A retrospective cohort studying integrative palliative care in a community oncology practice: Impact and outcomes.

Abstract

12047 Background: Embedding palliative care within oncology practice supports symptom management and shared decision-making; however, its impact on real-world healthcare utilization and end-of-life planning is less well characterized. In 2018, a dedicated palliative oncology service was implemented at a Midwestern community cancer institute. We evaluated whether this integrated model was associated with differences in healthcare utilization, hospice referral, and advance care planning (ACP) compared with standard oncology care prior to implementation. Methods: We conducted an IRB-approved retrospective cohort study of adults with cancer treated at a single institution. Two age-, stage-, and histology-matched cohorts were identified: patients treated after implementation of the embedded palliative oncology service (2018–2021; Cohort A, n=300) and patients treated prior to implementation (2012–2017; Cohort B, n=300). Clinical data were extracted from the electronic medical record. Outcomes included emergency department (ED) visits, inpatient and ICU admissions, inpatient and ICU length of stay, hospice referral, and documented ACP. Continuous variables were compared using two-sample t-tests and categorical variables using χ² tests (α=0.05). Post-hoc power analyses were performed for outcomes demonstrating between-group differences. Results: Both cohorts were predominantly White (88% in Cohort A vs 93% in Cohort B) with a slight difference in mean age (70.8 vs 75.0 years). Cohort A included a higher proportion of patients with stage III–IV disease (74% vs 53%, p≤0.05). Compared with standard oncology care, the integrated palliative oncology model was associated with a higher mean number of ED visits (4.1 vs 3.0, p<0.05). There were no significant differences in inpatient admissions, ICU admissions, or inpatient length of stay between groups. Hospice referral and ACP documentation occurred significantly more frequently in Cohort A (both p<0.05). Post-hoc analyses demonstrated adequate power (>80%) to detect observed differences in ED utilization, hospice referral, and ACP documentation. Conclusions: Implementation of an integrated palliative oncology service was associated with increased ED utilization alongside substantially higher rates of hospice referral and ACP documentation, without an accompanying increase in hospital or ICU admissions or length of stay. These findings suggest that embedded palliative oncology may promote earlier end-of-life planning while maintaining stable inpatient utilization. Prospective studies are needed to better understand drivers of ED use and to further refine integrated palliative oncology care models.