Objectives/Hypothesis: Sinonasal malignancies are a rare and biologically diverse group of tumors which comprise less than 5% of all head and neck cancers. Differences in patient demographics, health insurance status, and access to healthcare resources are known to affect survival across the spectrum of malignancies, but adequately powered studies historically rely on national database queries due to the rarity of the lesions. Understanding the availability and impact of community resources on a patient's diagnosis is thus quite difficult, especially in large, complex urban/metropolitan cities where socioeconomic status, life expectancy, and income can vary widely from street to street, even within the same zip code. Our objective was to analyze the association of a patient's residence, based on US Census Bureau data, with survival, tumor characteristics, health insurance status, race, gender, and distance to the treating medical center.
Study Design: Retrospective cohort study of patients at a single institution
Methods: Patients diagnosed and treated for a primary sinonasal malignancy between 2009 and 2019 at a single institution were identified and examined based on ICD-9 and ICD-10 diagnosis codes. Demographics, clinical presentation, health insurance, smoking & alcohol history, histopathology, location, staging at diagnosis, date of expiration, and recurrence were tabulated. Patients were assigned to specific United States Census tract codes based on their primary address during treatment. ANOVA statistical analysis was performed using Revman 5.4.1.
Results: After applying exclusion criteria, 50 patients were included for subsequent analysis. The cohort of patients were predominantly Caucasian, male, were covered by public health insurance, and lived 10-24.99 miles away from the treating medical center. Most of the analyzed patients presented with stage IV disease. Medicare subscribers, patients living more than 100 miles away from the institution, and whose who lived in US Census tracts with an average gross family income of less than $25,000 tended to present with higher staged disease that approached but did not reach significance (p = 0.23, 0.24, and 0.29 respectively). Females, Medicare subscribers, and patients with an income < $25,000 experienced worse 5-year-survival that also approached but did not reach statistical significance. Subgroup analysis based on specific histologic subtype did not yield any meaningful associations.
Conclusion: Our study serves as a proof of concept in the utilization of US Census Bureau data to understand how a community affects the incidence and treatment of a disease subtype. Utilizing US Census tracts may allow for a more detailed analysis of patient location, essentially subdividing zip codes into detailed parcels, and provide more community level information than is typically available in other large federal databases. The lack of a significant association between US Census tract location with survival and AJCC staging in our sample may be explained by the late presentation and aggressive nature of the disease in the patients we analyzed, as well as the expected behavior of many histologic subtypes of sinonasal malignancy. We anticipate this strategy may be used in the future to better allocate healthcare resources and screenings to locations with a higher degree of accuracy regardless of the disease entity in question.