Travel distance to treatment center does not independently impact survival in Acute Myeloid Leukemia (AML): A national cancer database (NCDB) analysis

Abstract

Abstract Introduction AML is an aggressive hematologic malignancy that requires frequent hospitalizations, close monitoring, and complex supportive care. Studies utilizing the NCDB have somewhat counterintuitively observed improved survival among patients with certain solid cancers living further away from hospitals. Comparable analyses on AML patients from countries with single-payer healthcare systems have found that distance does not impact access to care, choice of therapy, or survival. How distance to center impacts overall survival (OS) in patients with AML in the United States is unexplored. Methods Demographic, diagnostic and clinical data on patients diagnosed with AML from 2004-2019 were obtained from the NCDB. Cases where diagnostic, treatment and location data were unavailable were excluded. Crowfly distance was stratified by quartile and examined as a categorical variable. The primary outcome was OS in months. Patients without reported death dates were censored at time of last follow-up. Statistical analysis included Cox proportional hazard regression and Kaplan-Meier analysis. All multivariable models were adjusted for demographics (age, sex, race, ethnicity), socioeconomic status (median income, insurance status), geographic location of care center (urban/rural, US Census region), and clinical features (diagnosis year, Charlson Deyo comorbidity index, APL status, Core-binding factor status, chemotherapy, transplant status, multiagent therapy, palliative care). Results A total of 131,758 patients were identified. Adults aged 18-59 accounted for 33.8% of the cohort and 45.5% patients were female. The median crowfly distance was 12.9 mi, with an interquartile range (IQR) of 5.3 to 35.9 mi. Among cases reporting the type of treatment facility (N=118,606), 48.8% were treated at academic centers. In a univariable Cox regression analysis of quartiled crowfly distance using the first quartile (shortest distance) as the reference, greater distances were significantly associated with better OS (P<0.001 for all 3 quartiles). After multivariable adjustment, only the second and third crowfly quartiles demonstrated a significantly lower hazard ratio (HR): Quartile 2 HR 0.97 (95% CI 0.96 - 0.99, P = 0.005), Quartile 3 HR 0.96 (95% CI 0.95 - 0.98 P<0.001). Quartile 4 did not significantly differ from Quartile 1 (HR 1.00, 95% CI 0.97 - 1.02, P = 0.717). A subgroup analysis of only academic centers (N = 50,824) reflectedthis pattern but was not significant: Quartile 2 HR 0.99 (95% CI 0.96 - 1.0), Quartile 3 HR 0.98 (95% CI 0.95 - 1.0), Quartile 4 HR 0.97 (95% CI 0.94 - 1.0). The median elapsed time between diagnosis and initiation of treatment (5.0 days) did not vary with statistical significance across crowfly quartiles overall (P = 0.192), nor among academic centers (5.0 days, p = 0.758). Stratified analysis of median income ranges examined the relationship between SES and crowfly travel quartiles as it impacts survival outcomes. With distances held constant, there was no evidence that patients of the lowest median income tier (<$30,000) experienced worse survival: Quartile 2 HR 0.99 (95% CI 0.93 - 1.06), Quartile 3 HR 1.04 (95% CI 0.98 - 1.11), Quartile 4 HR 0.93 (95% CI 0.89 - 0.97). Conversely, a survival benefit was seen among patients of higher SES at the furthest distances from treatment centers (Quartile 4): median income of $35,000-$45,999 had a HR 0.91 (95% CI 0.87 - 0.95); median income of $46,000+ had a HR 0.82 (95% CI 0.78 - 0.86), when compared to median income <$30,000. Multivariable Cox model interaction analysis showed this interaction was statistically significant (p < 0.001). Conclusion Our NCDB analysis suggests that greater distance to treatment center was not associated with worse survival in AML. Our assessment of SES as a risk factor by subgroup stratification showed that patients with lower SES did not experience worse outcomes despite longer travel distances. Notably, higher-income patients living farther from care had better survival, suggesting that belonging to higher SES may help overcome any challenges related to travel distances. Our findings raise the question of whether efforts to minimize the burden of travel throughout the course of AML treatment could help patients overcome other known barriers to care, including socioeconomic, institutional, and geographic factors in such a way that prolongs survival.