Technology assessment in health care: group process and decision theory.

As biomedical technologies proliferate, it is incumbent upon the scientific community to monitor, validate, and encourage adoption of worthwhile procedures, drugs, and devices across the interface from research to practice. As the largest U.S. sponsor of biomedical research, the National Institutes of Health (NIH) formally established the Consensus Development Program and the Office of Medical Applications of Research (OMAR), in order to foster and improve the translation of biomedical research results into knowledge useful in the practice of medicine and public health. Individual technology assessments are conducted through a succession of consensus development conferences, which convene expert biomedical scientists, practicing clinicians, and public representatives in an effort to assess safety and efficacy, and to recommend clinical application of important medical technologies. The Consensus Development Program has evolved through three distinct stages. The "first generation," from 1977-82, initiated the experimental, untested concept of consensus development; the "second generation" (1982-84) stressed formulation of the code of standard operating principles enunciated above; and the ongoing "third generation" is testing the utility of formal data synthesis to augment assessments. OMAR is experimenting with an explicit, normative and analytic approach to aid technology assessments. Real-time microcomputer-based decision models are created to help the panel explore the implications of the data. This paper describes and discusses decision analysis and its potential applicability to medical technology assessment and consensus development. It explores OMAR's experience in testing the model during several consensus development conferences, as well as plans and projections for future investigation and implementation.