Predictors of access to palliative care services among patients who died at a comprehensive cancer center

Purpose: Palliative care services can decrease physical and psychosocial distress in patients with advanced cancer. However, most patients with cancer die without access to palliative care services (APCS), and patterns of referral are not well understood. The purpose of this study was to determine predictors of patients' access to palliative care. Patients and Methods: We reviewed patient records from the computerized database at UT M. D. Anderson Cancer Center over 2 (2003 and 2004) to determine differences in characteristics and outcomes between patients with and without APCS. APCS was defined as a palliative care consultation and follow-up or transfer to the palliative care unit. Results: A total of 499 of 1453 (34%) inpatients who died at our cancer center had APCS. There were no significant differences in race, age, or insurance status between the APCS groups. The two major predictors of a low rate of APCS were hematologic malignancies (20% rate of APCS versus 44% for solid tumors, p < 0.0001) and intensive care unit (ICU) admission (15% versus 52% for non-ICU admission, p < 0.0001). Patients with hematologic malignancies who were admitted to the ICU had the lowest APCS rate (10%, p < 0.0001). The median relative cost of care per patient for decedents with APCS was 0.62 compared to nonreferred patients (p < 0.0001). Conclusion: APCS was lower among patients with hematologic malignancies and those admitted to the ICU. APCS resulted in a lower cost of care. Mortality in comprehensive cancer centers is quite variable among different primary malignancies. More research is needed to better define patterns of referral.