of Interest: The New McCarthyism in Sci- ence:" I quote an excellent rebuttal from Rothman's article to Barnes and Bero's usage of funding source as an indicator of quality: Judging someone's work by the funding source, or by any other char- acteristic other than the content, raises an ethical problem. The ethical problem is similar in principle to the discredited practice of judging college applicants by their photographs.... Since there are no official boundaries on what could be the reason for a conflict of interest, whenever we stray from using anything but the substance of a work itself as the basis for judgment, we begin to substitute prejudice for reason; we abridge the rights of others and convert the free inter- inter- ..~ ..---___---- --. _ _ -h ab.._ _. matc- ..out pedigrees. (Roth- change of critical views irito a shouting- man 1993: 2784) Article Findings :i Although the limitations on a letter response do not permit a point-by- point review of the article's analysis, I believe a few observations regard- ing the authors' findings are appropriate. For example, the article pre- sents comparative data in three tables. Table 1 sets out topics of projects funded by CIAR. The text accompanying the table concludes that a greater percentage of applied than fundamental research projects inves- tigated ETS. The authors opine that a study measuring exposure to ETS would be much less damaging to the industry from a legal perspective than one showing that ETS causes disease, "thus it is 'safer' for the industry to sponsor exposure research:'2 If that is the case, CIAR is not being "safe;" because one-third (eight of twenty-four) of the studies related to ETS (SAB-recommended and applied) are classified by the authors as investigating "ETS Health Effects:' I am not qualified to speak on legal matters. Neither do the authors 2. In this response, I address the authors' data on their own terms. I have not attempted to determine whether the individual judgments that the authors have made using their classifica- tion scheme are valid.

Communications 1283 present their own legal credentials. However, a special committee of the National Academy of Sciences convened in 1983 (NAS 1983) considered exposure assessment to be one of four critical elements (hazard identifi- cation, dose-response determination, exposure assessment, and risk char- acterization) in conducting a systematic risk assessment to quantify the potential hazard for exposure to an environmental hazard such as ETS. Furthermore, determining a dose-response without exposure assessment data requires a host of unsubstantiated assumptions. By any objective standard, CIAR is providing critical information on exposure for an ETS risk assessment by funding fourteen of twenty-four ETS-related projects listed by the authors in Table 1. Table 2 sets out affiliations of principal investigators funded by CIAR. The authors give a greater percentage of academic affiliations for those receiving "peer-reviewed" than "special-reviewed" funding, although the latter exceeds 50 percent. I have not reevaluated CIAR's funding deci- sions, but I suspect the difference simply reflects the expertise of the principal investigator. Proposals submitted pursuant to the RFA (and thus reviewed by the SAB) are oriented toward basic research; a province of academia. On the other hand, those qualified to undertake applied stud- ies, which are applications- or fieldwork-oriented, are as likely to be engineering-type companies as those with academic affiliations. Scien- tific competence and integrity are not the sole possession of academia. Table 3 examines "quality and outcomes of publications" resulting from CIAR funding. As noted above, "quality" is examined through the use of questionable "surrogates:" The authors first examined whether arti- cles appeared in peer-reviewed publications, concluding that "[a]lmost one-half of articles resulting from 'special-reviewed' projects were pub- lished in non-peer-reviewed journals:' This compared to about one-fifth of the "SAB-reviewed" projects. While the authors' statistics may be correct about the amount of peer- reviewed publications, it says nothing about the quality of the papers. An example of this may be taken from the ETS literature. In 1985, scientists from R. J. Reynolds Tobacco Company (Eudy et al. 1985) presented results in a conference paper and later published similar results in con- ference proceedings (Eudy et al. 1986), establishing that nicotine in ETS resides predominantly in the aerosol vapor phase as opposed to being in the particulate phase, as found in mainstream smoke. Later, Neurath et al. (1991) and Esquier and Hee (1990) published conflicting data in peer- reviewed journals. By the authors' criteria, the peer-reviewed journal papers should be of better "quality." However, Ogden et al. (1993) repeated

1284 Journal of Health Politics, Policy and Law the peer-reviewed, published work of Neurath et al. and Esquier and Hee and showed that their experiments were in error. The original finding about the phase distribution of ETS nicotine that was given at a confer- ence and also published in conference proceedings was confirmed. The above example not only exposes the specious relationship between peer-reviewed publication and "quality" but also demonstrates that the venue where research results are reported has little bearing on the mer- its of the work. That is particularly so when the research in question is designed simply to provide missing data in the context of established methodology (i.e., applied research), rather than constituting an advance in basic scientific knowledge (i.e., fundamental research). The whole question of publication in peer-reviewed journals has become extremely muddled recently due to the written and unwritten policies of journals associated with public policy-oriented organizations such as the American Thoracic Society and the American Medical Asso- ciation. The Journal of the American Thoracic Society will no longer accept articles based on research funded by tobacco interests. While the Journal.oflheAmerican MedicalAssociation.has announced no-such for- mal policy, I can assure you that unwritten roadblocks make publication of "politically incorrect" science extremely difficult. Again in Table 3, the authors also found a difference between the two project groups with respect to whether the resulting articles were "pro- industry:" Even including their incredible criterion that an inconclusive result is pro-industry, the authors' own data belie their conclusion that "articles based on CIAR's special-reviewed projects were more likely than articles from its peer-reviewed projects to support the tobacco indus- try position:' The authors, focusing on "pro-industry" articles, look at only one side of the coin. Table 3 shows that articles resulting from "special-reviewed" funding were 50 percent more likely to be anti- industry than articles from "peer-reviewed" funding. I have other concerns with the article. For instance, Barnes and Bero question the propriety of members of the SAB receiving funding from CIAR. In creating the SAB, every effort was made to find the leading experts in the field of indoor air quality. The members of the SAB meet the high standards adopted by CIAR in its selection process. The com- ments of the authors regarding the SAB focus on a potential for conflict, but ignore the practical consideration-that automatically eliminating SAB members from consideration for funding would also eliminate some of the best researchers in the field. Furthermore, as the authors note, research resulting from SAB-recommended proposals is of high quality.

Communications 1285 In reviewing statements made at Occupational Safety and Health Administration (OSHA) hearings, the authors criticize a study of per- sonal ETS exposure of subjects in sixteen cities; the principal investiga- tors in this study were scientists at Oak Ridge National Laboratory (ORNL). The authors attempt to belittle the results, in part by criticizing the recruitment of subjects by an "interested" firm. This criticism regard- ing recruitment is simply wrong. All study participants in each of the six- teen cities were selected by local companies who had no knowledge whatsoever of the funding agency or the research objective of the study. The fact remains that among the data OSHA had to review, the ORNL study was the largest, most relevant, and most representative, and was accomplished using the most appropriate analytical methodology. The authors also criticize the use of laboratories at R. J. Reynolds. The investigators for ORNL were very interested in the project, but could not establish some of the necessary analytical methods in their laboratory (e.g., nicotine, solanesol) in the required time frame. Since R. J. Reynolds scientists had published widely on methods for measuring ETS exposure and had the-equipment in-place;--a collaboration between_ORNL.-and.- Reynolds scientists was proposed and accepted. After the protocol for this joint project was peer-reviewed and accepted by the scientists at ORNL and by CIAR's board of directors, the study was executed. CIAR welcomes additional studies of personal exposure to confirm or challenge any findings by ORNL. In fact, if OSHA had followed the advice of its consultants and conducted a personal exposure study in the first place, the CIAR-funded study may have been unnecessary. Implications of the Authors' Approach The bottom-line objective of the authors appears just before the conclu- sion: "Our findings support calls for investigators to refuse all tobacco industry financing.... All researchers associated with CIAR, even those who do not study tobacco-related issues, are contributing to the tobacco industry's agenda:' In their zeal to attack the tobacco industry, Barnes and Bero display their own bias, not only against tobacco interests, but against private industry in general. The call to "boycott" funding from CIAR is really directed toward all researchers receiving funds from private industry sources. This approach is based on logical absurdities. For example, many researchers entering into contracted funding with CIAR also receive money from government agencies. Indeed, one "special-reviewed"-funded research organization

1286 Journal of Health Politics, Policy and Law criticized by the authors is the Oak Ridge National Laboratory, a Depart- ment of Energy Research Laboratory. Are the authors asserting that ORNL biases its findings just for research funded by CIAR or for each funding source? Are researchers who receive funding from government sources biased toward reaching "politically correct" findings when con- ducting government-financed studies, merely because of the funding source? Or are they selectively biased only when conducting research funded by CIAR? The authors' conclusions appear to be that researchers are selectively biased only when receiving funding from CIAR or from private industry sources in general. Perhaps the greatest irony is that the Barnes and Bero article would not withstand their own scrutiny. The article is a result of funding pur- suant to the Tobacco-Related Disease Research Program funded by a sur- tax on the sale of tobacco products in Califoiaiia. Under the Barnes and Bero approach, it is not surprising that their conclusions are unfavorable to CIAR. Even so, I assert that each article resulting from such funding be judged on its own merits, not by surrogates of quality or its outcome. . The-current article_is-not-the-authors'.-first-foray regarding research relating to the tobacco industry. An article in Science addressed the authors' approach in a related context: To Bero, the only bias seemed to be one in favor of negative results in the symposia [related to tobacco]. But Marcia Angell, executive direc- tor of the New England Journal of Medicine, argued that Bero's data on quality may not fully explain the scarcity of negative results in the mainstream journals. Researchers have to be careful, said Angell, not to let "political correctness" drive their conclusions. "I hold no brief with the tobacco industry;" Angell said later, "on the other hand, sci- ence is science:" (Taubes 1993) Do not mistake my objection to the authors' call for a boycott of CIAR research funding, use of "surrogates" of quality, and their own bias as a statement that the source of research funding is wholly irrelevant. The fundamental considerations, however, should not be from whence the funding came, but the quality of the research itself and whether condi- tions were imposed by the funding entity with respect to the research and resulting reports. Furthermore, science is self-correcting. Those ques- tioning a scientific study always have at least two options: (1) to demon- strate how the study is technically flawed or (2) to conduct equivalent studies to generate countervailing (or confirming) data. The authors

Communications 1287 ignore these considerations and instead apply a pseudoscientific politi- cal analysis to works of science. In so doing, the authors add nothing to "science" and do a disservice to those interested in honest criticism of scientific endeavors. The criteria and analysis applied by the authors to CIAR may as eas- ily, and as wrongly, be applied to other industries, including the chemi- cal and pharmaceutical industries. Indeed, the authors' approach would apply with equal force to research funded by the government, particu- larly in a regulatory context, as well as to research funded by antismok- ing organizations. Research funded by CIAR is no different from research funded by other organizations receiving money from private industry, save perhaps for the fact that, as the authors note, CIAR is a nonprofit corporation that funds research through contracts and acts as an "intermediary" between investigators and the industry. "Science is science" and should be judged accordingly, not by unsub- stantiated "surrogates" of quality and irrelevant factors which lead to arbitrary conclusions. Judging science as science provides policy mak- ers with the research needed to make--informed--decisions.- A.nx._other approach elevates "political correctness" above scientific substance. Charles R. Green, R. J. Reynolds Tobacco Company References Barnes, D. E., and L. Bero. 1996. Industry-Funded Research and Conflict of Interest: An Analysis of Research Sponsored by the Tobacco Industry through the Center for Indoor Air Research. Journal of Health Politics, Policy and Law 21(3):515-542. Esquier, F., and J. HBe. 1990. Critical Study of Methods for Nicotine Measurements in Environmental Tobacco Smoke. Bulletin de l'Association pour la Recherche sur les Nicotianges pp. 19-36. Eudy, L. W., F. A. Thome, D. L. Heavner, C. R. Green, and B. J. Ingebrethsen. 1985. Studies on the Vapor-Particulate Phase Distribution of Environmental Nicotine by Selected Trapping and Detection Methods. Paper no. 38 presented at the 39th Tobacco Chemists' Research Conference, Montreal, Quebec, Canada, 2-5 October. Eudy, L. W., F. A. Thome, D. L. Heavner, C. R. Green, and B. J. Ingebrethsen. 1986. Studies on the Vapor-Particulate Phase Distribution of Environmental Nicotine by Selective Trapping and Detection Methods. In Proceedings of the 79th Annual Meeting of the Air Pollution Control Association. Pittsburgh: Air Pollution Control Association.

1288 Journal of Health Politics, Policy and Law National Academy of Sciences (NAS). 1983. RlskAssessment in the Federal Gov- ernment: Managing the Process. Washington, DC: National Academy Press. Neurath, G. B., S. Petersen, M. Diinger, D. Orth, and F. G. Pein.1991. Gas-Particulate Phase Distribution and Decay Rates of Constituents in Aging Environmental Tobacco Smoke. Environmental Technology 12:581-590. Ogden, M. W., K. C. Maiolo, P. R. Nelson, D. L. Heavner, and C. R. Green. 1993. Artifacts in Determining the Vapor-Particulate Phase Distribution of Environmen- tal Tobacco Smoke Nicotine. Environmental Technology 14:779-785. Rothman, K. J. 1993. Conflict of Interest: The New McCarthyism in Science. Jour- nal of the American Medical Association 269(21):2782-2784. Taubes, G. 1993. Peer Review Goes Under the Microscope. Science 25:252. Response: We would like the readers of the journal to know that the tobacco indus- try has a long history of publishing highly critical letters to the editor in response to published research and that it then cites those letters in other settings in an attempt to discredit the research (Rennie 1993; Davis and Chapman 1994; Rennie 1994). For_example, our previous research has shown that the tobacco industry often cites letters to the editor in policy settings to suggest that a particular research article is biased, flawed, or contains controversial findings (Bero and Glantz 1993). The letters that are cited are rarely accompanied by citations of the source article or by responses from the authors of the source article. It is therefore likely that Dr. Green's letter, regardless of the validity of his criticisms, will be used by the tobacco industry in an attempt to refute our findings. In his letter, Dr. Green makes several attacks on us personally, in addition to criticiz- ing the quality of our research. For example, he repeatedly refers to us as "biased:' states that our approach is "based on logical absurdities;" and criticizes us for not presenting our "legal credentials:" We feel that these sorts of personal comments are unprofessional and unwarranted, and, therefore, we limit our reply to the substantive criticisms raised in Dr. Green's letter. Objective of Our Article Dr. Green misrepresents the purpose of our study. In his letter, he claims that "the ultimate and clearly expressed objective of the authors is to dis- credit researchers receiving funding from tobacco-related interests and to silence all those who do not support the authors' own views:' In fact, our clearly expressed objective (as stated on page 518 of our article) was