|5U01CA258400-02 Interpret this number
|University Of Iowa
|Development of Small Area Interactive Risk Maps for Cancer Control Efforts
This study will use a Bayesian hierarchical modeling statistical framework to obtain rates of new
cancer cases and cancer deaths at the ZIP-code level, which will allow cancer control
researchers and practitioners to focus their efforts on populations at greatest risk. The Bayesian
hierarchical model will produce small area estimates by borrowing strength from neighboring
ZIP-codes as well as over time. Typically, the ZIP-code levels would need to be combined to
much larger regions to obtain sample sizes big enough to either de-identify individuals or
stabilize the variance, or both. We employ a zero-inflated model since many ZIP-codes will have
small and especially zero counts in a given year and age group. Given that the challenges of,
and needs for, calculating small area cancer estimates is greatest in rural areas, we will focus
on the largely rural states of Iowa, Kentucky, and New Mexico. We will then construct user
friendly maps and other interactive graphics that can be tailored and included on cancer
control/public health websites and develop communication and educational materials to promote
the use of these maps for cancer control purposes. The result is an interactive visualization
approach that displays age-adjusted cancer rates, risk of cancer relative to the state average,
and levels of confidence, all at the small area ZIP-code level even for rural areas with small and
zero counts. The probabilities and cancer risk estimates will be disseminated through a widely
available platform that is easy to use and understand.
Causal decomposition maps: An exploratory tool for designing area-level interventions aimed at reducing health disparities.
, Charlton M.E.
, Oleson J.J.
Biometrical journal. Biometrische Zeitschrift, 2023 Dec; 65(8), p. e2200213.