How Geographical Clusters of the Remaining Uninsured Can Inform Outreach and Enrollment Strategies in New Jersey

 

Background/Purpose: Despite millions of Americans being able to enroll in new health insurance options through the Affordable Care Act (ACA), there remain individuals who are uninsured. These individuals could be eligible for either subsidies through the Health Insurance Marketplace or for Medicaid expansion. The characteristics of the remaining uninsured is understudied. The objective of this study is to determine if there is a significant geographic clustering of the remaining uninsured in New Jersey.

New Jersey has both a Federally Facilitated Marketplace and Medicaid Expansion. In the most recent open enrollment period, 288,573 individuals selected health plans through the Marketplace. An additional 440,000 individuals have enrolled through Medicaid expansion. It is unclear why people that are eligible for these programs remain uninsured.

It has not been previously studied if there are significant geographic clusters of the uninsured in New Jersey. Research has indicated significant clusters of the uninsured in rural areas of Alaska. I selected New Jersey because it is more similar to other states than Alaska, and it has an active group of community organizations working to enroll the uninsured.

Methods: Data and samples: I used the 2014 American Community Survey estimates for the uninsured. It was calculated that there were 973,397 individuals between the ages of 18 to 64 who remained uninsured in New Jersey.

Measures: I first mapped the rate of the uninsured by census tract via QGIS software. Upon visual indication of geographic clustering, I used GeoDa to create a Queens Contiguity Matrix of the second order. I then remapped the census tracts by rates of the uninsured to confirm if the clustering was significant.

Results: Uninsured rates vary greatly across New Jersey by census tract (from 0 to 50%). Some of the highest concentrations of the uninsured are in urban areas. A spatial analysis showed that there was a significant clustering of census tracts with high rates of the uninsured. This was most dominant in the northern part of the state near major cities. Unexpectedly, there was also a much wider phenomenon of clustering of census tracts with low rates of the uninsured. The majority of Central New Jersey, as well as Northern areas away from major cities, had census tracts that clustered by low rates of the uninsured. The southern part of the state was the largest area without significant clustering.

Conclusions and Implications: My findings indicate that there is a geographic clustering of the uninsured in New Jersey. Both Central New Jersey and Northern areas away from urban centers that clustered by low rates of the uninsured are predominantly suburban. This clustering may indicate that the majority of people there have access to information about their enrollment options.  However, Northern urban areas of the state where census tracts cluster by high rates of the uninsured may indicate that people who reside there do not have adequate access to information about their coverage options. Understanding these two geographic clustering phenomena better can inform outreach and enrollment strategies for community organizations.