Targeting research project recruitment

Background 

The SUDDEN research project at the UNC School of Medicine investigates all sudden unexpected deaths among adults ages 18-64 that occur outside of a hospital. Piloted in Wake County in 2013, the project seeks to understand risk factors for both cardiac and non-cardiac deaths.

Ross J. Simpson, the lead SUDDEN PI and a cardiologist, wanted to expand his research study beyond Wake County. He approached Carolina Demography with a question: With limited budget and time, which counties in North Carolina, South Carolina, and Virginia should his team focus on in order to obtain a sample that would be representative of the at-risk population and yield enough sudden death cases for statistical analysis?

Approach  

The researchers wanted a list of counties that were representative of their states, contained specific sub-populations of interest, and took into account geographic dimensions.

Using multiple data sources, Carolina Demography compiled county indicators to identify the top counties to prioritize based on the objectives of the research team. The indicators included:

• Median household income
• Veteran population
• Group quarters population
• Unemployment rate
• Total Population, 18-64
• Sex (among 18-64)
• Educational attainment
• Racial/ethnic breakdown of 18-64 year olds
• Adults living alone
• CDC county-level incidence rates for select health conditions

The indicators were provided in the format of choice for the research team, along with county-level data in an interactive mapped format to explore in Tableau. In addition, our GIS specialist evaluated spatial methods for the SUDDEN team to use for data analysis after collection and developed a written summary for use in future proposals.

Results

The materials helped the SUDDEN team target their efforts to recruit participating agencies for data collection. They had limited time and budget and needed to know which communities to prioritize to ensure that the resulting data set was representative, sufficiently covered the at-risk population, and yielded enough sudden death cases for statistical analysis.