Early discharge program cost savings for patients with newly diagnosed acute myeloid leukemia after intensive induction chemotherapy

Abstract Introduction: Patients with newly diagnosed acute myeloid leukemia (ndAML) often undergo frontline intensive chemotherapy (IC), which is classically associated with long hospital stays to monitor for bleeding and neutropenic fever. This prolonged inpatient stay for monitoring comes at significant cost to the hospital, in addition to the opportunity cost of preventing inpatient bed use for patients with more emergent needs. Early discharge programs (EDPs) have been explored at select centers without impact on outcomes, but aggregated individual-level costs to hospital (i.e, microcosting) for these EDPs are not known. We sought to analyze major cost drivers for our center's EDP and evaluate the savings based on the patient's absolute neutrophil count (ANC) at discharge. Methods: We aggregated the relative costs of care from the individual components of both inpatient and outpatient clinical care based on cohorts of consecutive patients with ndAML who received intensive induction therapy from 2016-2024: ANC at discharge of ANC ≤0.1 x103 (early), ANC 0.101-0.499 (intermediate), ANC ≥0.5 x103 (traditional). Patients were excluded if they died before any planned discharge from the hospital, or if they received re-induction or salvage therapy before initial discharge. Costs were aggregated for risk exposure periods from date of induction (C1D1) through to start date of consolidation or second-line therapy. Direct costs of care to the hospital were established for 1) each day of antimicrobial therapy, 2) each platelet or red blood cell transfusion, 3) each night on the floor or in the ICU, 4) each visit to the transfusion/infusion clinic, and 5) each outpatient appointment based on hospital data. All costs were normalized to December 2024, with hospital-specific cost-to-charge ratio adjustment. Costs were aggregated in two phases: from C1D1 of induction chemotherapy to first hospital discharge (Phase 1: inpatient care), and from first hospital discharge to start of next therapy, either salvage/re-induction or consolidation (Phase 2: outpatient care). Chi-square analysis and Wilcoxon/Kruskal-Wallis rank-sum tests were used for categorical and continuous variables, respectively. Survival probabilities were estimated using the Kaplan–Meier method, with statistical significance established with a p-value <0.05. Results: In total 185 patients were identified from 2016 to 2024. The median days from C1D1 of IC to start of next therapy was comparable between the early, intermediate, and traditional cohorts (49 [41-58] vs 50 [45-59] vs 49 [43-56.8] days, p=0.37), as was the rate of proceeding to alloSCT (56% vs 60% vs 49%, p=0.49) and median OS (32.2 [12.1-NR] vs 26.1 [19.4-89.0] vs 38.7 [18.0-73.9] months, p=0.58). While these overall clinical metrics were comparable, lower ANC at discharge correlated significantly with more 1) transfusion clinic days (p<0.0001), 2) number of pRBC transfusions (p<0.0001), and 3) number of platelet transfusions (p<0.0001) during Phase 2. Conversely, lower ANC at discharge correlated with fewer days on intravenous antimicrobials (9 [7-14] vs 16 [10-20] vs 18 [9-24], p<0.0001) and fewer days in the hospital (24 [20-30] vs 29 [26-33] vs 28 [25-32], p=0.001), even when accounting for re-admissions before count recovery (n=12/54 early patients, n=2/45 intermediate patients). The median direct-to-hospital cost per patient was $19,473 in the early cohort, $23,095 in the intermediate cohort, and $22,267 in the traditional cohort. When comparing the early cohort to all others, the total cost savings per-median patient discharged early was $2,779 ($19,473 vs $22,252). When comparing the traditional cohort to all others, they incurred $896 more in costs on average ($22,267 vs $21,371). Conclusions: Early discharge from hospital after receipt of IC for ndAML provides modest savings, even with receiving considerable outpatient clinic and transfusion support, and also frees up beds for use by other patients.