Intensity of end-of-life cancer (EOL) care in Medicare Advantage (MA) vs Traditional Medicare (TM).

309 Background: EOL cancer quality metrics are often considered as indicators of overly intensive cancer management and are used to motivate EOL care de-escalation. However, these metrics assume patients received anticancer treatment at some point during their cancer trajectory—many Medicare enrollees with metastatic cancer do not. Moreover, some metrics are upstream and directly influenced by the oncologist, whereas others are downstream and directly influenced by the hospital. Prior studies comparing MA to TM enrollees found MA enrollees were less likely to be hospitalized and more likely to use hospice at EOL. However, studies also find MA enrollees with cancer are less likely to ever receive systemic anticancer therapy (SACT). We evaluated whether differences in the intensity of EOL care in MA vs TM are (1) due to MA enrollees being less likely to receive SACT in general, or less likely to receive SACT at EOL, and (2) in upstream metrics, downstream metrics, or both. Methods: We studied SEER-Medicare enrollees diagnosed with metastatic solid cancers, who survived ≥30 days post-diagnosis, and had ≥12 months of coverage prior to death in 2016-2018. We assessed NQF EOL metrics which we classified as upstream (SACT within 14 days of death, hospice enrollment <3 days before death, >1 ER visits) and downstream (>1 admissions, ≥1 ICU within 30 days of death, inpatient death). Multivariable logistic regression compared EOL outcomes by insurance in two groups: all patients and those receiving SACT in the prior 12 months, adjusting for sociodemographics, comorbidities, and the cancer diagnosis. Results: Of 32,488 TM and 16,563 MA enrollees with metastatic cancer, 56.1% (TM) and 60.5% (MA) received any SACT in the last year of life (OR [CI]: 0.77 [0.73 - 0.80], p<.0001). Among all enrollees (regardless of receiving SACT in the last year), MA enrollees had lower odds of all outcomes: SACT use (0.90 [0.82 - 0.98]), late hospice use (0.92 [0.86 - 0.99]), ER visits (0.88 [0.83 - 0.93]), admissions (0.80 [0.75 - 0.86]), ICU use (0.88 [0.83 - 0.92]), and inpatient death (0.69 [0.65 - 0.74]) (all p<0.05). Among those who received SACT within last year of life, MA enrollees had similar odds of upstream outcomes: SACT use (1.00 [0.91 - 1.10]) (p=0.98), late hospice use (0.96 [0.88 - 1.06]) (p=0.43), ER visits (0.93 [0.86 - 1.00]) (p=0.05). They had lower odds of downstream outcomes: admissions (0.78 [0.72 - 0.86]), ICU use (0.88 [0.82 - 0.94]), and inpatient death (0.71 [0.66 - 0.77]) (all p<0.001). Conclusions: Among all enrollees, MA performed better in EOL metrics than TM. However, these differences largely disappeared when evaluating patients who were receiving SACT and in metrics that oncologists have more influence over. Hence, although MA may decrease intensive EOL acute care use, this does not necessarily reflect improved cancer care quality.