Impact of hospital admissions through the emergency department versus direct admissions for patients with hematologic malignancies: A single site analysis

Abstract Introduction: Patients with hematologic malignancies often visit the Emergency Department (ED) more frequently than patients with chronic illnesses. They have complex care needs including inpatient and outpatient chemotherapy/immunotherapy, stem cell transplants, and treatment for complications. Due to episodes of high surges some hospitals have high ED to hospital admission times which can lead to delays in care. Our hospital system is able to provide in some situations direct admission (DA) in order to bypass the ED and begin work ups and treatment to prevent care delays. The aim of this study was to assess the benefit of DA as compared to admitting from the ED, in patients with hematologic malignancies. Methods: This retrospective study analyzed data from January 2017 to April 2024, incorporating records from UMass Memorial Medical Center. Patient encounters were separated with an inpatient qualifying diagnosis using ICD-10 codes for a hematologic malignancy, then filtered by known malignant heme attendings to confirm they were admitted to the correct team followed by filtering admission through ED versus directly to the wards. The study evaluated mean hospital length of stay (LOS) compared by heme malignancy subtype, demographic data, neutropenic fever (NF), and also evaluated mortality rate. Statistical analysis was performed by using unmatched student's t-tests. Results: This study included 2319 patient encounters, of which 1022 patients were admitted to the wards through the ED and 1276 were admitted by DA. In terms of comparisons, there were no observed differences by LOS (in days) between ED admission vs. DA (10.37 vs. 10.77, p = 0.41) in the overall cohort. Demographically there were no observed differences by LOS by sex [female: 9.96 (n = 435) vs. 11.14 (n = 485, p = 0.13); male: 10.67 (n = 487) vs. 10.54 (n = 791, p = 0.84), race [African-American: 10.32 (n = 53) vs. 12.35 (n = 57, p = 0.32); Asian: 11.54 (n = 22) vs. 7.77 (n = 41, p = 0.16); White: 10.24 (n = 841) vs. 10.87 (n = 1047, p = 0.25)], or ethnicity [Hispanic/Latino: 10.89 (n = 107) vs. 10.16 (n = 139, p = 0.60); Non-Hispanic/Latino: 10.36 (n = 906) vs. 10.71 (n = 1123, p = 0.51)]. In terms of malignant heme subtypes there were observed differences in LOS between ED admission vs. DA in acute myeloid leukemia (AML): 10.86 (n = 311) vs. 18.74 (n = 234, p<0.001), chronic myeloid leukemia (CML): 7.74 (n = 67) vs. 18.92 (n = 25, p = 0.029), diffuse large b-cell lymphoma (DLBCL): 13.45 (n = 90) vs. 7.3 (n = 280, p<0.001), and multiple myeloma (MM): 8.08 (n = 164) vs. 13.71 (n = 166, p<0.001). There were no observed differences in other subtypes [acute lymphoblastic leukemia: 9.33 (n = 48) vs. 7.19 (n = 91, p =0.17); burkitt lymphoma: 6.55 (n = 29) vs. 5.03 (n = 30, p = 0.33); CNS lymphoma: 11.75 (n = 8) vs. 7.16 (n = 67, p = 0.076); follicular lymphoma: 4.36 (n = 22) vs. 6.31 (n = 32, p = 0.104); hodgkin's lymphoma: 6.71 (n = 7), vs. 8.5 (n = 6, p = 0.67); mantle cell lymphoma: 7.35 (n = 17) vs. 7.71 (n = 82, p = 0.89); myelodysplastic syndrome: 18.15 (n = 13) vs. 29.46 (n = 28, p = 0.15), peripheral T cell lymphoma: 12.13 (n = 15), vs. 6.96 (n = 24, p = 0.229). NF was evaluated due to the close relationship between hospital LOS and mortality. AML was evaluated which showed observed LOS of 10d (n = 153) vs. 27d (n = 27, p<0.001). Mortality rate for in-hospital death was reported as twice-fold for ED admits with 3.33% (34/1022) vs. 1.18% DA (15/1276). As seen in other institutions the largest in-hospital mortality was seen in AML which showed a rate of 4.5% (14/311) for ED vs. 3.8% for DA (9/234). Discussion: This study revealed differences in LOS between an admission through the ED vs. DA to the wards for AML, CML, DLBCL, and MM. We suspect the reason behind a longer LOS for AML DA overall, and with AML patients with NF may be due to planned admissions which involve induction/reinduction chemotherapy, transplants which lead to complicated admissions. Similarly, patients with MM may have planned chemotherapy and/or transplants which involve a longer stay. Conversely patients with DLBCL are admitted for shorter courses of chemotherapy. The difference in overall mortality rate is small, but indicates a need for early specialized triaging with active involvement of the inpatient team throughout the admitting process in patients with heme malignancies. For NF patients we are currently exploring opportunities to shorten time to first dose of antibiotics.