The current extraordinary advances in basic biomedical and social sciences have unprecedented potential to improve the human condition. These insights, together with the human genome project and its successor proteomics, will require an enormous commitment to translational research to harvest their applications for medicine and public health. The progress in political and social theory, linguistics, statistics, psychology, and behavioral sciences generally deserves a similarly broad application of human research to reach full expression. These endeavors will require individuals to accept possible risk to themselves, benefiting the greater good with uncertain (or no) benefit accrued in return. Thus, those performing and overseeing research are obligated to provide the most reasonable assurances of safety possible.

Events in recent years have evoked considerable public concern about the safety of human studies and the measures in place to protect subject-participants. Isolated cases of unknown representativeness capture enormous public attention. The death of a subject during bronchoscopy, the death of a patient in a gene transfer experiment, the death of a healthy volunteer in an asthma study—these and similar occurrences are stridently announced and exhaustively analyzed for their heuristic guidance. As individual human tragedies they deserve all the attention they receive, including minute analysis as sentinel events from which to draw preventive guidance for future studies. But at this time, there is no way to know how representative these terrible events are and therefore no way to know what general lessons to infer from them for the uniquely human endeavor of studying some humans for the possible benefit of others. And, without quantification of the problem, there is no way to assess the appropriateness of new protections and their inevitable costs.

I believe any thinking person could be forgiven for assuming that there exists a reliable quantitative picture of the number of people harmed or injured each year as a result of their participation in research (let us call it the numerator)—and for assuming that there must similarly exist a comprehensive if not exhaustive registry of all individuals participating as “subjects” in biomedical or social science research (the denominator). Relating these to each other, society could reckon both the absolute magnitude of the problem and the fractional risk of harm confronting individuals considering participation. Frustratingly, neither figure is available, nor, without considerable effort, will they be forthcoming in the near future. This collective uncertainty has huge consequences.

An awareness of these problems led the Secretary of the Department of Health and Human Services to engage the Institute of Medicine for this study. We were asked to conduct an expeditious analysis of the possible value of accreditation of human research protection programs, which was published in April 2001. In addition, we were asked for a more comprehensive review of the present system for protecting human participants and to make suggestions for strengthening it. The latter is now offered.

Given the paradox of both the enormous potential and the uncertain reality in this area, what do we propose? We suggest that we address our current uncertainty by national data collection that in a few years should quantify and delimit the problem. However, until evidence is available, we believe it is necessary to pursue every promising mechanism to maximize the protection of individuals participating in research.

In order to accomplish this goal, we recommend that all research involving humans take place in settings or through organizational affiliations where the culture, announced and exemplified by the leadership, emphasizes the highest ethical standards and dedication to the welfare of every study participant. We further urge that all individuals responsible for the conceptualization and conduct of research be specifically trained in research ethics as well as techniques. We believe that the complexity of current science and, in the instance of clinical trials, the biological high stakes, require a review that begins with the scientific quality of the proposed research and a parallel examination of potential conflict of interest. These results would then lead into an explicit review of the ethical dimensions of the study. We urge a new approach to informed consent, one in which legal disclaimer and institutional self-protection are second to clear, simple, unclouded, unhurried, and sensitive disclosure that gives the potential participant all the information a reasonable person would need to make a well-informed decision, and the time to do so. Financial conflicts of interest at any stage or level of the process must be disclosed and managed so that the objectivity of research is preserved and the public trust is upheld.

We believe that the necessary efforts to accomplish these tasks are currently under-resourced, and we urge that that need for sufficient support be recognized and provided as a cost of the responsible conduct of research. And we believe that accreditation of programs should be explored as a means of achieving excellence, for it anneals two core mechanisms for improvement: self-assessment and the sharing of best practices.

We do not, however, urge a permanent accretion of new regulations and bureaucracy. Rather, we believe the protection system should be reexamined at a time when the steps described above can be accomplished, probably about five years from now. We strongly urge that a new look be taken when the magnitude of the problems and challenges facing the system can be appreciated and the appropriateness of protective mechanisms assessed. To do no harm is impossible. But to minimize harm while enabling the benefits of progress to emerge should be an attainable ideal.

Daniel D. Federman, M.D., Chair