LAUSR

PURPOSE Early palliative care (PC) improves outcomes for outpatients with advanced cancer. Its effect on hospitalized patients with cancer is unknown. Herein, we report on the influence of a novel, fully integrated inpatient medical oncology and PC partnership at a tertiary medical center during its first year of implementation. METHODS We conducted a retrospective, longitudinal, pre- and postintervention cohort study at Duke University Hospital. Pre- and postintervention cohorts were defined as all patients admitted to the solid tumor inpatient service from September 1, 2009, to June 30, 2010, and September 1, 2011 to June 30, 2012, respectively. We extracted patient data, including demographics, cancer diagnosis, disease status, length of stay, intensive care unit transfer rate, discharge disposition, time to emergency department return, time to readmission, and 7- and 30-day emergency department return and readmission rates. Nursing and physician surveys assessed satisfaction. Descriptive statistics, and Kruskal-Wallis and Χ2 tests were used to describe and compare cohorts. A generalized estimating equation accounted for repeated measures. RESULTS Pre- and postintervention analysis cohorts included 731 and 783 patients, respectively, representing a total of 1,514 patients and 2,353 encounters. Cohorts were similar in baseline characteristics. Statistically significant lower odds in 7-day readmission rates were observed in the postintervention cohort (adjusted odds ratio, 0.76; 95% CI, 0.58 to 1.00; P = .0482). Patients in the postintervention group had a decrease in mean length of stay (-0.30 days; 95% CI, -0.62 to 0.02); P = .0651). We observed a trend for increasing hospice referrals ( P = .0837) and a 15% decrease in intensive care unit transfers ( P = .61). Physicians and nurses universally favored the model. CONCLUSION A fully integrated inpatient partnership between PC and medical oncology is associated with significant and clinically meaningful improvements in key health system-related outcomes and indicators of quality cancer care.