Back

Convergence of Cancer Mortality Rates Across U.S. States: The Role of Socioeconomic and Behavioral Risk Factors

Olivei, A. C.

2025-05-13 epidemiology
10.1101/2025.05.12.25327403 medRxiv
Show abstract

BackgroundGeographic disparities in U.S. cancer mortality rates raise concerns about whether different states are equally benefiting from advances in best practices for cancer prevention, early detection, and care. This study assesses whether U.S. states are catching up to the frontier of best practices while accounting for state-level differences in cancer risk factors. MethodsWe analyze age-adjusted cancer mortality rates across 48 U.S. states and evaluate their convergence over the period from 1997 to 2021. Specifically, we examine whether states with initially higher mortality rates have experienced steeper declines in those rates over time, both with and without adjusting for state-level differences in cancer-related risk factors. We also investigate whether risk factors themselves have converged over time. ResultsThere is little evidence of unconditional convergence in cancer mortality across U.S. states. However, when controlling for risk factors such as smoking, obesity, share of manufacturing employment, and per capita GDP, convergence becomes evident. Conditional on those risk factors being similar, if one states mortality rate is higher than anothers, then roughly half of the gap closes within 12 years. Persistent disparities in cancer-related risk factors largely explain the absence of unconditional convergence. ConclusionsThese findings suggest states are making progress in best practices for cancer care. However, to reduce the dispersion of state-level cancer mortality rates around the national average, it will be necessary to address enduring state-level disparities in socioeconomic and behavioral determinants of cancer mortality.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.