HSR19-087: Intensity of Symptom Management After Palliative Care Referral Not Associated With Reduced Imaging Intensity for Cancer Patients During the End-of-Life Period

Abstract

Purpose: For patients with advanced cancer, palliative care (PC) referral itself results in improved quality of life and lower total health care costs. While our earlier research suggests that PC referral alone does not appear to affect imaging utilization, the effect of the intensity of services for PC symptom management on utilization of imaging services is unknown. We sought to evaluate whether an association exists between the intensity of PC symptom management services and utilization of high-cost imaging at the end-of-life. Methods/Materials: Adult cancer patients who died between January 1, 2012–May 31, 2015 were identified. A PC symptom management intensity score was determined for each patient during the last year of life based upon the number of inpatient and outpatient PC service visits and the number of problems addressed within the 8 defined domains for palliative care. Frequency of utilization of nonemergent oncologic imaging was determined for the last 3 months and the last month of life. Using chi-square and Wilcoxon-rank sum tests, effects of low (<20th percentile) and high (>80th percentile) PC symptom management intensity score on the proportion of patients imaged and mean number of studies per patient (mean imaging intensity [MII]) were compared for the last 3 months and last month of life. Results: 3,772 decedents were included, with 3,523 (93.4%) never referred to PC and 249 (6.6%) referred to PC, largely before the last month of life (70.3%). Within the last 3 months and last month of life, the proportion of patients imaged was significantly greater (P<.001) in patients with high PC symptom management intensity scores when compared to those with low PC symptom management intensity scores. There was no significant difference in the mean number of imaging studies received by either group in the last 1 month or 3 months of life. (P>.05). Conclusions: The intensity of PC symptom management in end-of-life oncologic care was not associated with decreased use of nonemergent, high-cost cancer imaging in our cohort. This may suggest an opportunity for reduced high-cost imaging through further clarification of the goals of care with both patients and oncologists, although the role of imaging in the setting of PC during the end-of-life period warrants further study as well.