There is a need for more hospital-based ethics education and more interdisciplinary and cross-subspecialty discussion of inherently complex and stressful pediatric end-of-life cases. Education should focus on establishing appropriate goals of care, as well as on pain management, medically supplied nutrition and hydration, and the appropriate use of paralytic agents. More research is needed on clinicians' regard for the dead-donor rule.

The oversight of research involving human participants is widely believed to be inadequate. The U.S. Congress, national commissions, the Department of Health and Human Services, the Institute of Medicine, numerous professional societies, and others are proposing remedies based on the assumption that the main problems are researchers' conflict of interest, lack of institutional review board (IRB) resources, and the volume and complexity of clinical research. Developing appropriate reform proposals requires carefully delineating the problems of the current system to know what reforms are needed. To stimulate a more informed and meaningful debate, we delineate 15 current problems into 3 broad categories. First, structural problems encompass 8 specific problems related to the way the research oversight system is organized. Second, procedural problems constitute 5 specific problems related to the operations of IRB review. Finally, performance assessment problems include 2 problems related to absence of systematic assessment of the outcomes of the oversight system. We critically assess proposed reforms, such as accreditation and central IRBs, according to how well they address these 15 problems. None of the reforms addresses all 15 problems. Indeed, most focus on the procedural problems, failing to address either the structure or the performance assessment problems. Finally, on the basis of the delineation of problems, we outline components of a more effective reform proposal, including bringing all research under federal oversight, a permanent advisory committee to address recurrent ethical issues in clinical research, mandatory single-time review for multicenter research protocols, additional financial support for IRB functions, and a standardized system for collecting and disseminating data on both adverse events and the performance assessment of IRBs.