Evaluating depth of disparities in relation to relative survival rates: Analysis of SEER data.

Abstract

6602 Background: Prior studies have identified widening survival disparities between Black vs White patients (pts) in solid tumors with higher 5-year relative survival rates (RSR), a surrogate for higher amenability to treatment (Tehranifar et al., 2009). Given major advancements in treatment for hematologic malignancies leading to higher RSR, we sought to evaluate whether disparities worsened with higher RSR, especially among Black pts compared to White pts, in the modern era of therapy. We hypothesized that racial disparities would be larger in malignancies with higher amenability, as measured by 5-year RSR. Methods: We identified first hematologic malignancies among pts ages ≥18 years diagnosed between 2009-2019 in a collection of 17 Surveillance, Epidemiology and End Results (SEER) Program registries. We treated RSR as a continuous variable. Cox proportional-hazards model was used to correlate RSR and race with overall survival (OS) controlling for other pt characteristics. The interaction between RSR and race was included in the model to test if the racial effect on OS differed at various RSR levels. Results: We identified 363,340 pts who met inclusion criteria. Median age was 66 years and 55% were male. 82% were White and 10% Black. With a median follow-up of 48 months (mos), median RSR was 68.2% and median OS was 101 mos (95% CI 100-102 mos). Black pts had a lower median RSR than White pts, 63.8% vs 68.9%, and lower median OS, 90 mos (95% CI 87-93 mos) vs 99 mos (95% CI 98-100 mos). On univariable analysis, all variables were statistically significant given the large sample size. On multivariable analysis (MVA), Black pts had a higher risk of death than White pts. The interaction between 5-year RSR and race was highly significant (p=<0.01, not shown in Table), as indicated by slightly increasing hazard ratios (HRs) reflecting Black-White survival differences with increasing RSR level (i.e., as RSR increased, Black pts had slightly worse OS than White pts) (Table). Conclusions: We hypothesized that with increasing RSR, Black-White disparities might increase, as all effective therapies which contribute to improved RSR might not be equitably available to all populations. We observed that statistically significant disparities were present across RSR levels and increased only minimally for those with hematologic malignancies with higher RSR. In conclusion, disparities are prevalent in hematologic malignancies despite the degree of amenability to treatment. [Table: see text]