FRAIL-HRU-AML: Impact of Frailty Assessment on Health Resource Utilization in Acute Myeloid Leukemia Patients: A Population-Based Study from Ontario, Canada

• Published in: Blood 2024-11, Vol.144 (Supplement 1), p.449-449
• Main Authors: Gupta, Gopila, Podolsky, Sho, Liu, Ning, Cheung, Matthew C., Bankar, Aniket
• Format: Article
• Language: English
• ISSN: 0006-4971; 1528-0020
• DOI: 10.1182/blood-2024-209217

Introduction The management of acute myeloid leukemia (AML) involves significant healthcare resource utilization (HRU) due to frequent and prolonged hospitalizations for chemotherapy and supportive care. Frailty, which encompasses overall fitness beyond just comorbidities, is associated with poorer outcomes in various cancers. Assessing frailty can enhance treatment decision-making in oncology by adding valuable context to disease-specific factors. However, the specific impact of frailty on HRU in AML has not been well studied. Therefore, this study aims to evaluate the impact of frailty on HRU in AML patients. Methods This retrospective cohort study from population-based health administrative databases in Ontario, Canada (ICES) included all patients (pts) ≥18 years newly diagnosed (ND) with AML between 2006 and 2021 and treated within 90 days after diagnosis. Patients were followed from date of first chemotherapy to 1- year after maximum follow up until March 31, 2023, for HRU outcomes. Patients were censored at the time of allogenic stem cell transplant (ASCT). The primary predictor, frailty was measured using McIsaac's frailty index (MFI) and categorized as fit (FT), pre-frail (PFR), or frail (FR). HRU outcomes included length of stay for all hospitalizations in days (Total-LOS), intensive care unit stay in days (ICU-LOS), and number of hospital admissions including emergency visits (HA) within first year after starting chemotherapy. These outcomes were measured as per person year (PPY) to adjust for variability in length of follow-up. Association of frailty with HRU outcomes was measured as rate ratios (RR) using multivariable negative binomial models. Model co-variates included age, sex, rurality, neighborhood income quintile, Ontario marginalization (ON-MARG), co- morbidities, ethnicity, prior non-AML malignancy, and treatment intensity {classified as intensive (IT) or non-intensive (NIT) based on standard practices}. Results This study included 5450 pts with ND- AML, with a median age of 65 years (IQR 54-74), 55.8% being males. 3543 (65%) patients received IT and 1907 (35%) received NIT. In entire cohort, 1750 (32.1%) patients were FT, 1874 (34.4%) PFR, and 1826 (33.5%) FR. In 2,035 (37%) patients ≤ 60 years, 44.5% (905) were FT, 36.7% (746) PFR, and 18.9% (384) FR. In 3,415 (63%) patients > 60 years of age, 24.7% (845) patients were FT, 33.0% (1,128) PFR and 42.2% (1,442) FR. 39.0% (683) of FT, 28.7% (537) of PFR, and 18.7% (342) of FR patients und