Prior studies have revealed deficiencies in the care provided to patients dying from advanced medical illnesses in acute care hospitals. These deficiencies are best addressed through system change, which may include the development of clinical pathways and quality improvement models. The Palliative Care for Advanced Disease (PCAD) pathway was developed by an interdisciplinary team and includes a carepath, a daily flowsheet, and a physician order sheet with standard orders for symptom control. To evaluate the utility of PCAD, the clinical pathway was introduced on three hospital units (Oncology, Geriatrics, and an inpatient palliative care/hospice unit) as part of a quality improvement initiative and outcomes were compared to two general medical units receiving usual care. A chart audit tool (CAT) was used to review medical records of 101 patients who died on one of these five units during the year prior to implementation (baseline) and 156 who died during the nine months of the PCAD intervention. Four indices from CAT evaluated change over time: the mean number of 1) symptoms assessed, 2) problematic symptoms, 3) interventions consistent with PCAD, and 4) consultations requested. Nine of 27 (33%) patients on the Oncology/Geriatrics units and all 50 patients who died on the palliative care/hospice unit were placed on PCAD. During the PCAD intervention, dying patients who resided on Geriatrics, Oncology and palliative care/hospice units were more likely to have DNR orders than the comparison units, whereas the comparison units were more likely to use "morphine infusions" and cardiopulmonary resuscitation than the units that received the PCAD intervention. The mean number of symptoms assessed increased significantly in all units (P < 0.001 for all comparisons). The number of problematic symptoms identified (P=0.014) and the number of interventions consistent with PCAD increased only on the palliative care/hospice unit (P=0.021). The number of medical consultations declined on all units and reached significance on the Geriatrics and Oncology units (P=0.037). Although these results reflect less than one year of the PCAD intervention and must be considered preliminary, they suggest that 1) a clinical pathway such as PCAD can serve as a managerial and educational tool to improve the care of the imminently dying inpatient; 2) a PCAD clinical pathway can be implemented on hospital units as a quality improvement initiative--a "PCAD intervention;" 3) a PCAD intervention can change outcomes in a positive direction, as measured using a chart audit tool; 4) a PCAD intervention can promote aggressive symptom assessment and treatment when goals of care are aimed at comfort; and 5) changes may occur in units that do not directly receive the intervention, a phenomenon that suggests the possibility of diffusion. Further study of this systems-oriented approach to change is warranted and should include direct assessment of patient and family outcomes, as well as measures of process.