For a better experience, click the Compatibility Mode icon above to turn off Compatibility Mode, which is only for viewing older websites.

NIH R01 Funding Granted to Improve Small Area Estimation for State and Local Health Departments

screenshot of graph

September 29, 2022

In July 2022, researchers at the Dornsife School of Public Health (DSPH) were awarded an National Institutes of Health (NIH) R01 grant to develop and disseminate statistical methods that will help state and local health departments study neighborhood-level health disparities.

As principal investigator, Harrison Quick, PhD, assistant professor of biostatistics at DSPH, will lead a team of researchers at DSPH, the Philadelphia Department of Public Health, and the Centers for Disease Control and Prevention.

Despite the substantial evidence of health disparities between neighborhoods in cities like Philadelphia, neighborhood-level data is often limited, particularly for rare health outcomes. This phenomenon – referred to as small area estimation – is exacerbated when inference is desired for multiple demographic subpopulations, such as data from different racial/ethnic backgrounds and different age groups.

“There can be sizable disparities in health outcomes and their underlying risk factors between communities that need to be observed and addressed, but data at the community level can be limited,” said Quick. “Through the use of Bayesian statistical models, however, we can leverage spatiotemporal structure – i.e. relationships between adjacent neighborhoods and across time – to obtain estimates that are more accurate and more precise, thereby allowing us to make inference at finer levels of geography than the data alone would allow.”

This study builds on prior work by the research team – including the recent State of Cancer in Philadelphia report and the SALURBAL research study – and aims to develop, apply, and disseminate statistical tools that will help state and local health departments better study health disparities in their communities.

"Data at a community level can be limited." Bayesian statistical models "allow us to make inference at finer levels of geography than the data alone would allow." - Principal Investigator Harrison Quick, PhD

First, the team will research and develop novel statistical methods for the analysis of small area data which leverage both spatial and temporal dependencies that exist in the data. For instance, while data from a particular neighborhood in particular year may be limited, data from surrounding neighborhoods and previous years can be used to produce estimates with greater accuracy and precision.

Second, the team will apply these methods to study trends in heart disease mortality and its risk factors in Philadelphia census tracts over a 10-year period. To align with the city’s progress toward the American Heart Association’s 2020 Impact Goals, which sought to reduce heart disease mortality and its risk factors by 20 percent, researchers also plan to publish a report similar to the State of Cancer in Philadelphia report.

Third, the team will partner with the Centers for Disease Control and Prevention (CDC) to produce easy-to-use software and help train state and local health departments to conduct small area estimation in their communities.

This work dovetails with Quick’s recently funded NSF CAREER award focusing on the intersection of spatial statistics and data privacy, which aims to develop methods to improve access to public health data at sub-county levels. Pilot funding for this research was provided by the American Statistical Association and the Centers for Disease Control and Prevention’s National Center for Health Statistics.