Using a Bayesian analytic approach to identify county-level ecological factors associated with survival among individuals with early-onset colorectal cancer

PLoS One. 2024 Oct 29;19(10):e0311540. doi: 10.1371/journal.pone.0311540. eCollection 2024.

Abstract

Background: In the United States (US), incidence of early age of onset colorectal cancer (EOCRC, diagnosed <50 years of age) has been increasing. Using a Bayesian analytic approach, we evaluated the association between county-level ecological factors and survival among individuals with EOCRC and identified hotspot and coldspot counties with unexplained low and high survival, respectively.

Methods: Principal component (PC) analysis was used to reduce dimensionality of 36 county-level social, behavioral, and preventive factors from the Centers for Disease Control and Prevention data. Survival information was derived from the Surveillance, Epidemiology, and End Results Program data from January 1, 2000 to December 31, 2019. The association between the identified PCs and survival was evaluated using multivariable spatial generalized linear mixed models. Counties with residual low and high survival (i.e., unexplained by the PCs) were classified as hotspots and coldspots, respectively.

Results: Four PCs were used to explain the spatial variability in 5-year survival among 75,215 individuals with EOCRC: PC1) poverty, chronic disease, health risk behaviors (β = -0.03, 95% credible interval (CrI): -0.04, -0.03); PC2) younger age, chronic disease-free, minority status (β = -0.01, 95% CrI: -0.02, 0.00); PC3) urban environment, preventive services (β = 0.02, 95% CrI: 0.00, 0.03); and PC4) older age (-0.04, 95% CrI: -0.06, -0.02). Among individuals with distant malignancies, the residual spatial variability remained high for two US counties: 1) Salt Lake County, UT residents experiencing 26.5% (95% CrI: 1.5%, 47.8%) lower odds of survival [hotspot], and 2) Riverside County, CA residents experiencing 37% (95% CrI: 7.97%, 78.8%) higher odds survival [coldspot] after adjustment for county-level factors.

Conclusions: County-level ecological factors are strongly associated with survival among individuals with EOCRC. Yet there is some evidence of survival disparities among individuals with distant malignancies that remain unexplained by the included factors.

MeSH terms

  • Adult
  • Age of Onset
  • Bayes Theorem*
  • Colorectal Neoplasms* / epidemiology
  • Colorectal Neoplasms* / mortality
  • Female
  • Humans
  • Male
  • Middle Aged
  • Principal Component Analysis
  • Risk Factors
  • SEER Program
  • United States / epidemiology

Grants and funding

Sunny Siddique (SS) was supported by Clinical and Translational Science Awards Grant Number UL1 TR001863 from the National Center for Advancing Translational Science, a component of the National Institutes of Health (NIH). SS was also supported by the National Cancer Institute of the NIH under the Ruth L. Kirschstein National Research Service Award Individual Predoctoral Fellowship (award number 1F31CA281335-01). Contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official view of NIH. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.