Given the diversity of assessment methodologies in previous studies, and given the increasing attention being shown in studies in which nausea and vomiting are used as outcome measures, it is reasonable to expect that there are more questions on the assessment of nausea and emesis than there are answers. There is little consensus on the efficacy of various approaches used to help control chemotherapy-induced nausea and emesis. There is also little consensus on the appropriate assessment techniques for their measurement. While there is room for much controlled experimentation and study of the ways to assess nausea and emesis, a review of past studies indicates several guidelines that should be kept in mind. The first of these is that nausea is a subjective phenomenon. It exists only insofar as it is defined by the patient. The clinical reality is that patients' decisions to delay or terminate chemotherapy are based on their subjective evaluations of the side effects, not on evaluations made by a clinician or others. With this in mind, a self-report approach to the assessment of nausea appears the most reasonable at this time. On the other hand, while vomiting does have subjective aspects, it represents a behavior that may be reliably observed by others. Thus, it is appropriate that observer ratings be obtained whenever feasible. The use of such indirect measures of nausea and vomiting as time to food intake, appetite change, and other measures of patient well-being should be considered whenever feasible. As with all measures, the cost in patient time needs to be balanced against the scientific needs of any study quantifying nausea and vomiting. While direct, physiologic measures would provide an important external reference to the assessment of nausea and vomiting, at the present time, these appear useful only in animal studies. The fact that nausea and vomiting can occur independently points to the validity of their assessment as separate behaviors. It is also important to note that their frequency, severity, and duration are separable phenomena. A complete assessment of nausea and vomiting responses should include separate measures of frequency, severity, and duration. Anticipatory nausea and vomiting are clinical phenomena that appear to occur with a reasonable enough frequency so that their measurement would be important in clinical studies.(ABSTRACT TRUNCATED AT 400 WORDS)