American Association for Cancer Research
Browse (13.69 MB)

Supplementary Materials and Methods from Patterns of Cancer-Related Healthcare Access across Pennsylvania: Analysis of Novel Census Tract-Level Indicators of Persistent Poverty

Download (13.69 MB)
posted on 2024-04-03, 07:20 authored by Jennifer L. Moss, Nathaniel R. Geyer, Eugene J. Lengerich

The data elements assessed in the current study, as well as study code, are available in Supplementary Materials and Methods.


Penn State College of Medicine



Persistent poverty census tracts have had ≥20% of the population living below the federal poverty line for 30+ years. We assessed the relationship between persistent poverty and cancer-related healthcare access across census tracts in Pennsylvania. We gathered publicly available census tract-level data on persistent poverty, rurality, and sociodemographic variables, as well as potential access to healthcare (i.e., prevalence of health insurance, last-year check-up), realized access to healthcare (i.e., prevalence of screening for cervical, breast, and colorectal cancers), and self-reported cancer diagnosis. We used multivariable spatial regression models to assess the relationships between persistent poverty and each healthcare access indicator. Among Pennsylvania's census tracts, 2,789 (89.8%) were classified as non-persistent poverty, and 316 (10.2%) were classified as persistent poverty (113 did not have valid data on persistent poverty). Persistent poverty tracts had lower prevalence of health insurance [estimate = −1.70, standard error (SE) = 0.10], screening for cervical cancer (estimate = −4.00, SE = 0.17) and colorectal cancer (estimate = −3.13, SE = 0.20), and cancer diagnosis (estimate = −0.34, SE = 0.05), compared with non-persistent poverty tracts (all P < 0.001). However, persistent poverty tracts had higher prevalence of last-year check-up (estimate = 0.22, SE = 0.08) and screening for breast cancer (estimate = 0.56, SE = 0.15; both P < 0.01). Relationships between persistent poverty and cancer-related healthcare access outcomes differed in direction and magnitude. Health promotion interventions should leverage data at fine-grained geographic units (e.g., census tracts) to motivate focus on communities or outcomes. Future studies should extend these analyses to other states and outcomes to inform public health research and interventions to reduce geographic disparities.

Usage metrics

    Cancer Epidemiology, Biomarkers & Prevention



    Ref. manager