Demographic, clinical, and rural/urban factors associated with care outcomes in hospitalized patients with cancer: A retrospective cohort study.

213 Background: Patients with cancer often require hospital-level care to address symptoms and concerns, yet a dearth of research exists to identify factors associated with care outcomes in this population. Methods: We conducted a retrospective cohort study including adult patients with cancer admitted to the University of Oklahoma Medical Center from 1/2017-12/2022. We extracted patients’ demographics (age, sex, race), clinical factors (cancer type, Charlson Comorbidity Index [CCI, categorized as 0, 1-2, 3+]), and rural/urban residence (metropolitan vs. non-metropolitan, defined by Rural-Urban Continuum Codes) via the electronic health record. We used multivariable Cox regression models to explore associations of patients’ demographic and clinical factors with their care outcomes (i.e. length of stay [LOS], readmission risk, and overall survival [OS]). Results: We identified 20,683 hospitalized patients with cancer during our study period (mean age 62.2 ± 15.4, 52% female sex, 41% CCI 3+, 31% non-metropolitan residence). Races represented include White (80%), Black (10%), Native American (5%), Asian (3%), and other (3%). Median LOS was 4.0 days, with 30- and 90-day readmission rates of 15% and 223%, and median survival of 5.0 years. Factors associated with longer LOS were male sex (HR = 0.93, p < .001), CCI 3+ (HR = 0.68, p < .001), gastrointestinal (HR = 0.77, p < .001), thoracic (HR = 0.75, p < .001), hematologic (HR = 0.42, p < .001), head and neck (HR = 0.72, p < .001), breast (HR = 0.83, p < .001), and sarcoma (HR = 0.79, p < .001) cancers. Factors associated with greater readmission risk were CCI 3+ (HR = 2.87, p < .001), GI (HR = 1.30, p < .001), hematologic (HR = 1.61, p < .001), CNS (HR = 1.32, p < .001), and sarcoma (HR = 1.43, p < .001) cancers; older age (HR = 0.99, p < .001) and patients in non-metropolitan areas (HR = 0.86, p < .001) were associated with lower readmission risk. Factors associated with worse OS were older age (HR = 1.02, p < .001), CCI 3+ (HR = 1.55, p < .001), gastrointestinal (HR = 2.11, p < .001), gynecologic (HR = 1.39, p < .001), thoracic (HR = 2.16, p < .001), hematologic (HR = 2.01, p < .001), head and neck (HR = 1.30, p < .001), breast (HR = 1.47, p < .001), non-melanoma skin (HR = 1.73, p < .001) cancers, and patients in non-metropolitan areas (HR = 1.08, p < .001). Conclusions: In this large cohort of hospitalized patients with cancer, we found associations among demographic, clinical, and rural/urban factors with patients’ clinical outcomes. Specifically, we found patients in non-metropolitan areas experience fewer readmissions and worse survival. These findings inform future work seeking to understand factors of Oklahomans with cancer as they relate to clinical outcomes.