Methods, Models, and Interpretations for Spatial-Temporal Public Health Applications

Abstract

Improving the health of communities and individuals around the world is one of the great challenges of this densely connected global era which finds itself rife with disparity. In order to make the best use of our limited resources, spatially-resolved and time-specific estimates of health indicators are required to make well-informed decisions regarding resource allocation and policy implementation. The availability of the health data used to make these estimates is often too limited compared to the spatial and temporal heterogeneity of the population of interest for traditional methods to produce reliable estimates. To address these difficulties, different data types and sources are frequently harmonized to achieve reliable high-fidelity estimates and their associated uncertainty. This dissertation describes methods and models used to perform Bayesian spatial-temporal smoothing to leverage the complete set of sparse health data to make granular predictions across the space-time domain of interest. In particular, we provide a statistical introduction to Template Model Builder, a flexible inferential tool for mixed effects model estimation that proves to be well-suited to spatial-temporal applications, and use it to jointly estimate European breast cancer incidence and mortality with data from local cancer registries and national databases. We conclude with a discussion of the limitations of interpreting mixed effects model uncertainty intervals and propose a novel unbiased coverage probability estimator that can be used to aid in dissemination and interpretation of the results from these models.