Traditional epidemiological approaches based on biomedical models may be limited with respect to their response to "outbreaks of concern" among work-force or community populations. Three published Canadian "outbreaks" were reviewed in this study. In all three instances, research was initiated because lay persons were concerned about either nonspecific symptoms or hazardous exposures, and individuals publicly called for a response. Epidemiologic analyses were inconclusive as to the reasons for the outbreaks, and they contributed little toward the resolution of concerns. There is a need for a fuller recognition of the role of social context and of the action-oriented nature of such research. The elucidation of multifactorial and culturally mediated causation, as well as the development of remedial actions, require a rethinking of research methods. We specifically call for (a) an expansion of the disciplinary base of research teams to include social scientists, and (b) the adoption of combined qualitative and quantitative research approaches.