The concept of worker notification has evolved over the years to a prevailing view that workers should be notified of all epidemiologic studies that make use of their medical information. To adequately communicate health information to workers, epidemiologists must consider the intended audience along with its need for risk information that is scientifically valid and clearly outlines areas of uncertainty. This goal is facilitated by full disclosure to workers of information pertinent to the planning, conduct, and interpretation of research efforts. Subjectivity in choosing unanticipated research findings for inclusion in worker notifications may permit bias toward a personal viewpoint, but this can be minimized by prestudy agreement on appropriate criteria for selecting which research findings to communicate. Epidemiologic theory and principles of causal inference should guide the development of appropriate criteria. The timeliness of worker communications has received limited attention, and workers have often been the "last to know" about important studies. This may influence workers' receptivity to the risk message. Sponsoring organizations should ensure that an acceptable communications plan is included in research protocols and that the plan accords priority to notifying workers about study results.