Serious illness communication practices in glioblastoma care at an academic medical center.

28 Background: Glioblastoma (GBM) is an aggressive primary brain tumor with a disease course marked by both cognitive and physical decline. Early conversations about end-of-life (EOL) preferences are important given the average life expectancy of 14 months. Little is known about patterns of communication in GBM. Objective: To describe current practices surrounding EOL conversations in GBM patients at an academic medical center. Methods: We reviewed the medical records of 240 patients treated for GBM at the University of Pennsylvania who died between 2017 and 2019. We systematically identified and characterized EOL conversations, as defined by documented discussions of prognosis, goals of care, EOL planning and/or code status. Results: Patients in this study were predominantly male, white, privately insured and had relatively high performance status (median KPS 80 at initiation of therapy). Median survival was 487 days and 96% of patients received both chemotherapy and radiation. Nearly all patients (96%) had at least one EOL conversation documented. The median number of documented conversations per patient was 4. The first EOL conversation occurred at a median of 24 days after diagnosis (range 0 to 3,883) and 362 days before death (0 to 2,401). The last EOL conversation occurred at a median of 32 days before death (0 to 1,514). EOL conversations took place earlier among older patients, with each additional year of age at diagnosis associated with a first conversation 8 days earlier (95% CI: -12 to -4 days; p<0.01). At least one EOL conversation occurred at an outpatient visit for 89% of patients; 33% had conversations during hospital admissions; and 38% had EOL conversations via telephone. Medical oncology was the most common specialty to document EOL conversations (89% of patients), followed by palliative care (19%), radiation oncology (13%) and neurosurgery (10%). Conclusions: EOL conversations appear to have occurred relatively early and often, but further work is needed to evaluate the quality of documented discussions and to determine whether this pattern is unique to this single center or widespread. The variety in provider specialty and conversation setting reflects the complexity of the care environment in GBM.