Data-Driven Assessment of the County-Level Breast Cancer Incidence in the United States: Impacts of Modifiable and Non-Modifiable Factors

Tingting Zhao,Qing Han,Jinfeng Zhang
DOI: https://doi.org/10.48550/arXiv.2401.09660
2024-01-18
Applications
Abstract:Female breast cancer (FBC) incidence rate (IR) varies greatly by counties across the United States (US). Factors responsible for such high spatial disparities are not well understood, making it challenging to design effective intervention strategies. We predicted FBC IRs using prevailing machine learning techniques for 1,754 US counties with a female population over 10,000. Outlier counties with the unexpectedly high or low FBC IRs were identified by controlling the non-modifiable factors (demographics and socioeconomics). Impacts of the modifiable factors (lifestyle, healthcare accessibility, and environment) were mapped. Our study also shed light on hidden FBC risk factors at the regional scale. Methods developed in our study may be used to discover the place-specific, population-level, modifiable factors for the intervention of other types of cancer or chronic diseases.
What problem does this paper attempt to address?