Variation in physician recommendations, knowledge and perceived roles regarding provision of end-of-life care

Background: There is high variability in end-of-life (EOL) treatments. Some of this could be due to differences in physician treatment recommendations, their knowledge/attitude regarding palliative care, and their perceived roles in treating patients with advanced serious illness (ASI). Thus, the objective of this paper was to identify potential variation in physician recommendations, their knowledge/attitude regarding palliative care and perceived roles in treating ASI patients. Methods: A cross-sectional survey consisting of vignettes describing patient characteristics that varied by age, expected survival, cognitive status and treatment costs and asked physicians whether they would recommend life-extending treatments for each scenario, was administered to 285 physicians who treat ASI patients in Singapore. Physicians were also assessed on their knowledge/attitude in palliative care. They were administered a best-worst scaling exercise requiring them to select their most and least important role as a physician caring for an ASI patient. Results: There was a wide variation in physician recommendations for life-extending treatments for patients with similar profiles, which can partly be attributed to physician characteristics (years of experience and place of training). Only about one-fourth of the physicians answered all knowledge/attitude questions correctly. Statements assessing knowledge/attitude regarding pain management had the fewest correct responses. The most important perceived role regarding provision of EOL care concerned symptom management. Conclusions: Results suggest that variation in physician treatment recommendations may be partly related to their own characteristics, raising concerns regarding the EOL care being provided to patients. Efforts should be made to better understand this variation and to provide the physicians with additional training in key aspects of palliative care management.