Estimating Small Area Statistics and Developing a Novel Mapping Tool to Display Them Using a User-Centered Design...

Area of Science:

• Epidemiology
• Biostatistics
• Geographic Information Systems (GIS)

Background:

• Cancer registries face challenges in providing reliable cancer data for small geographic areas.
• Small populations and low case counts often lead to suppressed or unstable estimates, hindering targeted interventions.

Purpose of the Study:

• To develop a visual analytics platform and interactive graphics for displaying reliable modeled cancer risk estimates in small geographic areas.
• To support cancer control planning by overcoming data limitations in areas with small populations and case counts.

Main Methods:

• A user-centered design process informed platform development through focus groups and surveys with cancer registry and public health professionals.
• A Bayesian hierarchical model was employed to generate reliable cancer risk estimates by borrowing strength from neighboring areas and temporal data.
• The Cancer Analytics and Maps for Small Areas (CAMSA) tool was developed to visualize these estimates.

Main Results:

• CAMSA displays age-adjusted cancer incidence, mortality rates, and risk probabilities for eight cancers at county and ZIP-code tabulation area levels.
• The tool enables identification of high-incidence areas, including within specific sex and race/ethnicity subgroups.
• End users expressed enthusiasm for CAMSA's potential in local and state cancer control efforts and suggested enhancements like map overlays and data export.

Conclusions:

• CAMSA successfully presents cancer rate and risk estimates for small geographic areas, addressing previous data suppression issues.
• The developed statistical models and visualizations cater to the needs of diverse end users in cancer control and public health.
• The user-centered approach ensured the platform's relevance and utility for cancer surveillance and intervention planning.