End-of-life care in a voluntary hospitalist model: effects on communication, processes of care, and patient symptoms.

PURPOSE: To assess the effects of hospitalist care on communication, care patterns, and outcomes of dying patients.
METHODS: We examined the charts of 148 patients who had died at a community-based, urban teaching hospital, comparing the end-of-life care provided by community physicians and hospitalists.
RESULTS: Patients of hospitalists and community-based physicians were similar in age, race, severity of acute illness, and difficulties with activities of daily living. After admission, hospitalists had discussions with patients or their families regarding care more often than did community physicians (91% [67/74] vs. 73% [54/74], P = 0.006) and were more likely to document these discussions themselves. Among patients who were "full code" at admission, there was a trend towards patients of hospitalists receiving comfort care more frequently at the time of death (50% [25/48] vs. 37% [15/40], P = 0.14). Although there were no differences in the use of medications such as long-acting opioids, no symptoms in the 48 hours prior to death were more likely to be noted for patients of hospitalists (47% [n = 35] vs. 31% [n = 23]), P = 0.03). After adjustment for confounding factors in multivariable models, only findings regarding documentation of discussions and symptoms remained statistically significant.
CONCLUSION: Hospitalists at a community-based teaching hospital documented substantial efforts to communicate with dying patients and their families, which may have resulted in improved end-of-life care.