Abstract
The treatment setting influences acute myeloid leukemia (AML) outcomes. Most cancer patients receive care in the community, yet few studies have evaluated the effectiveness of clinical programs outside of academic or National Cancer Institute (NCI)-designated cancer centers. This was a multi-level, case-controlled study of real-world outcomes for initial AML treatment in a community-based network for 1,391 patients with AML between 2011 and 2018. We benchmarked survival within our network against the Surveillance, Epidemiology, and End Results (SEER) database. Coarsened exact matching was performed against 17,186 chemotherapy-treated patients in the SEER database. Cox proportional and accelerated failure time multivariable modeling were performed to identify patient, disease, therapy and center characteristics associated with the risk of AML mortality. Within the network, 799 patients were treated at six specialized blood cancer centers and 592 at 63 other hospitals. Patients receiving high-intensity induction at specialized centers had improved median survivals of 31 months versus 18 months [P<0.001] at non-specialized centers. Median survivals were 13 for non-specialized centers versus 10 months for SEER [P<0.001], and 18 for the entire network versus 10 months for SEER [P<0.001]. Multivariable modeling showed significant impacts from age (HR = 1.025), high-intensity induction therapy (HR= .695), diagnosis year (HR= .937), neighborhood income (HR = .997; P<0.01), higher acuity (HR = 1.002) and Charlson comorbidity score (HR = 1.117). AML treatment may be effectively delivered in the community hospital setting, with specialized centers producing better outcomes for higher intensity treatments.