When we study symptoms over time, we have a number of goals. One goal may be to characterize the average trajectory of symptom severity over time. Often we are asking questions such as: Do symptoms change over time? If so, what factors (eg, treatment) modify that trajectory? We also may be interested in characterizing the impact of symptoms and answering questions such as: Which symptoms have the greatest impact on subjects? How do changes in symptoms affect patient reports of health status and quality of life? How does the level of severity influence that impact? Finally, the goals of the investigation may be to characterize between-subject and within-subject variation, to answer questions such as: Are there groups of symptoms that change within individuals in a similar manner? Are there biological factors (eg, cytokines) that change in conjunction with the development of symptoms? For example, we might be interested in the relationship of inflammatory cytokines to fatigue, pain, and other symptoms of cancer and cancer treatment.

These questions differ from traditional analysis, in which we would typically test whether the average change in symptoms is a function of biological factors by examining the mean scores for groups of subjects with high versus low values of the biological factors. When our interest becomes focused on change and variation at the level of the individual rather than the group, we might be interested in questions such as: Do individuals with generally higher levels of interleukin (IL)-6, an inflammatory cytokine strongly associated with lung cancer, report more pain?