Spatial and temporal modeling of breast cancer mortality in Kansas: An R-INLA approach.
Based on Breast Cancer Statistics, 2025, breast cancer is a leading cause of death among women in the United States. Geographic disparities and associated risk factors influence breast cancer mortality over time and across spatial areas within the state of Kansas.
This study investigates the spatial and temporal distribution of breast cancer mortality in Kansas, analyzing associations with socioeconomic, healthcare, and behavioral characteristics while accounting for geographic heterogeneity and temporality.
Using data from 105 counties within Kansas, breast cancer mortality was modeled using known count distributions. Within these model frameworks, two approaches to spatial units were implemented: using county-level units and creating spatial clusters of counties. These models incorporated both spatially structured and unstructured effects with different correlation structures. Key socioeconomic, healthcare, and behavioral factors were analyzed. Model performance was evaluated using the Deviance Information Criterion (DIC), Widely Applicable Information Criterion (WAIC), and Marginal Log Likelihood.
The Poisson BYM2 model provided the best fit for the county analysis (DIC = 1305.02, WAIC = 1308.40) and the spatial cluster analysis (DIC = 2435.90, WAIC = 2420.70). The percent of females who binge drink alcohol was significant in the county analysis. In contrast, the average percent of females who binge drink alcohol, the average percent of females who smoke tobacco, the average percentage of females with diabetes, and the average percent of females were significant in the spatial cluster analysis. The relative risk of breast cancer mortality did not change significantly over time in the county analysis, but it did in the cluster analysis.
Spatial and temporal models provide valuable insights into the risk of breast cancer mortality in Kansas, within the county analysis and the spatial cluster analysis. Public health officials should focus on providing resources and equitable healthcare in high-risk counties and high-risk spatial clusters through targeted interventions to improve access to healthcare and breast cancer outcomes.
This study investigates the spatial and temporal distribution of breast cancer mortality in Kansas, analyzing associations with socioeconomic, healthcare, and behavioral characteristics while accounting for geographic heterogeneity and temporality.
Using data from 105 counties within Kansas, breast cancer mortality was modeled using known count distributions. Within these model frameworks, two approaches to spatial units were implemented: using county-level units and creating spatial clusters of counties. These models incorporated both spatially structured and unstructured effects with different correlation structures. Key socioeconomic, healthcare, and behavioral factors were analyzed. Model performance was evaluated using the Deviance Information Criterion (DIC), Widely Applicable Information Criterion (WAIC), and Marginal Log Likelihood.
The Poisson BYM2 model provided the best fit for the county analysis (DIC = 1305.02, WAIC = 1308.40) and the spatial cluster analysis (DIC = 2435.90, WAIC = 2420.70). The percent of females who binge drink alcohol was significant in the county analysis. In contrast, the average percent of females who binge drink alcohol, the average percent of females who smoke tobacco, the average percentage of females with diabetes, and the average percent of females were significant in the spatial cluster analysis. The relative risk of breast cancer mortality did not change significantly over time in the county analysis, but it did in the cluster analysis.
Spatial and temporal models provide valuable insights into the risk of breast cancer mortality in Kansas, within the county analysis and the spatial cluster analysis. Public health officials should focus on providing resources and equitable healthcare in high-risk counties and high-risk spatial clusters through targeted interventions to improve access to healthcare and breast cancer outcomes.
Authors
Colwell Colwell, Chalise Chalise, Gajewski Gajewski, Ratnayake Ratnayake, Mudaranthakam Mudaranthakam
View on Pubmed