Abstract
Introduction
Access to diagnostic imaging varies widely across the United States, but whether distance to imaging facilities independently predicts cancer screening and diagnosis at advanced stages has not been tested across multiple states. This study assessed that relationship for seven cancer outcomes across 16 states.
Methods
We analyzed 1,304 counties across 16 states (166.1 million population), grouped into five geographic tiers. The primary exposure was median distance from county centroids to the nearest imaging center. Outcomes included mammography and colorectal screening rates (CDC PLACES) and late-stage incidence for breast, lung, colorectal, prostate, and cervical cancers (NCI State Cancer Profiles). Ordinary least squares regression with robust standard errors was run for each state separately, adjusting for poverty, racial composition, age, rurality, and uninsurance.
Results
Median imaging distance ranged from 2.3 miles (New Jersey) to 46.2 miles (North Dakota). In adjusted models, imaging distance significantly predicted higher late-stage colorectal cancer in 6 of 15 states (40%), lung cancer in 4 of 15, cervical cancer in 3 of 6, prostate in 2 of 15, and breast in only 1 of 14. Every significant association was positive. Across all outcomes and states, imaging distance reached significance in 21 of 112 state-outcome pairs. When all counties were pooled, every bivariate correlation between imaging distance and each outcome was statistically significant.
Conclusion
Imaging distance independently predicts lower screening and more advanced diagnoses, with colorectal cancer showing the strongest evidence. The minimal effect for breast cancer likely reflects the success of mobile mammography. These findings support expanding mobile imaging for other cancers and targeting facility placement to imaging deserts