Terminal admission phenotype in acute myeloid leukemia: A national analysis of inpatient care intensity (NIS 2016–2023).

Abstract

e23207 Background: Inpatient mortality among patients with acute myeloid leukemia (AML) remains substantial; however, terminal hospitalizations are poorly characterized in administrative data. Inpatient death alone does not distinguish expected end-of-life trajectories from potentially avoidable, high-intensity care. This study aims to define and evaluate a terminal hospitalization phenotype in AML based on inpatient death with markers of high-intensity, life-sustaining care. Methods: A retrospective, survey-weighted analysis of the National Inpatient Sample (2016–2023) was performed. Adult AML hospitalizations were identified using ICD-10-CM C92.0*. Terminal admission was defined as inpatient death with ICU-level care proxies, including invasive mechanical ventilation, dialysis, or shock. Primary outcome was terminal admission prevalence. Secondary outcomes included care intensity, length of stay (LOS), and costs. Multivariable survey-weighted logistic regression evaluated factors associated with terminal admissions. Results: Among 484,550 national AML hospitalizations, 8.7% resulted in in-hospital death. Notably, 43.7% of these deaths occurred in the setting in the ICU level of life-sustaining care, defining a terminal admission phenotype. Although terminal admissions accounted for only 3.8% of all AML hospitalizations, they represented a disproportionate concentration of care intensity. Across all AML hospitalizations, mechanical ventilation occurred in 4.9%, dialysis in 2.1%, and shock in 2.0%. Among the high-risk subgroup of patients who died, care intensity was markedly concentrated: 37.9% received mechanical ventilation, 9.7% dialysis, and 10.9% experienced shock during the terminal hospitalization. Mean LOS was 11.7 days overall (95% CI 11.5–11.9) and 13.4 days among deaths (95% CI 12.9–13.9). Mean hospitalization cost was $47,114 overall (95% CI $44,611–$49,618) and $73,562 among deaths (95% CI $67,115–$80,009). Mean charges among deaths reached $255,771 (95% CI $239,769–$271,773). After adjustment, terminal admissions were independently associated with weekend admission (adjusted odds ratio [aOR] 1.34, 95% CI 1.22–1.46) and urban teaching hospitals (aOR 2.20, 95% CI 1.65–2.93). Increasing age (aOR 1.01 per year, 95% CI 1.00–1.01) and male sex were also associated with higher odds of terminal admission. Conclusions: Over two in five inpatient AML deaths occur in the setting of high-intensity life-sustaining care, defining a distinct terminal phenotype requiring substantial resources. This nationally representative analysis moves beyond mortality alone to identify a reproducible marker of end-of-life care intensity in AML. The terminal admission phenotype provides a scalable framework for evaluating care delivery patterns, hospital variation, and opportunities to better align inpatient AML care with prognosis and goals.