Incorporating Residential Histories into Space-Time Models for Health Geographic Analysis

Project Details

Description

This research project will develop and test new spatial statistical methods for health and disease mapping that incorporate residential history data. These statistical methods will enable researchers to assess risk of chronic diseases, such as cancer, and other health outcomes, such as pre-term births, as a function of the geographic-specific exposures associated with residential history. The new statistical methods will provide researchers with a robust and powerful tool for using residential histories when they test hypotheses about geographic exposures over time and space and their impacts on health and disease. The project will increase basic understanding of the amount of information bias introduced when residential histories are ignored. Project results will provide empirical examples for geographers, public officials, and other scientists that demonstrate why residential history data should be used in health and disease surveillance systems, how these data can be incorporated, and why this information should be included when conducting health geographic analysis. Project methods and findings will assist those addressing a broader set of health-related issues. The use of the data from a state cancer registry and a birth registry will provide insights regarding the use of administrative databases to obtain residential histories for health geographic analysis. The new methods will be adaptable for use by researchers and by public health practitioners and medical personnel in addressing problems besides long-latency diseases. By analyzing daily-scale movement, for example, it will be possible to map acute diseases like salmonella or health events like asthma attacks. The project also will provide education and training opportunities for undergraduate and graduate students in health and medical geography, computer science, and epidemiology.

Common methods for assessing risk factors based on geographic-specific exposures or the clustering of health and disease events generally have relied on static data limited to a single point in time and space, such as a person's location at the time of diagnosis. Ignoring residential history is a significant shortcoming in such analyses because of the latency period between causative exposures and resulting health and disease events. The investigators will address this shortcoming by providing a framework for combining multipoint, longitudinal residential history data with health and disease data that is normally based on a single time point. Building upon previous research in health geography, geographical information sciences, and data mining, they will develop hierarchical Bayes models that assess risk of disease while accounting for latency and temporally changing social and environmental exposures, geographic uncertainty, and missing data. They will test and demonstrate these new statistical models using both synthetic data and empirical secondary datasets of cancer and birth outcomes that include residential histories.

StatusFinished
Effective start/end date6/1/168/31/20

Funding

  • National Science Foundation: $300,000.00

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.