Polypharmacy still is a problem in the large majority of patients with terminal cancer. Further studies should be focused on the patients' quality of life, drug interactions, and adverse events.

Study-based guidelines on thromboprophylaxis are not available for palliative care patients. The authors asked a panel of academic medical experts in palliative care, oncology, blood coagulation, and intensive care to select a prophylactic regimen out of 5 predefined options for a virtual patient with advanced bronchial cancer in different clinical settings. Primary prophylaxis for venous thromboembolism was withdrawn by all physicians when the patient had a Karnovsky's index of 10 and was described as dying. It was given by 25% of physicians when the patient had a Karnovsky's index of 20 and by 85% when Karnovsky's index 40 was still 40. Similar results were obtained in the situation of secondary prophylaxis of venous thromboembolism and when the patient was described as having a history of chronic atrial fibrillation. This data clearly show that thromboprophylaxis is delivered according to a compound estimate of risks and benefits of such prophylaxis in a specific palliative care situation.

Our data reveal a scarce demand for AD in our population of hospitalized cancer patients. Patients who wanted to issue an AD had a high HADS-D, which is associated with a low performance status.

Our data reveal that about half of all cancer patients in palliative care units are treated with thromboprophylaxis. Low performance status was the most frequent contraindication for thromboprophylaxis.

Oncologists differ widely in their attitudes towards palliative care and services. These attitudes depend on a number of individual and society-based variables. It is recommended that palliative care be started early in the disease trajectory of patients with a life-threatening disease but in Austria we lack data on oncologists' adherence to this recommendation. We surveyed 785 oncologists in Austria by presenting the clinical course of a hypothetical patient with primary metastatic breast cancer from diagnosis until death. The majority of oncologists would involve palliative care services when the patient's Karnofsky index (KI) was < 50, and hospice services when the KI was < 40. Special training in palliative care was significantly associated with early use of hospice services. Reasons for not involving palliative care and hospice services earlier than indicated were systematically evaluated and included, among others, "fear of destroying the patient's hopes" (36% of respondents with regard to palliative services, 57% with regard to hospices). Overall, 67% of the oncologists would inform the patient about the malignant nature of her disease and the anticipated limitation of her life expectancy at the time of diagnosis. Issuing an advance directive would be discussed by only 25% at that time. Our data show that oncologists involve palliative care services at an advanced stage of disease in patients with primary metastatic cancer and that information about malignancy and the incurable nature of the disease is not uniformly provided at the time of diagnosis.