Inpatient palliative care (PC) consultation in hospitalized patients with cancer.

Abstract

317 Background: PC improves quality of life and patient outcomes, yet substantial gaps in delivery persist among hospitalized patients with cancer. We sought to examine associations among receipt of inpatient PC consultation, patient characteristics, and clinical outcomes in hospitalized patients with cancer. Methods: We conducted a retrospective study of adults with cancer hospitalized from 1/2021-12/2024 at a tertiary health system. We abstracted patient demographics and clinical characteristics from the electronic health record. We used regression models to describe associations of inpatient PC receipt with patient factors and clinical outcomes (hospital length of stay [LOS; hazard ratio {HR} for risk of discharge], intensive care unit [ICU] admission, readmissions, overall survival, and discharge location), with death treated as a competing risk when appropriate. Results: Among 11,008 patients (median age=65 years [range: 18-102], 51% female sex, 87% English-speaking, 43% with metastatic disease), most common cancer types were gastrointestinal (GI; 19%), gynecologic (15%), and genitourinary (GU; 14%). For documented race, patients were 77% White, 9% Black, 6% American Indian/Alaska Native, and 2% Asian/Pacific Islander. Overall, 10% (n=1,139) of patients received inpatient PC consultation. Factors associated with increased odds of inpatient PC receipt included Black race (ref: White; OR=1.35, p=.004), “Other” language spoken (ref: English; OR=2.34, p=.003), cancer type (ref: GU; GI: OR=1.91, p<.001; thoracic: OR=2.00, p<.001; CNS: OR=2.05, p<.001; hematologic: OR=2.66, p<.001), and metastatic disease (OR=2.35, p<.001). Patients living in rural areas (OR=0.72, p=.030) had lower odds of receiving inpatient PC compared with those in metropolitan areas. Patients who received inpatient PC had shorter LOS (HR=1.60, p<.001), higher ICU admission risk (HR=1.26, p<.001), greater readmission risk (HR=1.16, p=.001), and worse survival (HR=1.75, p<.001). For discharge location, those who received inpatient PC had higher odds of discharge to hospice (OR=18.52, p<.001), discharge to a facility (OR=2.53, p<.001), and inpatient death (OR=10.75, p<.001) compared with those discharged home. Conclusions: In this large sample of hospitalized patients with cancer, one-tenth of patients received inpatient PC consultation. We found associations of inpatient PC receipt with patient factors and clinical outcomes, including shorter hospital LOS, higher risk of ICU admission, greater readmission risk, and worse survival, which likely relates to the high acuity of patients receiving inpatient PC. Notably, inpatient PC receipt correlated with discharge location, underscoring the importance of PC consultations in assisting with patient disposition and goals of care. Findings from this work should help inform future efforts to optimize PC integration and enhance inpatient cancer care delivery and outcomes.