PURPOSE:  This workshop will introduce participants to path modeling and its ability to examine direct and indirect effects of symptoms of cancer and its treatment on health-related quality of life (HRQL) outcomes. Workshop participants will become familiar with the concepts of direct, indirect, and total effects and the importance of conceptualizing and analyzing multidimensional phenomena appropriately. DESCRIPTION:  It’s not uncommon to find little or no relationship between symptoms or their change and selected HRQL outcomes when analyzing oncology trial data. One reason, we argue, is that the relationships among these variables are more complex than the analyses would imply or can evaluate. That is, the hypothesis that symptoms affect overall quality of life (QoL) may not be supported (or only marginally so) because the analyses assume a direct effect of symptoms on QoL. However, thinking through the logic among these variables suggests a “causal cascade,” whereby intervening variables mediate the causal pathway between symptoms and QoL or other dimensions of HRQL. Thus, as will be demonstrated, symptoms, such as pain or appetite loss, affect fatigue, which affects QoL. Path analyses are needed to evaluate these more complex causal relationships and obtain total effects of symptoms on the various domains of HRQL. The result is a more accurate understanding of treatment effects and how that affects HRQL. Dr. Boye will discuss the implications this conceptual and analytic approach has for refining our understanding of efficacy of oncology treatments. Dr. Stull will provide background on path analysis, and Ms. Ainsworth will give an empirical example from two trials of treatment for lung cancer. The audience will be encouraged to discuss and explore their own examples as they relate to endpoint models and conceptual frameworks and how these methods may help shed light on null or marginal results they have obtained in trials.