Introduction

Scientists in the Intramural Research Program of the National Institutes of Health generally are responsible for conducting original research consonant with the goals of their individual Institutes, Centers, and Divisions.

NIH scientists, as all scientists, should be committed to the responsible use of the process known as the scientific method to seek new knowledge. While the general principles of the scientific method—formulation and testing of hypotheses, controlled observations or experiments, analysis and interpretation of data, and oral and written presentation of all of these components to scientific colleagues for discussion and further conclusions—are universal, their detailed application may differ in different scientific disciplines and varying circumstances. It is clear, however, that only by adherence to the highest standards of intellectual honesty in formulating, conducting and presenting research does science advance and do scientists fulfill their contract with the community at large.

These Guidelines state general principles that NIH scientists are expected to follow in their research activities with regard to supervision of trainees, data management, publication practices, authorship, peer review and use of privileged information, and clinical investigations in order to promote the uniform application of the highest ethical standards to the conduct of all scientific research. It is the responsibility of each Laboratory or Branch Chief, and successive levels of supervisory individuals (especially Institute, Center, and Division Intramural Research Directors), to insure that each NIH scientists is cognizant of these Guidelines and to resolve issues that may arise in their implementation.

These Guidelines supplement existing NIH policies on the conduct of research. In particular, those policies concerning Institutional Review Board oversight of clinical research protocols; animal use; radiation, chemical and other safety issues; and other aspects of the Standards of Conduct for all federal employees remain parts of the canon of conduct for each scientist.

The formulation of these Guidelines is not meant to codify a set of rules, but rather to make explicit patterns of scientific practice that have been developed over many years and are followed by the vast majority of scientists, and to provide benchmarks when problems arise. Although no set of guidelines, or even explicit rules, can prevent willful scientific misconduct, it is hoped that formation of these Guidelines will contribute to the continued clarification of the application of the scientific method in changing circumstances.

The community will ultimately judge the NIH by its adherence to these intellectual and ethical standards, as well as by its development and application of important new knowledge through scientific creativity.

Supervision of Trainees

Research training is a complex process, the central aspect of which is an extended period of research carried out under the supervision of an experienced scientific mentor. This supervised research represents not merely performance of tasks assigned by the supervisor but rather a process wherein the trainee takes on an increasingly independent role in the choice of research projects, development of hypotheses and the performance of the work. Indeed, if training is to prepare a young scientist for a successful career as a research investigator, it must be geared toward providing the trainee with the aforementioned skills and experiences. It is particularly critical that the mentor recognize that the trainee is not simply an additional laboratory worker.

Each trainee should have a designated primary scientific mentor. It is the responsibility of this mentor to provide a training environment in which the trainee has the opportunity to acquire both the conceptual and technical skills of the field. In this setting, the trainee should undertake a significant piece of research, chosen usually as the result of discussions between the mentor and the trainee, which has the potential to yield new knowledge of importance in that field. The mentor has the responsibility to supervise the trainee's progress closely and to interact personally with the trainee on a regular basis in such a way as to make the training experience a meaningful one. Styles of research differ, both among fields and among investigators in a given field, so that no specific rules should be made about the number of trainees that is appropriate for a single mentor to supervise. Nonetheless, mentors should limit the number of trainees in their laboratory to the number for whom they can provide an appropriate research experience.

There are certain specific aspects of the mentor-trainee relationship that deserve emphasis. First, mentors must be particularly diligent in avoiding the involvement of trainees in research activities that do not provide meaningful training experiences but which are designed mainly to further research or development activities in which the mentor has a potential monetary or other compelling interest. Second, training must impart to the trainee appropriate standards of scientific conduct. The mentor conveys these standards by instruction and by example. Third, mentors have a responsibility to provide trainees with realistic appraisals of their performance and with advice about career development and opportunities.

Data Management

Research data, including detailed experimental protocols, primary data from laboratory instruments, and procedures of reduction and analysis of primary data, are the essential components of scientific progress. Scientific integrity is inseparable from meticulous attention to the acquisition and maintenance of these research data.

It is expected that the results of research will be carefully recorded in a form that will allow continuous access for analysis and review. Attention should be given to annotation and indexing of notebooks to facilitate detailed review of data. All data, even from observations and experiments not directly leading to publication, should be treated comparably. Research data should always be immediately available to scientific collaborators and supervisors for review. In collaborative projects involving different units, all investigators should know the status of all contributing data and have direct access to them.

Research data, including the primary experimental results, should be retained for a sufficient period to allow analysis and repetition by others of published material from those data. In some fields, five or seven years are specified as the minimum period of retention but this may vary under different circumstances.

Publication Practices

Publication of experimental results is an integral and essential component of research. Other than presentation at scientific meetings, publications in a scientific journal should normally be the mechanism for the first public disclosure of new findings. Although appropriately considered the end point of a particular research project, publication is also the beginning of a process in which the scientific community at large can substantiate, correct and further develop a particular set of results.

Timely publication of new and significant results is important for the progress of science, but fragmentary publication of the results of a scientific investigation or multiple publications of the same or similar data are inappropriate. Each publication should make a unique and substantial contribution to its field. As a corollary to this principle, tenure appointments and promotions should be based on the importance of the scientific accomplishments and not on the number of publications in which those accomplishments were reported.

Therefore, each paper should contain all the information that would be necessary for the scientific peers of the authors to repeat the experiments. This principle requires that any unique materials (e.g., monoclonal antibodies, bacterial strains, mutant cell lines), analytical amounts of scare reagents and unpublished data (e.g., protein or nucleic acid sequences) that are essential for repetition of the published experiments be made available to other qualified scientists. It is not necessary to provide materials (such as proteins) that others can prepare by published procedures, or large quantities of materials (such as polyclonal antisera) that may be in limited supply, although it is desirable to do so.

Authorship

Authorship refers to the listing of names of participants in all communications, oral and written, of experimental results and their interpretation to scientific colleagues. Authorship is the fulfillment of the responsibility to communicate research results to the scientific community for external evaluation.

Authorship is also the primary mechanism for determining the allocation of credit for scientific advances and thus the primary basis for assessing a scientist's contributions to developing new knowledge. As such, it potentially conveys great benefit, as well as responsibility. For each individual the privilege of authorship should be based on a significant contribution to the conceptualization, design, execution, and/or interpretation of the research study, as well as a willingness to take responsibility for the defense of the study should the need arise. In contrast, other individuals who participate in part of a study may more appropriately be acknowledged as having contributed certain advice, reagents, analyses, patient material, support, etc., but not be listed as authors. It is expected that such distinctions will be increasingly important in the future and should be explicitly considered more frequently now.

In recent years, there has been a rapid increase in the average number of authors per communication. In part, this increase is due to the needs of modern research projects for contributions from many individuals, frequently those with different specialized skills. While multi-authorship is not a problem in itself, it raises many issues such as criteria for inclusion as an author, ability of each author to evaluate and defend all aspects of a study, sequence of listing of authors, and separation of various experimental results to increase numbers of communications and authorship citations. To clarify some of these concerns, consideration should be given in interdisciplinary studies to preparing brief statements of the exact contribution of each author to the work described in each communication.

Because of the variation in detailed practices among disciplines, no universal set of standards can be easily formulated. It is expected, however, that each research group and Laboratory or Branch will freely discuss and resolve questions of authorship before and during the course of a study. Further, each author should review fully material that is to be presented in public forums or submitted (originally or in revision) for publication. Each author should be willing to support the general conclusions of the study and be willing to defend the study.

The submitting author should be considered the primary author with the additional responsibility of coordinating the completion and submission of the work, satisfying pertinent rules of submission, and coordinating responses of the group to inquiries or challenges. The submitting author should assure that the contributions of all collaborators are appropriately recognized and must be able to certify that each author has reviewed and authorized the submission of the manuscript. The recent practice of some journals in requiring approval signatures from each author before publication is felt to be a useful step in regard to fulfilling the above.

Peer Review and Privileged Information

Peer review can be defined as expert critique of either a scientific treatise, such as an article prepared or submitted for publication, a research grant proposal, a clinical research protocol, or of an investigator's research program, as in a site visit. Peer review is an essential component of the conduct of science. Decisions on the funding of research proposals and on the publication of experimental results must be based on thorough, fair and objective evaluations by recognized experts. Therefore, although it is often difficult and time-consuming, scientists have an obligation to participate in the peer review process and, in doing so, they make an important contribution to science.

Peer review requires that the reviewer be expert in the subject under review. The reviewer, however, should avoid any real or perceived conflict of interest that might arise because of a direct competitive, collaborative or other close relationship with one or more of the authors of the material under review. Normally, such a conflict of interest would require a decision not to participate in the review process and to return any material unread.

The review must be objective. It should be based solely on scientific evaluation of the material under review within the context of published information and should not be influenced by scientific information not publicly available.

All material under review is privileged information. It should not be used to the benefit of the reviewer unless it previously has been made public. It should not be shared with anyone unless necessary to the review process, in which case the names of those with whom the information is shared should be made known to those managing the review process. Material under review should not be used in any manner by the reviewer unless specifically permitted by the journal or reviewing organization and the author.

Clinical Research

Clinical research, for the purposes of these Guidelines, is defined as research performed on human subjects as part of human experimentation. All of the topics covered in the Guidelines also apply to the conduct of clinical research; clinical research, however, entails further responsibilities for investigators.

The preparation of a written research protocol (''Clinical Research Protocol") according to existing guidelines prior to commencing studies is almost always required. By virtue of its various sections governing background; patient eligibility and confidentiality; data to be collected; mechanism of data storage, retrieval, statistical analysis and reporting; and identification of the principal and associate investigators, the Clinical Research Protocol provides a highly codified mechanism covering most of the topics covered elsewhere in the Guidelines. The Clinical Research Protocol is generally widely circulated for comment, review and approval. It should be scrupulously adhered to in the conduct of the research. The ideas of the investigators who prepared the protocol should be protected by all who review the document.

Clinical investigators are responsible for assuring that the proposed clinical research will be conducted only if the Clinical Center, or other clinical facilities, has the appropriate capability and support structure to insure that the research can be done safely and efficiently. The principal investigator should be familiar with the functioning of the clinical unit and should allow the investigation to continue only if the unit can provide adequate clinical care.

Investigators who are neither clinicians nor trained in clinical research may perform laboratory research on material derived from humans. To conform to the requirement of working under approved human experimentation guidelines, they should ordinarily be advised by or collaborate with trained clinical investigators.

The supervision of trainees in the conduct of clinical investigation is complex. Often the trainees are in fellowship training programs leading to specialty or subspeciality certifications as well as in research training programs. Thus, they should be educated in general and specific medical management issues as well as in the conduct of research. The process of data gathering, storage, and retention can also be complex in clinical research and sometimes not easily subject to repetition. The principal investigator is responsible for the quality and maintenance of the records and for the training and oversight of all personnel involved in data collection.

Concluding Statement

These Guidelines are not intended to address issues of misconduct, i.e., fabrication, falsification, plagiarism or other practices motivated by intent to deceive. Rather, their purpose is to provide a framework for the fair and open conduct of research without inhibiting scientific freedom and creativity.

Footnotes

NOTE: Issued in 1990; reprinted courtesy of the National Institutes of Health, Bethesda, Md.