| |||||||||||||||||||||||||||||||
Our sometimes misplaced sense of optimism is at its core an expression of confidence in science and the scientific method. We believe that honest intellectual inquiry can lead us to theories or laws that, after testing and prodding, will yield truths that will light our path to the future.
For good reason, the public, through its elected representatives, has been willing to devote more and more resources to science and research. Despite the best efforts of the Reagan administration, public spending for research -- and I grant that much of it was for defense -- continued to rise in real and relative terms.
The foundation of public support for science, or for any public endeavor, is trust -- in this case, trust that scientists and research institutions are engaged in the dispassionate search for truth. We are willing to spend great sums in the service of a higher value., and no value is believed to be more dear to a scientist than the truth.
Our support for science reflects our fascination with it. We avidly devour the news of the latest discoveries. Full sections of The New York Times and other newspapers are devoted to scientific coverage. The recent revelations about the big-bang theory of the cosmos led the network news broadcasts. When 60 Minutes airs a story on a study detailing the health benefits of drinking red wine, shelves are emptied of wine and advertising campaigns are launched. Sitting in traffic listening to the radio, you are likely to hear a news story on the latest cholesterol study published in the New England Journal of Medicine.
Increasingly, in the modern era, cutting-edge scientific discoveries move rapidly from the laboratory to the legislature and the regulatory agencies. Scientists, hospitals, and institutes convene press conferences to publicize their findings and satisfy a public both eager for scientific advance and troubled by every new report of a threat to health and the environment. Scientific findings largely determine the agendas of government agencies with thousands of employees -- the Environmental Protection Agency, the Food and Drug Administration, and the Occupational Safety and Health Administration, to name but a few.
In this climate, the publication of a paper claiming that a food additive causes cancer or that a building material retards the mental development of infants can trigger a massive public outcry and demands for government action. Americans are not a patient people. The public is rarely content to wait 5, 10, or 15 years for further research and study.
The consequences can be serious. A report that an additive causes cancer may drive a manufacturer out of business, throwing hundreds or thousands out of work. A report about the toxicity of paint or insulation can cause homeowners to spend thousands on testing kits (some of them fraudulent) and on the removal and replacement of the offending material. Housing units may even be closed, leaving low-income people homeless. School districts will strip carpet, paint, or insulation from buildings, leaving themselves unable to afford new textbooks.
Some of these regulatory actions will ultimately prove unwarranted; that is an inherent and understandable risk of relying on cutting-edge research. But there is a difference between honest error and various forms of misconduct, such as plagiarism and fraud. Blatant forms of misconduct attracted the interest of the Subcommittee on Oversight and Investigations of the U.S. House of Representatives, and the subcommittee's interest has in turn caused the scientific community to pay attention to the issue of misconduct.
What, exactly, is misconduct? Legitimate questions have been raised about precisely how misconduct should be defined, but the subcommittee has never sought to explore the outer reaches of this concept. We leave those efforts to others. Rather, we have looked only at clear-cut cases involving fabrication, falsification, or plagiarism. We follow these cases carefully to see how the system deals with them, how speedily complaints are addressed, and how fairly and accurately issues are resolved. Although many cases are referred to our subcommittee, once preliminary investigation establishes that only differences of opinion or mistakes arising from honest error are at stake, we leave the matter to the scientific community. The subcommittee does not assess the substance of scientific theories themselves, concentrating instead on such factual matters as whether data really exist and whether allegations have been seriously and properly handled.
Subcommittee investigators have pursued cases in which published papers claimed to present data that, in fact, never existed or described the results of experiments that, in reality, had never been performed. They have pursued cases in which researchers abused the peer-review process to obtain advance copies of papers and appropriated their insights or accomplishments. They have followed cases in which evidence of misconduct was covered up or whitewashed by institutions apparently more interested in the appearance of integrity than in the reality of it.
Unfortunately, the nature and the purpose of the subcommittee's inquiries have been misrepresented, particularly by persons who have everything to lose from the truth. Some have said that we are putting science itself on trial. That is false. Others have said that we seek to punish people for honest errors. That is false. It has been suggested that we do not understand that mistakes are inevitable in scientific inquiry. False again.
Science is the search for truth at the boundaries of what is known about the natural world. It is a difficult and perplexing task, one in which we make progress by finding and correcting errors. In the uncharted waters of discovery, mistakes are bound to happen. To condemn scientific error would be tantamount to stopping scientific progress.
There is a corollary to the inevitability of errors, however, and it is that when errors are recognized, scientists have a duty to make them known. A "no fault" process of recognizing and correcting error is essential to progress. But in some of the cases examined by the subcommittee, this proposition was turned on its head: it was said that since errors are inevitable, there is no need to acknowledge or correct them.
For a time, the Subcommittee on Oversight and Investigations was excoriated as a band of philistines incapable of understanding science and as a troop of inquisitors engaged in a senseless witch hunt. Indeed, some of the more intemperate accounts portrayed the very concept of scientific independence as under siege by the subcommittee's efforts to uncover the truth.
As a result of this campaign, at times orchestrated for profit by public-relations firms, the subcommittee was deluged with well-meaning but misinformed letters urging us to leave scientists and universities alone. (Many advised that we go after corrupt defense contractors, a suggestion more amusing to us than to General Dynamics, Northrop, or TRW, all of which were the subjects of extensive subcommittee investigations. In fact, those are among the investigations for which we are best known.) Despite this barrage of criticism, the subcommittee has persisted, because what is at stake is too important to abandon.
Why are the inquiries by our subcommittee so important? I have already mentioned the policy implications, but there is another, equally important reason. We have an obligation to the American taxpayer. As chairman of an oversight subcommittee with responsibility for the National Institutes of Health (NIH), I have a duty to see that taxpayer dollars appropriated for scientific research are actually used for research. Currently, taxpayers spend about $9 billion per year on the NIH. Most biomedical research performed in this country is federally funded.
Every time a researcher takes taxpayer money and publishes fabricated, falsified, or plagiarized findings, the taxpayer has in effect been swindled. Furthermore, given our budget deficit, there is never enough money to go around. Each dollar wasted on a dishonest researcher is a dollar that might have gone to another, more worthy candidate who might have made a real contribution. In short, there is an opportunity cost for each grant that is abused.
Moreover, the nature of scientists and the scientific method is to build on the interesting results obtained by others. Every paper published with fabricated or falsified data will spur other scientists using still other federal grants to try to replicate or extend the results, wasting even more money and time.
How widespread, then, is the problem of scientific misconduct? The only honest answer is that we do not know. One indicator is a study conducted by a small research group called the Acadia Institute1,2. This study found that approximately 40 percent of the deans of the nation's major graduate schools knew of confirmed cases of scientific misconduct occurring in their own institutions within the previous five years.
Another indicator is the recent survey sponsored by the American Association for the Advancement of Science3. It found that 27 percent of a group of scientists surveyed said that they had personally encountered, during the previous 10 years, research that they suspected was falsified, fabricated, or plagiarized. Those 27 percent reported, on average, witnessing at least two such incidents. Furthermore, close to half of the respondents said that the incidence of fraud was on the rise, whereas only 2 percent thought it was declining. More than half characterized university investigations of misconduct as lax. Although I would not characterize these findings as scientific, it is difficult to call them inconsequential.
Equally notable, however, were the scientists' responses when asked what they had done about the misconduct they had recognized. The vast majority had done little or nothing. The few who took action generally confined themselves to discussing the incident privately with a few people. Only 2 percent brought the matter to the attention of the public. Ironically, although large numbers had seen misconduct, although virtually none who had seen it had acted, and although the majority viewed university investigations as ineffectual, nearly all claimed that scientists should monitor themselves and that outsiders should not become involved.
The subcommittee also found disturbing attitudes when we contacted some 20 leading scientists to solicit their views on misconduct. In private interviews, almost all these scientists cited examples of misconduct they had witnessed, whistle-blowers they had seen harassed, or other matters engendering concern. Yet none were willing to testify, write open letters, or even have their names used publicly, for fear of retaliation.
Whether misconduct is widespread or not, let us examine the argument of those who say that misconduct need be taken seriously only if it is widespread. The suggestion that a single bank robbery need not be investigated unless and until it can be proved that that particular form of bank robbery is rampant would be correctly regarded as ludicrous. Believe it or not, I have found that the vast majority of members of Congress are honest. But does that mean that a particular member of Congress should not be investigated if there are allegations of wrongdoing? Of course not. Very simply, every professional -- lawyer, journalist, mechanical engineer, member of Congress -- has a clear ethical obligation to handle each and every allegation of misconduct fairly. My concern is that in the field of science, that has until recently not happened.
I am sorry to say that the case studies pursued by the subcommittee have provided still more evidence of the disturbing reality. My descriptions here are of necessity quite condensed, but I urge all interested parties to obtain the subcommittee's hearing reports, which run approximately 1000 pages, for more detail.
One of the first cases to capture the subcommittee's attention was that of Dr. John Darsee of Harvard Medical School4,5. Dr. Darsee's federally supported research in cardiology and his long list of publications marked him as a rising star until it was discovered almost by chance that he had committed an act of fraud. Two investigations were conducted at the medical school -- the first by Dr. Darsee's immediate supervisor and department chairman, and the second by a committee of faculty members from Harvard and elsewhere that the dean had appointed. Both reported no misconduct in Darsee's published research. A third committee, appointed by the NIH, discovered a massive fraud: the data for a number of Dr. Darsee's published experiments did not exist. This was particularly awkward for Harvard, which had maintained up to that point that its investigators had reviewed the data fully -- the same data that later turned out not to exist.
Another major case involved Dr. Stephen E. Breuning, a professor at the University of Pittsburgh who had risen to national prominence as an expert on drug treatment for the mentally retarded5,6,7. Dr. Breuning's research purportedly showed that the condition of severely retarded children improved markedly when they were taken off certain tranquilizers. From 1979 to 1983, Dr. Breuning reportedly published a third of all the articles in his field. Some states changed their treatment protocols in response to his purported findings.
In 1983, however, in a letter to the National Institute of Mental Health, Dr. Breuning's mentor, Dr. Robert Sprague, raised serious questions about the integrity of Dr. Breuning's research. The University of Pittsburgh investigated and reported to the government that it found no problem. The National Institute of Mental Health sat on the matter for years, and the whistle-blower, Dr. Sprague, suffered apparent reprisals. During this three-year period, Dr. Breuning traveled the country propagating his theories further to an unknowing public. It turned out that, among other things, Dr. Breuning's data had in large part never existed and that subjects had not been tested. Dr. Sprague's allegations were eventually substantiated in almost every detail. Dr. Breuning eventually pled guilty to two felonies and served time in a halfway house. Yet even after Dr. Breuning was publicly unmasked, several scientific journals actually fought the efforts of coauthors to retract articles on which he had worked.
In a more recent case involving faculty members from Tufts University and the Massachusetts Institute of Technology (MIT), a scientist named Dr. Margot O'Toole was vilified and effectively driven from her profession after she revealed that a paper in Cell coauthored by, among others, her supervisor, Dr. Thereza Imanishi-Kari, and Nobel laureate Dr. David Baltimore relied in large part on data that were falsified8,9,10,11. Senior scientists at both MIT and Tufts University contended that Dr. O'Toole's allegations were unfounded and reported that the paper was virtually error-free. These senior scientists informed Dr. O'Toole that the article could not be corrected because any correction would damage Dr. Imanishi-Kari's career5 and because the scientific literature was so full of error anyway that one more error would not matter. Dr. Baltimore himself said, in a letter to Dr. Herman Eisen, that he did not favor the retraction of the paper because of possible adverse consequences for another scientist who had been listed as an author, even though he, Dr. Baltimore, admitted that the paper was grossly defective and that he would never fully trust Dr. Imanishi-Kari's work again. Even after forensic analysis by the Secret Service had established that the data Dr. Imanishi-Kari did have were largely a recent fabrication, she and her coauthors continued to portray themselves as innocent victims.
At the end of a lengthy investigation by the Office of Scientific Integrity, Dr. Baltimore retracted the Cell article. He later resigned from the presidency of Rockefeller University. Still later, he announced his intention to retract the retraction. Dr. Imanishi-Kari came under prolonged investigation by a grand jury12,13,14.
In another case, involving the Cleveland Clinic, a researcher, Dr. Rameshwar K. Sharma, was ultimately found to have falsified his grant application, but was initially exonerated by the clinic's own investigation on the strange theory that all he had done was engage in "anticipatory writing"15. In anticipatory writing, a researcher claims to have already performed certain experiments and obtained certain data, and then justifies the false claims on the grounds that he or she "anticipated" being able to achieve the results at some point in the future.
It later emerged that the Cleveland Clinic's ready exoneration of the anticipatory writer had been based on a two- or three-hour inquiry in which no data were reviewed and no witnesses were interviewed. The person who handled this investigative effort, Dr. Bernadine Healy, now directs the NIH. Among her early actions in that post were efforts to undermine the Office of Scientific Integrity and compromise its independence -- efforts that included inappropriate verbal attacks on a critical draft report and on the office itself, a demand that a second important draft report be rewritten, an unjustified attempt to initiate an investigation of one of the office's chief investigators, and an instruction that the office reverse its decision to remove a scientific-panel member who had failed to inform the office of a possibly serious conflict of interest.
Finally, there is the matter of Dr. Robert Gallo, the NIH's world-famous AIDS researcher. Because the subcommittee has not yet completed its investigation, I cannot tell you what our conclusion will be. I can say, however, that one of the things that puzzles and troubles us is Dr. Gallo's entanglement in a large number of unusual circumstances that he claims are misunderstandings and coincidences.
After Dr. Zaki Salahuddin, one of his longtime laboratory scientists, was convicted of a felony in connection with his activities at Dr. Gallo's laboratory, Dr. Gallo explained that he had been unaware of Dr. Salahuddin's activities. In short order, Dr. Prem Sarin, Dr. Gallo's deputy laboratory chief, was indicted for activities unrelated to those of Dr. Salahuddin but also stemming from work at the laboratory16,17. Dr. Gallo explained that he knew nothing of his deputy laboratory chief's misconduct and that these two separate criminal cases involving his laboratory scientists were unfortunate coincidences.
Then we learned that two subjects described in an article in the Lancet coauthored by Dr. Gallo and French researcher Dr. Daniel Zagury had died, but that Dr. Gallo had failed to report these deaths to the NIH as was required by grant regulations and had erroneously reported in the Lancet that he had observed no adverse reactions in the human subjects18,19. Dr. Gallo again had an explanation. He explained that the statement in the Lancet was an inadvertent error and that his failure to comply with NIH procedure was a result of unfamiliarity with the regulations -- this despite some 20 years of employment at the NIH.
More recently, in the controversy over the AIDS blood test, Dr. Gallo is under investigation because of, among other things, allegations that statements he made in the patent application and thereafter in the patent dispute were deliberately misleading. Dr. Gallo first stated that the virus he used was definitely different from that used by the competing French team. When genetic sequencing proved that the viruses were identical, he suggested that the French must have taken his virus. When that claim was challenged, Dr. Gallo explained that there must have been an inadvertent contamination in his laboratory. Meanwhile, there were also questions about the cell line in which Dr. Gallo grew his viruses. Initially, Dr. Gallo seemed to suggest that the cell line was his own development. It eventually emerged that the cell line belonged to Dr. Adi Gazdar, a researcher at another NIH institute. Dr. Gallo explained that this was a misunderstanding and that he had never intended to deprive Dr. Gazdar of credit, but merely renamed the cell line for convenience.
This is a deeply troubling case. An eminent scientist who heads one of the most important laboratories in the world is embroiled, directly or indirectly, in many different serious situations, which Dr. Gallo himself invariably characterizes as inadvertent errors, miscommunications, and unfortunate coincidences., and the NIH's now defunct Office of Scientific Integrity found Dr. Gallo innocent of fraud or serious misconduct, a finding about which I had serious questions. We understand that the NIH itself receives millions of dollars each year in royalties from the blood test, but that should be no impediment to its assessing the charges against Dr. Gallo objectively.
Although the cases the subcommittee has examined differ considerably in their particulars, common threads run through them. First, in most cases, the dishonest scientist engaged in misconduct for years. Those in a position to know ignored the warning signs or turned a blind eye. Second, in nearly every case, the whistle-blower was treated badly, often far worse than the offender, for having put truth ahead of personal convenience or career advantage. Third, in virtually every case we examined, the internal procedures that are supposed to allow universities and other research institutions to police themselves failed. At best, internal investigations appeared careless and inept; at worst, they appeared to be conducted in bad faith. Fourth, the offices within the NIH that are responsible for monitoring issues of scientific integrity showed a marked disinclination to perform their job, a reluctance to pursue scientific-misconduct cases aggressively, a dilatoriness in developing evidence, and an overall lack of dedication to the goal of confronting scientific misconduct. Fifth, many prominent scientific-misconduct cases are accompanied by a barrage of erroneous and misleading reports in the media in which the long-suffering whistle-blower is labeled as vindictive and irresponsible, those investigating the allegations are lambasted as ignorant troublemakers, and blatant acts of dishonesty -- such as the wholesale theft of another researcher's article or the fabrication of data -- are trivialized as minor mistakes, misrepresented as differences of interpretation, or dismissed as mere communication problems.
What, then, can be done about misconduct in science? Every case of misconduct is difficult for those accused, for those making the allegations, for the institutions involved, for the funding agencies, and for the profession. To their credit, a great many scientists and professionals have begun addressing the issue. They realize that public esteem for science and scientists can only be harmed when ego and career are valued more highly than the accuracy of the scientific literature and the welfare of the public. The reality of scientific misconduct is painful for everyone who loves science and respects those who devote their lives to it -- myself included. But if the reality is unpleasant, the solution is to change it.
Some have proposed that the problem be attacked with ethics courses for junior scientists. But no ethics seminar can counteract the example set by leading scientists who engage in misconduct or who excuse it in others. Dr. Margot O'Toole can tell you that it is often idealistic young scientists who pay the heaviest price for honorably seeking to protect the integrity of science.
Some have claimed that more "due process" is needed. Due process is certainly appropriate and desirable. But the problem is not so much the amount of process as its quality. It is difficult for a group of professors to investigate impartially someone who is a colleague and, in many instances, a personal friend. It is asking a great deal to hope that an institution will assiduously pursue charges that, if substantiated, may cause it embarrassment, bad publicity, and the loss of funds. In many instances, it is unrealistic to expect the NIH to investigate aggressively the possibility that it erred in its selection of grant recipients.
Who, then, should resolve allegations of misconduct? And what procedures should be followed? These questions are being debated intensely today, and I have no ready answer. Scientists and their institutions are better equipped than others to do this job, and I would greatly prefer to see the scientific community handle the problem. The worst place to resolve these questions is in a congressional hearing. Congressional hearings are rather blunt instruments, poorly suited to making fine distinctions of fact. They are a forum of last resort, whose function is to shine light on severe problems not being addressed. The subcommittee would not have needed to pursue any of the cases I have described here had the institutions and people involved pursued them properly from the start. I would go even further: there is absolutely no good reason why a congressional subcommittee should be doing a job that scientists can and should do themselves.
Encouraging science to police itself is far preferable to the alternatives. I have heard the phrase "science police" bandied about, and that is a ghastly idea. But with every Darsee or Breuning whose case is covered up or mishandled, pressure builds for such extreme measures. Scientists are allowed to monitor themselves on the theory that they are honorable and devoted to truth. They are granted considerable leeway on the theory that their work is of transcendent value and serves the public good. Only so many times can prominent scientists refuse to correct an article on the grounds that a correction may hurt a wrongdoer's career or is not worth the trouble because the journals are riddled with error anyway, before the public begins to wonder why it should fund literature that scientists themselves treat with disrespect. Scientists need to understand that the best way, perhaps the only way, to avoid the threat of "science police" is for scientists themselves to show that they have the ability and the will to police themselves. It is a matter of morality, but also of self-interest.
I hope that the roots of my interest in this issue are understood, and not just the policy or budgetary roots. In 1943 my father, then a member of Congress from Michigan, cosponsored a pioneering bill on national health insurance, and throughout his career he was one of Congress's strongest supporters of federally funded research and of the NIH. My brother is a research scientist, and my wife devotes much time to efforts on behalf of medical research.
One of the most upsetting and wrongheaded criticisms of the subcommittee's work on scientific misconduct is the charge that it would cause young idealists to pursue careers elsewhere. In all too many of the cases with which we are familiar, young, idealistic scientists challenged the system and blew the whistle. We do not need courses in ethics to teach young scientists to be idealistic. We need institutions that will say to them, Don't be afraid to ask hard questions. Don't ever lose that idealism. After all, I remember what it was like to be a young, idealistic member of Congress.
I began this discussion by talking about our innate American sense of optimism. After years of fighting uphill battles, after accumulating layers of scar tissue, I may no longer believe in the perfectibility of human institutions or societies, but I do believe in their improvement. One of the most valuable tools we have toward that end is science.
Addendum
Since I gave the Shattuck Lecture last year, judgments have been handed down in three of the cases I cited. In the case of Dr. Imanishi-Kari, in July 1992 the U.S. attorney in Baltimore formally declined to prosecute her. In statements to the press as well as in correspondence with the subcommittee and the Secret Service, U.S. Attorney Robert Bennett stressed that the declination was based on the difficulty of presenting the complex scientific facts to a lay jury, which would have to understand them fully to find guilt beyond a reasonable doubt in a criminal trial. Bennett emphasized that the declination was not an exoneration of David Baltimore or Thereza Imanishi-Kari, nor did it reflect doubt on the part of the prosecutor's office that the data had been falsified.
In the case of the Cleveland Clinic, the Office of Research Integrity of the Department of Health and Human Services ruled late last year not only that the scientist, Dr. Sharma, had committed scientific misconduct, but also that the initial inquiry "lacked competence, thoroughness, and objectivity" and "was hurried and superficial."
After a six-day trial, a jury convicted Dr. Prem Sarin on July 8, 1992, on three counts of embezzlement and false statements to the NIH.
Finally, in the case of Dr. Robert Gallo, the Office of Research Integrity found last December that Dr. Gallo had intentionally misled the scientific community by claiming that he had not grown the lymphadenopathy-associated virus obtained from French investigators in a permanent cell line.
With respect to Dr. Gallo's service as a referee on a manuscript submitted to a journal by scientists from the Institut Pasteur, the Office of Research Integrity found that "Dr. Gallo's actions reflect Dr. Gallo's propensity to misrepresent and mislead in favor of his own research findings or hypotheses." The office also faulted Dr. Gallo's laboratory supervision, asserting that it "reflects irresponsible laboratory management that has permanently impaired the ability to trace the important steps taken" in Dr. Gallo's early AIDS research. The Office of Research Integrity noted that the importance for science and the public health of the work relating to AIDS "imposed an obligation in reporting the methodologies and results of this groundbreaking research" -- an obligation, the office said, that Dr. Gallo "failed to meet."
I am indebted to Reid P.F. Stuntz, Peter D.H. Stockton, Bruce F. Chafin, Dennis Fitzgibbons, and Janina A. Jaruzelski for their assistance as congressional staff members in the subcommittee's investigations and the preparation of this lecture, and to Dr. Margot O'Toole, Dr. Ned Feder, Mr. Walter Stewart, Dr. Suzanne Hadley, and other subcommittee witnesses courageous enough to speak the truth, for their invaluable contributions to the subcommittee's work.
Source Information
Presented as the 102nd Shattuck Lecture to the Annual Meeting of the Massachusetts Medical Society in Boston on May 9, 1992.Congressman Dingell (D-Mich.) is chairman of the Subcommittee on Oversight and Investigations of the U.S. House of Representatives.
References
| |||||||||||||||||||||||||||||||
Related Letters:
Shattuck Lecture -- Misconduct in Medical Research
Baltimore D., Wortis H. H., Huber B., Woodland R., Onek J. N., Delaney M.
Extract |
Full Text
N Engl J Med 1993;
329:732-734, Sep 2, 1993.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
