Disparities in outcomes among hospitalized unhoused patients with cancer in the US.

Abstract

6502 Background: Cancer is the 2nd leading cause of death among death among unhoused adults in the US. Use of aggressive medical interventions and costs during hospitalizations remains unstudied in the unhoused population. Methods: All hospitalized adults age ≥18 with a principal cancer diagnosis were identified in the 2016-2020 National Inpatient Sample (NIS). Logistic regression tested associations between unhoused status and: cost of care, mortality, and receipt of invasive procedures/systemic therapy. Adjusted analysis accounted for patient demographics, socioeconomic status, comorbidities, and effect modification of housing status by length of stay (LOS). Results: A total of 9,030 unhoused and 2,758,693 housed adults with cancer were included in the study. There were significant ( p<0.01) differences in age <65 years (77% unhoused vs 41% housed), male sex (75% vs 53%), race (Black, 25% vs 13%; White, 58% vs 71%), and insurance type (Private, 6% vs 27%; Medicaid, 53% vs 11%). There were also differences by primary cancer diagnosis, with higher rates of lung (17% vs 14%) and liver (8% vs 3%) cancers. Unhoused persons received less chemotherapy inpatient (5% vs 7%) and fewer overall procedures (48% vs 58%), while experiencing longer LOS (median 6 vs 4 days). On adjusted analyses, unhoused adults were had lower odds of receiving invasive procedures (aOR [95%CI], 0.34 [0.27-0.42]) or systemic therapy (0.41 [0.20-0.85]); and were less likely to have a higher-than-median cost of stay 0.47 [0.39-0.57]). Additionally, unhoused adults were 35% and 20% less likely to have a higher-than-median cost of stay or die while hospitalized (95% CI 0.59 – 0.71 and 0.67 – 0.96, respectively). When accounting for the possibility of housing status being modified by LOS, unhoused persons remained significantly less likely than housed persons to receive inpatient procedures (aOR 0.38, 0.47, and 0.60 for short, medium, and long LOS, respectively; p<0.001 for each) or have high cost of stay (adjusted OR 0.55, 0.71, 0.70, p<.001 for each), though the differences by housing status were attenuated with longer length of stay. Conclusions: In this first nationwide analysis of hospitalizations among unhoused adults with cancer, unhoused patients remained significantly less likely to receive invasive procedures or systemic therapy. These disparities in inpatient management despite the higher prevalence of more aggressive cancers and comorbidities among the unhoused highlight missed opportunities to improve cancer care.