Assessing survival effects of persistent poverty and payer status in non-elderly adults with lung or breast cancer.

Abstract

1608 Background: Persistent poverty (PP), defined as area poverty ≥20% over 30 years has been associated with increased cancer incidence and worse survival. We examined 1) whether residing in a PP area has larger survival effects than areas with similar contemporary poverty rates; 2) whether adjusting for sociodemographics and payer status reduces PP effects; and 3) whether effects differ for relatively poor (lung and bronchus, LCa) vs good prognosis (female breast; BrCa) cancers. Methods: We used Surveillance, Epidemiology, and End Results (SEER) data linked to Medicaid and Medicare enrollment records (2007-2013), with census-tract level indicators for PP and current poverty (CP) as of 2010. We selected adults aged 19-64, diagnosed with LCa or BrCa. We created a hierarchical poverty status measure (PP, CP>20%, CP 10-20%, CP <10%). Primary payer at diagnosis (SEER) was edited using Medicaid and Medicare enrollment records. Covariates include patient sociodemographics (including age, race, sex (LCa only), marital status), rurality, and region. We used Cox proportional hazard models to estimate effects of poverty status and payer on survival, starting with a limited Model I (poverty status, rurality, and cancer stage), adding sociodemographics, region, and diagnosis year (Model II), plus primary payer (Model III). Results: Among LCa (BrCa) 8.8% (5.3%) of 70,964 (246,405) patients lived in PP areas, with 19.1% (13.7%) in areas with CP >20%; primary payers included 51.9% (75.0%) private, 28.7% (16.9%) Medicaid, and 3.1% (1.1%) uninsured. For both cancers, adjusted hazard of death increased with increasing poverty levels; PP areas showed the highest risk (table). The magnitude of the poverty gradient was smaller for LCa vs BrCa. Adjusting for sociodemographic measures and primary payer reduced the magnitude of poverty status effects, particularly for BrCa. Conclusions: We observed significant but distinct associations of poverty status, including PP, with survival for non-elderly adults with LCa and BrCa. With better prognosis, BrCa patients have longer exposure to PP, which may increase the impact on outcomes. Person-level characteristics and health insurance are also key factors affecting care access and survival. These results highlight the need to consider both area context and personal socioeconomic characteristics when targeting and evaluating interventions to enhance survival outcomes. [Table: see text]