Background: End-of-life chemotherapy (EOL-C) is associated with a decline in quality of life and increased cost of care. Despite these findings and recommendations of professional and national organizations, EOL-C remains a common practice. Studies have suggested that early referral to palliative care may reduce the likelihood of receiving chemotherapy within the last 30 days of life. We sought to describe the palliative care referral pattern and cancer-directed agent administration at end-of-life among breast cancer patients at a safety-net health system. Methods: Parkland Health (PH) is the safety-net system in Dallas County and is affiliated with the University of Texas Southwestern Simmons Comprehensive Cancer Center. After Institutional Review Board approval, we retrospectively identified patients who died from metastatic breast cancer at PH between 2017 and 2021. Data on demographics, disease and treatment characteristics, and palliative care referral/evaluation were obtained by chart review. EOL-C was defined as planned or administered treatment with cytotoxic agents, immunotherapy, and/or targeted agents (oral or intravenous) within the last 30-days of life. Results: Between 2017 and 2021, 109 patients died from metastatic breast cancer at PH. Median age at diagnosis and death were 52 and 55, respectively. Most patients belonged to racial and ethnic minorities: 41.3% Hispanic, 41.3% Black. Breast cancer was hormone receptor positive/HER2 negative in 53, HER2-positive in 29, and triple negative in 27 patients. Thirty-three patients (30.3%) received or were planned to receive EOL-C: 15 IV (11 administered, 4 planned) and 18 PO. An additional 3 patients received clinic administered non-chemotherapy treatments (leuprolide, fulvestrant, zoledronic acid). Rate of EOL-C was 33.3% (15) in Hispanics and 26.7% (12) among Black patients. Majority of patients were referred to palliative care (79, 72.5%), however, only 46 (42.2%) completed a visit. In the EOL-C group, 84.8% (28) of patients were referred to palliative care and 42.4% (14) completed a visit. Median time between palliative care referral and death was 509 days among patients who completed a visit vs. 53 days in those who did not. Conclusions: Compared to other studies, we observed a higher rate of EOL-C in our patient population. In addition, less than half of the patients in our study completed a palliative care visit before death. Late palliative care referral, timely access, understanding the benefit of the service, low health literacy, language barrier, cultural considerations, and physicians’ perception are likely contributors to pattern of care. As our next step, we plan to study systemic barriers that impact access to and utilization of palliative care service among cancer patients at our safety-net system.