SPECIAL SESSION: Moving From The Means To The Standard Deviations in Symptom Management Research

Description/Overview: Most longitudinal studies of symptoms in patients with chronic medical conditions report means scores and standard deviations to describe changes in symptom occurrence or severity over time. However, most clinicians know that a large amount of inter-individual variability exists in patients' reports of their symptom experiences. For example, in oncology patients receiving chemotherapy, while some patients report very few symptoms, other patients report every conceivable symptom with the highest severity scores. It is important for clinicians to be able to identify these high risk patients in order to target more aggressive symptom management interventions. In order to be able to identify patients are higher risk for a more severe symptom burden, nurse researchers need to use statistical procedures that go beyond the simple reporting of means and standard deviations. Newer approaches to the analysis of longitudinal data, including hierarchical linear modeling and latent class analysis, provide methods to identify patients who are at higher risk for a more severe symptom burden. In addition, the demographic, clinical, and molecular characteristics that are associated with increased risk can be determined. If these risk factors are confirmed in future studies, they can be used to build predictive risk models that will assist clinicians to pre-emptively identify high risk patients. The focus for this presentation is to describe these newer methods of longitudinal data analysis using the symptoms of fatigue and sleep disturbance by oncology patients as the exemplars. Fatigue and sleep disturbance are common symptoms in patients with a variety of a chronic medical conditions. Therefore, using these two symptoms as exemplars will provide information to both clinicians and researchers on the most common phenotypic and molecular characteristics associated with the most severe levels of fatigue and sleep disturbance. As part of this presentation, the purposes for using hierarchical linear modeling and latent class analysis will be compared and contrasted. In addition, approaches for integrating molecular markers into symptom management research will be discussed. This presentation will assist clinicians to perform better assessments of symptoms in patients with chronic conditions. In addition, it should provide essential information to guide the development of future symptom management studies.