End-of-Life Systemic Oncologic Treatment in the Immunotherapy Era: The Role of Race, Insurance, and Practice Setting.

Receipt of antineoplastic systemic treatment near end of life (EOL) has been shown to harm patient and caregiver experience, increase hospitalizations, intensive care unit and emergency department use, and drive-up costs; yet, these rates have not declined. To understand factors contributing to use of antineoplastic EOL systemic treatment, we explored its association with practice- and patient-level factors. We included patients from a real-world electronic health record-derived deidentified database who received systemic therapy for advanced or metastatic cancer diagnosed starting in 2011 and died within 4 years between 2015 and 2019. We assessed use of EOL systemic treatment at 30 and 14 days before death. We divided treatments into three subcategories: chemotherapy alone, chemotherapy and immunotherapy in combination, and immunotherapy (with/without targeted therapy), and estimated conditional odds ratios (ORs) and 95% CIs for patient and practice factors using multivariable mixed-level logistic regression. Among 57,791 patients from 150 practices, 19,837 received systemic treatment within 30 days of death. We observed 36.6% of White patients, 32.7% of Black patients, 43.3% of commercially insured patients, and 37.0% of Medicaid patients received EOL systemic treatment. White patients and those with commercial insurance were more likely to receive EOL systemic treatment than Black patients or those with Medicaid. Treatment at community practices was associated with higher odds of receiving 30-day systemic EOL treatment than treatment at academic centers (adjusted OR, 1.51). We observed large variations in EOL systemic treatment rates across practices. In a large real-world population, EOL systemic treatment rates were related to patient race, insurance type, and practice setting. Future work should examine factors that contribute to this usage pattern and its impact on downstream care. [Media: see text].