Abstract A124: Neighborhood, race and insurance predict for hospital admission during radiation therapy

Background: Hospital admission during radiotherapy (ADRT) is associated with increased cost, interrupted treatment, and inferior outcomes. The purpose of this study was to benchmark patient ADRT rates, define socioeconomic predictors for ADRT, and geographically map ADRT rates on the neighborhood level across a large Mid-Southern catchment region served by a single academic cancer referral center. Methods: Demographic, clinical and treatment information were collected for all patients treated with radiation therapy (RT) at our center from January 1, 2015 to December 31, 2017. Occurrence of ADRT included inpatient and emergency room admissions. ADRT was categorized as causing “minor interruption” if ADRT was associated with postponement in 1-4 RT treatments. “Major interruption” was defined as postponement in 5 or more treatments. Patients with Medicaid or no insurance were categorized as “At-Risk”. Patient predicted income (PPI) was modeled using 2017 US Census data for annual household income by patient residence census tract, categorized into low ($67k) thirds. ADRT rates were compared across variables, analyzed using Pearson’s Chi square testing, and geomapped by patient residence at the neighborhood (census tract) level. Results: 3,729 patients were included. 2,032 (54.5%) were Caucasian, 1,577 African American (42.3%), and 120 (3.2%) other. Insurance status was defined as Commercial, Medicare, or At Risk in 1,794 (48.1%), 1,503 (40.3%), and 432 (11.6%) patients. The mean PPI was $49,951 (range $10,871-$177,857). A total of 83,306 fractions (median 24, IQR 11-30) were delivered with 7,107 (8.5%) total interruptions. 727 interruptions (mean 0.19, range 0-21) were due to ADRT in 197 patients (5.3%). Minor interruption rates were significantly elevated in At Risk patients v. those with Commercial or Medicare insurance (7.4% v 3.5% p=