A multicenter study involving six palliative care units in Italy was carried out on 159 terminal cancer patients seen at home or in hospital. The physician or the nurse completed independently from the patient the Therapy Impact Questionnaire (TIQ), a questionnaire devised for quality-of-life evaluation in terminal cancer patients. The patient's assessment was used as the valid reference measurement to compare with the health-care workers' evaluation to assess the validity of the latter. The results showed that percentages of agreement were higher for physical than for psychological and cognitive symptoms, and that there was a greater agreement on the absence rather than on the presence of a problem. None of the characteristics of the patient nor of the proxy showed any statistically significant relationship with the two disagreement indexes. The results suggest that caution is needed in the use of health-care workers as alternative sources of information regarding patients' quality of life.