It deserves to be emphasized that correction of a published article does not necessarily indicate misconduct on anyone's part. Frequently new techniques and better methodologies are discovered, so that an old problem can be re-examined with more forceful means. Continuous upgrading of the literature is part of the scientific process. We are now fortunate that computerized literature data banks, such as MedLine from the Library of Medicine, tie any retraction or errata into the original article so that a researcher requesting a citation will automatically get the citation plus the subsequent corrections. - The responsibility of journals in dealing with fraudulent data is identical to its responsibilities in dealing with error it's just harder to do. The journal does not have the authority to require an author to appear on its doorstep, open his or her files or answer questions. All it can do is reject or retract a paper. That is why the investigation of fraud is a responsibility properly assigned to an author's institution and/or funding agency. A journal can help initiate an investigation by consulting an author's institution if allegations of fraud are first brought to the editor, but resolution often requires the weight of officialdom. In some cases in which reproducibility of information is challenged, an author will resist retraction, and a journal that unilaterally announces such retraction exposes itself to libel lawsuits. At some future time there may be some need to modify libel laws so that a statement by some of the authors that they no longer have confidence in their own article is not actionable by the remaining authors. In some cases that problem can be circumvented, but it can be difficult and expensive to get the necessary information to the reader. Nevertheless, it is the responsibility of the journal to do so, and it is the primary responsibility of the journal which published the original data to publish the retraction. Differences among Journals It is worth adding that not all journals are identical and scientists read articles with considerable skepticism. There are certain premier journals in which peer review is rigorously applied and the most exacting criticism is made of any article submitted. There are other journals that are more lenient and allow more speculative discussions. A variety of journals with different standards of rigor is good for science, just as a variety of magazines ranging from frivolity to sober facts in non-scientific areas is good for society. Scientists read some journals with great trust, others with more doubt. Skepticism increases even in a very rigorous journal when a novel approach or a new theory is presented. Readers are rarely looking for fraud; they understand, however, that pioneers beating their way through the jungles of difficult experiments may sometimes be overly optimistic. In seeking to minimize fraud, misconduct and honest error, it is important to realize that a blind faith in all of the literature is not characteristic of scientists. A journal tries to publish the best, but it is far better to publish an imaginative idea that turns out to be wrong than to reject a creative paper that later turns out to be seminal. Importance of Peer Review The role of journals is to save the time of the reader by screening out those claims of accomplishment that are likely to be incorrect. At journals such as Science this is done by peer review, and the quality of that peer review is crucial: it must be thorough and candid to alert the journal to sloppy scholarship, poorly designed experiments or erroneous reasoning; it must also be fair, so that authors are protected against conflict of interest, irresponsible criticism or inappropriate delay in presenting their discoveries to the scientific community. It is the responsibility of the editors of journals to obtain individuals of the highest caliber as peer reviewers, to protect their confidentiality and to expect them to be receptive and unbiased. Although there is some criticism of peer review, there is no doubt that it is essential if journals are to have any general standards. No modern editor can keep up with all of modern science; it is necessary to get the advice of experts in the field. Peer review is like democracy it has many imperfections but, when compared with the alternatives, it is clearly the best device. Honorary Authorship - On the issue of honorary authors, there is inevitably some disagreement. My own view is that no one should be listed as an author of a scientific article who did not contribute significantly to the work it describes. Simply raising money for the research is not enough. Likewise, a person who prepares a sample for part of the experiment and does not have any other role in the project is not entitled to be an author. However, it is not easy to draw the line between that person and another who does one part of a complicated experiment extremely well, understands its place in the whole and contributes an essential part of the investigation, but is not necessarily conversant with the rest of the work. For example, a hormone may be constructed by molecular biologists, tested by physiologists who understand animal assays, and pursued through clinical trials by physicians. Each will be clearly aware of the others' responsibility, but may not be conversant with details of the techniques of their collaborators. Such multi-disciplinary cooperation is the kind of research in which the chance of error is greater than in small groups in which all the contributors know all procedures. Nevertheless, in the modern specialized world, multi-discipline investigations have been extremely productive, and are examples of the kind of cooperation that often gets praise from outsiders. Multi-disciplinary research teams demand more careful administration than simpler projects, but it would be tragic to ban them under rigid definitions of authorship because they account for much of the progress of modern science. Legislation in this area could not be precise and might be harmful to the process of science, but consensus within the scientific community can create a climate to minimize abuse. For example, the scientific community should be tougher than it has been in agreeing that those who benefit from the praise when work goes well should suffer when error or fraud is found. In some cases heads of large groups have protested that they didn't know about co-workers' failures. If that alibi were greeted with the condemnation it deserves, the "honorary author" problem might become a thing of the past. Need for Legislation Indeed, There are some who say there is need for immediate legislation. there is evidence that some university and industry procedures for punishing fraud and protecting whistleblowers need improvement. Universities and granting agencies are proceeding to codify and develop better procedures and are putting them into practice, but we must not sacrifice due process in quest of speed. It would seem, therefore, that Congress should keep a watchful eye, but not rush into a new legislation unless they perceive the institutions most directly involved cannot handle the situation. A few cases of fraud have received high publicity, but there are about 40,000 scientific journals in the world and over 500,000 Ph.D. scientists in the United States alone. The number of cases of fraud are small against this large background. Fraud is deleterious to science and must be deterred, but it would be tragic to set up a cumbersome bureaucratic machinery which did more to retard science than the fraud we are trying to prevent. Summary Scientific journals play an important role in protecting against misconduct through the devices of publishing retractions, corrections and errata. They create the climate in which errors of any significant sort are eventually going to be exposed. They can sometimes be useful in differentiating deliberate fraud from honest error, but usually that assignment will have to be determined by the institution that employs the scientist. The journals' primary responsibility rests in maintaining the integrity of the scientific literature. Most of the literature is accurate and the amount of deliberate, high-impact fraud is small. Scientists are adventurers, exploring the frontiers of knowledge. If scientific pioneers return from adventures deliberately misrepresenting the Promised Land in falsely glowing colors, they should be denounced as frauds and mountebanks, and be drummed out of science. If, however, in pursuing these grand adventures, they come back triumphant but with dirty fingernails, we should not exaggerate that imperfection to make laws which would decrease future discoveries. Mr. JOHNSTON. Thank you, Dr. Koshland. We have the pleasure of Dr. George Lundberg introducing our next speaker. Doctor. TESTIMONY OF DR. GEORGE D. LUNDBERG, VICE PRESIDENT FOR SCIENTIFIC INFORMATION, THE AMERICAN MEDICAL ASSOCIATION Dr. LUNDBERG. Thank you, Mr. Chairman. My name is George D. Lundberg, M.D., and I present this statement on behalf of the American Medical Association. For the past seven and one-half years I have been Vice-President for Scientific Information of the American Medical Association. In this position I am responsible for the editorial content of all 10 scientific journals of the AMA. The nine specialty journal editors report directly to me. I myself edit the Journal of the American Medical Association, so-called JAMA. These journals have very large circulations, and they help to further two of the AMA's highest priorities-the education of the profession and the public and the maintenance of high ethical standards. At the request of the Committee and in the interest of time constraints, I will not read the main segments of my statement which have been submitted to the committee. JAMA has published many papers on this subject, and we have provided copies of eight relevant documents for the committee. Dr. Drummond Rennie was once a Deputy Editor of the New England Journal of Medicine, but has been an editor with JAMA since 1982, and he now holds the newly created position of Deputy Editor (West) of JAMA, at the same time, being a professor of medicine at the University of California, San Francisco. At that institution he is setting up a unique program for the study of biomedical publication at the Institute for Health Policy Studies, a program which is consonant with the AMA's academic and ethical goals. Dr. Rennie recently organized and directed the first International Congress on Peer Review in Biomedical Publications sponsored by the AMA, held six weeks ago in Chicago and attended by more than 260 editors from 24 countries. 98 Dr. Rennie has studied this problem and has made several suggestions to improve the process. These suggestions are not as yet policy for our journals, but all are under active consideration. I hope your Committee gives careful consideration of Dr. Rennie's proposals, especially the part about an experimental audit, which suggests a way to get at some of the questions which the Committee has, I believe, duly been asking through the course of today. Thank you, Mr. Chairman. I will be happy to answer any questions that you might have later. [The complete prepared statement of Dr. Lundberg follows:] 98 The conference program and abstracts appear in supplemental materials submitted by Dr. Rennie. |