Improving population representation through geographic health information systems: mapping the MURDOCK study

Area of Science:

• Population health
• Geospatial analysis
• Clinical study recruitment

Background:

• Ensuring participant representativeness in population health studies is challenging.
• Recruitment bias is inevitable despite efforts to create representative registries.
• Geospatial referencing of participants offers data for monitoring and guiding recruitment.

Purpose of the Study:

• To highlight geospatial analysis methods for assessing and improving study recruitment.
• To apply these methods to the Measurement to Understand Reclassification of Disease of Cabarrus/Kannapolis (MURDOCK) study.
• To identify factors influencing recruitment representativeness in a longitudinal population health study.

Main Methods:

• Geospatial analysis of participant data from the MURDOCK study.
• Chi-squared tests comparing MURDOCK population with 2010 Census data.
• Logistic regression modeling to characterize low-recruitment Census tracts, considering distance to enrollment sites.

Main Results:

• Low-recruitment tracts were associated with higher minority and younger populations.
• Increased distance from enrollment sites correlated with lower recruitment rates.
• Geospatial analysis identified specific areas and factors hindering representative recruitment.

Conclusions:

• Geospatial methods are effective for assessing and enhancing recruitment representativeness.
• Targeted strategies are needed to improve recruitment in underrepresented areas.
• Distance to enrollment sites presents a significant barrier to participation in population health studies.