Political determinants of US states' screening-amenable cancer stage at diagnosis and premature cancer mortality

Abstract

Background: Political determinants of cancer risk are largely unexplored, conceptually and empirically.

Methods: Observational analysis of associations present in 2017-2021 between 5 state-level political metrics and 4 age-standardized cancer outcomes (regional and distant stage at diagnosis for breast, cervical, and colorectal cancer among screening-age adults and premature cancer mortality), overall and in standardized linear regression models adjusting for state-level poverty and medical uninsurance.

Results: In fully adjusted models (adjusted for state-level poverty and state-level medical uninsurance variables: % working age adults [age 35-64] without medical insurance; number of years of state Medicaid expansion), each 1 SD shift toward a more liberal political ideology (measured by voting record) among elected officials in the US House of Representatives was associated with decreased risk of diagnosis with regional and distant breast and colorectal cancer (respectively: -0.76, 95% confidence interval [CI] = -1.26 to -0.25; -0.75; 95% CI = -1.5 to 0). Risk of premature cancer mortality likewise was lower, in the fully adjusted models, with each 1 SD shift toward more liberal scores for the state electoral college vote (-2.01, 95% CI = -3.68 to -0.33), the state liberalism policy index (-2.51, 95% CI = -4.48 to -0.54), and political ideology of elected officials in the US Senate (-1.93, 95% CI = -3.71 to -0.14).

Conclusion: Our state-level analyses suggest that political metrics are associated with preventable cancer outcomes. Efforts to reduce population burdens of cancer and inequities in these burdens could benefit from analyses of sociopolitical drivers of cancer risk across the cancer continuum.

Figure 1.

Maps of state-level data for study cancer outcomes, political metrics, and covariates: (1) outcomes: age-standardized cancer stage at diagnosis (breast [female], cervix, colorectal) for adults at or above age at which screening is first recommended (% regional + distant) and cancer premature mortality rate (death before age 65) (panels A-D), (2) political metrics (panels E-I), and (3) covariates (panels J-L), for 50 US states and the District of Columbia (2017-2021).

Figure 2.

Correlations among the cancer outcomes (age-standardized cancer stage at diagnosis [% local; % regional + distant]) for adults at or above age at which screening is first recommended, age-standardized cancer premature mortality rate (death before age 65 per 100 000 person-years), political metrics, and covariates, 50 US states and the District of Columbia, 2017-2021.

Figure 3.

Cross-sectional standardized associations, at the state level, of current % regional + distant cancer stage at diagnosis (breast [female], cervix, colorectal) for adults at or above age at which screening is first recommended and premature cancer mortality rate (death before age 65) with the concurrent political exposures, for 50 US states and the District of Columbia, age-standardized and additionally separately and together for: (A) percentage below poverty, and (B) percentage uninsured (adults 35-64) and state Medicaid expansion, 2017-2021.

Figure 4.

Lagged standardized associations, at the state level, of current % regional + distant cancer stage at diagnosis (breast [female], cervix, colorectal) for adults at or above age at which screening is first recommended and premature cancer mortality rate (death before age 65) for the cancer outcomes (2017-2021) with the lagged political exposures and covariates, for 50 US states and the District of Columbia, age-standardized and additionally separately and together for (A) percentage below poverty (2012-2016), and (B) percentage uninsured (adults 35-64) (2012-2016) and state Medicaid expansion (2014-2016).