A longitudinal quality of life (QOL) study was performed on patients with advanced melanoma during chemotherapy. The purpose was to describe QOL in this palliative context and to compare the clinical outcome variables and patients' self-assessed QOL. QOL was assessed by the EORTC core questionnaire technique (QLQ-C36), a study-specific melanoma (MM module and the Hospital Anxiety and Depression (HAD) scale. The questionnaires displayed good psychometric qualities and the technique proved to be applicable in this longitudinal study of severely ill patients. Only six patients, out of 95 in total, complied with the full one-year study. Drop-outs occurred early in the course of treatment, most of them due to progressive disease or death. Pretreatment, patients reported a low level of dysfunction and symptom burden but 9 weeks later they exhibited significant deterioration in all QOL measurements, with the exception of pain and emotional functioning. The mean duration of response was short and there was considerable observed treatment-related toxicity. However, no correlation was found between physician-rated clinical outcome variables and QOL measurements, except for neuropathy. Our results are in accordance with earlier data on the supplementary value of QOL measurements to define endpoints in clinical trials.