Predicting the Vulnerability of New Jersey Communities to Opiate Overdoses Using People, Places & Things: A Spatial Risk Modeling Approach
I presented this lightning talk at the Bloustein School’s annual Research Day on a spatial risk model I was developing to predict which New Jersey communities face the greatest vulnerability to opiate overdoses.
The opioid crisis hits different communities in different ways, and standard public health data often misses the geographic dimension. This project aimed to build a model that could identify high-risk areas before overdose clusters emerge, giving public health officials and emergency services a basis for directing resources where they’re needed most.
People, Places & Things Framework
The model organized its inputs around three categories. “People” captured socio-demographic factors — age distributions, poverty rates, insurance coverage, and historical overdose data. “Places” accounted for the built environment — proximity to treatment centers, pharmacies, transportation access, and neighborhood characteristics. “Things” drew on operational data sources like EMS call records, hospital admissions, prescription monitoring data, and law enforcement reports.
Bringing these together required GIS for spatial data integration, statistical modeling to isolate which factors actually predict risk, and machine learning for the prediction layer. The point of combining all three categories was that no single data source tells the full story — a community might look fine on demographic indicators but have almost no nearby treatment options, or it might have good healthcare access but a concentration of social risk factors.
The preliminary results showed significant geographic variation across the state. Vulnerability wasn’t just a function of poverty or urbanization; it was driven by interactions between demographic conditions, healthcare access, and environmental factors that simple county-level statistics would miss. The goal was to produce maps and scores that public health officials and emergency services could actually use to target prevention resources before overdose clusters emerge, rather than reacting after the fact.
Resources
About the Author
