RESPONDING TO ALLEGATIONS OF MISCONDUCT IN THE UNIVERSITY 8. Peer review: its adequacy to detect fraud without the necessity of special regulations was still a subject of argument in 1985. There remains disagreement whether well-publicized procedures for handling allegations of misconduct would in themselves constitute a deterrent to improper activities. But few institutions have a sufficiently searching independent peer review of research results to detect deliberate deception or scientific error. Peer awareness of improper activities is a different matter, and raises again the question of how to educate peers to report misconduct or error as a matter of scientific responsibility. B. Influence of federal agencies on policies and actions As noted above, both the NIH and the NSF have made recommendations concerning the responsibilities of grantee institutions with regard to research misconduct, and are further concerned whether more active oversight should be pursued. The DHHS invited comments on a wide range of possible actions intended to "protect against scientific fraud or misconduct," in the Federal Register of September 19, 1988 (FR 53: 36344), with the intent of implementing far-reaching regulations. The thoroughness of the FDA in reviewing the research records of new drug trials is a model of what a government agency can do when it has a crucial concern for the accuracy of specific research results. The model which is thereby enforced research using fixed protocols, relying minimally on the initiative or originality of the investigator - is not one appropriate to the exploratory science of investigator-initiated grants. It could be argued that contract research is somewhere inbetween, and in fact the auditing of such research, such as cancer therapy trials, is more thorough. Policy decisions on research regulation by these major agencies, normally made in response to a combination of political pressure and informed public opinion, have great impact on research institutions. The protection of human and animal subjects of research, the latter still the topic of much maneuvering by pressure groups, has markedly changed as a result of decisions to require grantees to meet certain formal standards of review, with policies and procedures subject to approval by the granting agency. It is difficult to argue that these changes were not for the better, even though they have impaired the freedom of investigators to try out their ideas, by imposing extra delays and expenses for every new initiative or protocol modification, and have created a costly bureaucracy in research institutions. The ethical standard of treatment of both human and animal research subjects is far better than formerly because of these requirements imposed by outside agencies, but under procedures developed to some degree by each institution, and implemented by its own faculty and staff. If the federal agencies decide that guaranteeing the reliability of research results requires a comparable organizational effort to monitor investigator's proposals and publications, then institutions will be forced to devote considerable faculty and staff resources to such an effort. Less costly and burdensome means of ensuring the quality and ethics of science would be preferable. It is hoped that the several granting agencies will at least develop consistent or identical requirements to facilitate institutional compliance. RESPONDING TO ALLEGATIONS OF MISCONDUCT IN THE UNIVERSITY A. VII. COMPLICATING FACTORS Extra-procedural judgments in screening charges and complaints As discussed above, it requires subjective judgment to decide which complaints or charges should be treated as serious issues that require use of the formal procedures for responding to allegations of misconduct. Regulations cannot specify how to treat all the variations of research misconduct defined in the first section of this paper, nor can they help decide how to approach the accused or his employees in the preliminary information-gathering part of an inquiry. Evaluation of the credibility and motives of the accuser and other informants is difficult and subjective, and, though necessary, does not automatically answer questions about the validity of the charges. It may require an individual who is disgruntled or unsuccessful as a researcher to take the professional risk of coming forward with negative information. Someone who is on good terms with a scientist is certainly less likely at present to report suspected misconduct. The sympathetic authority learning of various abuses may find that no real action can be taken, because the abuses are not sufficiently egregious or documentable to be prosecutable, or the complainant does not want to draw attention to himself or herself by making formal charges. Individuals whose sharp practices have alienated others may be quite aggressive in their countercharges, so that an acknowledgment of misconduct or wrongdoing seems a less likely response than a lawsuit. In contrast, some of the exploitative practices of senior investigators are hallowed by custom, so that it may take a generation of enhanced ethical training to make it possible to change them. B. Deleterious aspects of reviews or audits by outsiders One solution to questions about the objectivity of internal review or investigative committees is to let an outside agency do the whole job. Fiscal auditing, for example, may be undertaken by a granting agency in addition to whatever supervision is provided by the research institution. Drug and cancer therapy trials are usually audited by outside organizations hired for this purpose by the supervisory or funding agency. Review of the validity of original research, with much more diverse protocols, however, requires technical expertise not likely to be available at the staff level. There is a risk that outside reviews would receive less cooperation from faculty or other employees of the institution than internal committees. There can also arise the attitude that the outside auditors are there to be fooled or cheated, and local ethics may actually suffer. c. Loss of due process in trial by public hearing or the media As much as due process is emphasized in the policies and procedures currently being developed by agencies and institutions, these may be completely ignored by less responsible persons or groups intent on making a point. A recent example was the one-sided RESPONDING TO ALLEGATIONS OF MISCONDUCT IN THE UNIVERSITY testimony before a Congressional committee that went beyond the general to a particular case of alleged research misconduct, in which neither formal charges nor an opportunity to reply to a great deal of innuendo was provided to the senior investigators involved in the project. The media may also be responsible for one-sided reporting, such as when an accused person follows legal counsel's advice and refuses to engage in a public dialogue, or when it quotes uncritically the remarks of individuals who are significantly biased. It is not clear how to guard against these complications by internal policies or procedures. D. Intimidation by legal challenges and processes Scientists don't really understand the workings of the law, and may not approve of them when they do. They distrust the adversary nature of legal proceedings as opposed to a cooperative attempt to find the truth. At several steps in the process, the threat of legal action, or of failing to meet the requirements of a legally sound process and decision, can shadow the faculty making or responding to allegations of misconduct. The administration is more familiar with legal standards and may have more ready access to legal advice, but finds some of the requirements of the law to be unrealistic in the academic setting, and its standards for proof and evidence to be uncongenial or incomprehensible. A complication of the process of correction of misconduct is the question of the liability of the accused. Legal advice may be that it is safer to acknowledge no wrongdoing, and thereby avoid liability, unless the institution is willing to pursue its case in court, which it is unlikely to wish to try. This leaves an unsatisfactory outcome, with the individual unable to pursue a research career in the face of an essentially unresolved dispute. Probation or rehabilitation are precluded by the failure to admit wrongdoing. An institution's actions are further complicated by concerns about the law's strong protections of an individual's entitlement to employment or pursuit of one's career. Disciplinary actions are difficult to impose in the absence of the consent of the accused. E. Factors inhibiting significant change in customary response The customary response has been to cover up or ignore misconduct and it remains the least painful way out of many situations. However high the ethical standards of an institution, it is still an uncomfortable process to accuse and investigate a member of the faculty, especially when it is likely to concern something that some other faculty member should have prevented or detected at an earlier stage. It is embarrassing to admit to having been deceived and it is worse to read the newspaper treatment of the issue. The current defensive posture that emphasizes the sanctity of science makes it harder to accept the possibility of misconduct, however, by imposing an unrealistic standard of behavior. Institutions should not allow their preambles to indulge in hyperbole about the purity of science. In Forgive and Remember (Chicago, University of Chicago Press, 1979) the sociologist Charles L. Bosk discusses the RESPONDING TO ALLEGATIONS OF MISCONDUCT IN THE UNIVERSITY professionalization of surgeons with regard to recognition of and response to error. There are significant similarities and differences between their training experience and that of graduate students in research. There are strong controls within a surgical service, like a laboratory, over what is considered acceptable practice, and several kinds of error were described. A technical error, openly acknowledged within the group, was readily forgiven, and blame was not attached to individuals. A judgmental error, which in the context of a trainee was usually a failure to notify the responsible surgeon of a problem, or to fail to show total dedication to the tasks of the service, was considered a moral failing, and had the most serious career implications. Normative errors were equivalent to failing to follow "good laboratory practice" of a general type. Bosk also identified a failure to follow the idiosyncratic practices of the particular service as quasi-normative errors. How trainees learn to identify and avoid these errors is a process that could bear comparable study in the research realm. The surgeon learned to value procedures and proper responses even more than results, which are after all not entirely within the power of the physician to control. Researchers may be more intent on results than on how they are achieved, and clearly have less moral indoctrination than surgeons. Bosk noted, however, that the strong personal moral sense that was inculcated was not matched by a corporate conscience, and he concluded that surgeons would do a better job with a stronger sense of public responsibility and accountability. Some of these generalizations transfer readily to scientific research. VIII. A. RECONSIDERATION OF THE GOALS OF POLICIES ON RESEARCH MISCONDUCT Correction versus prevention It would be better for science to correct more errors and make the admission of error less portentous than to catch more people guilty of deliberate deception. The recent surge of interest in research fraud and other misconduct has led to an emphasis on detecting and fixing problems, rather than on preventing them, with the notable exception of the Institute of Medicine's project on the Responsible Conduct of Research. But if the fundamental goal is to improve the reliability of research, not to bring scoundrels to justice, then this emphasis may be directing us away from the measures that would have the best results. It appears that dishonesty in research commonly starts with sloppiness and a lack of regard for scientific rigor. While fear of being caught may inhibit some from egregious falsehood, every scientist knows that it is only a matter of personal conscience that guarantees that what goes into the notebook is complete and accurate, and that the shaky experiments get repeated rather than accepted just because the numbers are favorable. Development of this conscience, socialization in the ethos of scientific research, and learning how to be as skeptical of one's own work as that of others, are processes that are not understood in depth. Attention to these educational processes and their availability to research trainees would pay off in better quality RESPONDING TO ALLEGATIONS OF MISCONDUCT IN THE UNIVERSITY science, and as a direct consequence, more ethical science. Improving the rules for handling allegations of misconduct, structuring more effective investigative committees or doing prepublication research audits have been discussed in depth, but by their nature can only detect but not prevent error or fraud. Even wide notice of investigative and corrective mechanisms for fraud are not likely to impact on the incidence of the "little sins". And as we have seen in several major fraud cases, fear of sanctions has failed to inhibit terribly self-destructive actions. Errors of the grade of "that was a stupid thing to do" may be indistinguishable from casual falsification or minor deliberate deception; it is more corrective, less divisive, and ultimately more scientifically productive, to handle these as errors. The benefit of the doubt must be given to the honesty of all concerned. On this score, perhaps, this approach differs from that of those dedicated to "cleaning up science." The foregoing argument does not say that there should not be sanctions against those who are fairly judged to have abused their positions as researchers, supported by scientific and educational institutions. The argument is that even aggressive administration of sanctions is unlikely to improve ethical standards, and its bureaucratic trappings will have a chilling effect on the spirit of the great majority of honest investigators. B. New mechanisms for dealing with complaints and charges But to strengthen the quality of science, it is important to maximize openness of criticism, and this should include means of making it easier to report the detection or suspicion of improper activities to responsible authorities. Some institutions have designated an official who is not attached to the laboratory or even the department as a potential recipient of complaints. MIT has an ombudsman (though that clearly has not prevented explosion of a contested charge into the public arena), and others have a designated associate dean. It should be possible for a student or faculty member to talk over an issue and set in motion a process that is not a regular investigation, but is an informal sounding of expert and ethical opinion, that can still preserve at this stage the anonymity of the person raising the question. Such discussions must lack the due process provisions of the more definitive preliminary inquiry and formal investigations spelled out in most procedures adopted in response to NIH or NSF guidelines. An individual may not be satisfied with the results of such an informal process; indeed, the information accessible might be limited. The next step should be to go further in the institution's response procedures, not to jump to outside parties, as happened in the MIT case. The reception accorded the complainant should encourage the preferred path to resolution. Complaints and accusations may appear to be a major nuisance to an institution; to treat them this way would be a serious error. They are a real part of the self-regulatory process of science, a function which we are trying to preserve. If the substance of the complaint suggests misconduct as opposed to error or incorrect procedures, then the institution is obliged to try to pursue the matter, even if the accuser does not wish to play any further role, as urged in the AAU report. Third parties have no standing unless formally invited by |