BACKGROUND: The purpose of this study is to examine the concordance of symptom assessment among the multiple raters in French-speaking elderly patients with an advanced cancer benefiting from palliative care. PATIENTS AND METHODS: This study was conducted in a geriatric hospital with palliative care specificity. During 6 months, patient, nurse and physician completed the Edmonton symptom assessment system on two consecutive days. RESULTS: 42 patients with an advanced oncological disease were included. Mean age was 72+/-9.04 (range 52-88) and 23 were females. Mean mini mental status examination (MMSE) was 27.5+/-1.6. First assessment was completed at a median of day 8 after admission. Nurses, physicians and patients assessments were reproducible between days 1 and 2 (P>0.05). Pearson correlation coefficient significantly associated nurse assessment with patient assessment for pain, depression, anxiety, drowsiness, appetite and wellbeing (P<0.05). Physician assessment was associated with patient assessment for pain, depression, drowsiness, appetite, wellbeing and shortness of breath (P<0.05). However, regression analysis looking for patient score from both physicians and nurses scores weakly correlated all these factors (R2<0.6), except for appetite (R2 for day 1/day 2: 0.79/0.64). CONCLUSIONS: French-speaking elderly cancer patients without cognitive failure and in stable general condition are consistent in their symptom assessment, and they have to be considered as the gold standard. Nevertheless, interdisciplinary assessment is probably a valid surrogate to self-assessment by the patient but only when the latter is truly impossible.