Impact of Fee-for-Service Versus Managed Care Medicare Insurance on the Quality of End-of-Life Care Among Older Adults with Blood Cancers

Introduction: Patients with hematological malignancies often experience suboptimal end-of-life (EOL) care, with significantly lower rates of hospice enrollment and higher rates of intensive healthcare use near death compared to patients with solid malignancies (Earle, JCO 2008). Moreover, patients with blood cancers from minoritized racial and ethnic groups may be more likely to experience suboptimal EOL care. These disparities may be partly driven by modifiable factors such as insurance; however, little is known regarding the potential influence of insurance type on EOL care for this patient population.

Methods: We conducted a retrospective cohort study using the Centers for Medicaid and Medicare Services national claims database. We included patients ≥ 66 years of age who had a blood cancer, survived ≥ 30 days after their diagnosis, died between 2016 and 2020, and were insured by either Medicare fee-for-service (FFS) or Medicare Advantage (MA; managed care). Unlike FFS, MA plans receive fixed payments per enrollee and are not responsible for hospice-related costs following hospice enrollment. We examined established EOL care quality indicators acceptable for blood cancers (Odejide, JCO 2016), including (a) high-intensity healthcare utilization near the EOL (≥ 2 emergency department (ED) visits, ≥ 2 hospital admissions, or any intensive care unit (ICU) admission within the last 30 days of life, chemotherapy use within the last 14 days of life, and in-hospital death), (b) hospice enrollment, and (c) claims for advance care planning (ACP). We conducted univariate and multivariate analyses assessing the impact of FFS vs MA insurance on EOL care. Our multivariate logistic regression models adjusted for age, sex, race/ethnicity, urban/rural status, census-derived household income, comorbidity status, and risk of 90-day mortality (Greenwald, Anesthesiology 2022).

Results: We identified 69,275 eligible blood cancer decedents, of which 67% had FFS insurance and 33% had MA. Of these, 30.2% had lymphoma, 14.9% had myeloma, and 55% had leukemia or myelodysplastic syndromes. About half was female (52%). With respect to race/ethnicity, 81.9% was non-Hispanic White (NHW), 8.9% non-Hispanic Black (NHB), 2.1% Asian/Pacific Islander, and 5.6% Hispanic.

Compared to patients with FFS insurance, those with MA were more likely to be NHB (11.1% vs. 7.8%, p<0.001), Hispanic (8.3% vs. 4.3%, p <0.001), and to have comorbidity index of ≥ 4 (43.9% vs 41.8%, p<0.001).

Of the entire cohort, 15.3% had ≥ 2 ED visits, 15.9% had ≥ 2 hospitalizations, and 29.4% had ≥ 1 ICU admission in the last 30 days of life; 2.0% received chemotherapy in the last 14 days of life; and 22.9% died in the hospital. The rate of hospice use was 54.8%, with 61.5% enrolling for > 7 days. Only 14.5% of the cohort had ACP claims.

Multivariate regression analysis revealed that patients with MA insurance were significantly less likely to have either ≥2 ED visits (Odds Ratio [OR]: 0.80, 95% Confidence Interval [CI]: 0.76 to 0.84) or ≥1 ICU admissions (OR: 0.83, 95% CI: 0.80 to 0.86) within the last 30 days of life, and had a lower likelihood of in-hospital death (OR: 0.74, 95% CI: 0.71 to 0.77) compared to patients with FFS. There were no significant differences in hospital admissions or chemotherapy use near death by insurance type. MA was significantly associated with a higher likelihood of hospice enrollment (OR 1.11, 95% CI 1.08 to 1.15) and lower odds of short hospice stays (≤ 7 days) compared to FFS insurance (OR 0.94; 95% CI 0.90 to 0.98). Finally, MA patients were less likely to have claims for ACP (OR 0.84, 95% CI: 0.80 to 0.88) compared to FFS patients.

Conclusions: In this large national cohort of blood cancer decedents, MA insurance was associated with lower rates of high-intensity healthcare utilization near death, and higher rates of timely hospice use compared to Medicare FFS, suggesting potentially higher quality EOL care for the former. Interestingly, patients with MA were more likely to be people of color. Our paradoxical finding of lower odds of ACP claims among MA enrollees despite other evidence of high-quality EOL care may reflect underutilization of ACP claims even when ACP discussions occurred. Given that our analysis suggests that insurance type may impact the quality of EOL care, future work characterizing which elements of MA promote high-quality EOL care may help to improve equitable access to high-quality EOL care for all patients with blood cancers.