In this article we briefly review the experience of the National Cancer Institute of Canada (NCIC) Clinical Trials Group (CTG) with respect to the assessment of patient reported outcomes in clinical trials, and illustrate issues important to assessing symptom palliation in clinical trials of cancer therapy. We highlight a standard approach taken by the NCIC CTG, and illustrate how this approach may be applied to the complex problem of symptom control analysis in patients with locally advanced NSCLC. We further illustrate how variations in this analysis yield different apparent rates of palliation. Apparent rates of palliation critically depended on the outcome measures used: single symptom response across patients (5–32%, depending on the symptom of interest), symptom response in specific symptomatic patients (37–100%), symptom control (45–82%), index symptom response (60%), proportion of patients experiencing improvement in all symptoms (21%), or health-related quality of life (HRQoL) improvement (23%, global). Rates also varied substantively depending on which cohort of patients was considered relevant to each analysis (i.e., was included in the respective denominator). Substantive discordance in patients' apparent palliation was seen when HRQoL data were compared with symptom diary data. Appropriate and valid descriptions of palliative outcomes in clinical trials are complex undertakings. We conclude that several measures are required for a textured clinical description of outcome, and recommend reporting palliation according to individual symptom response rates and HRQoL response rates, in order to address each construct of palliation success.