Improving geocoding outcomes for the Nebraska Cancer Registry: learning from proven practices

Area of Science:

• Geographic Information Systems (GIS)
• Public Health Informatics
• Cancer Epidemiology

Background:

• Previous geocoding efforts in the Nebraska Cancer Registry yielded unacceptable results.
• Identified deviations from established geocoding procedures contributed to poor performance.

Purpose of the Study:

• To evaluate and improve geocoding accuracy for the Nebraska Cancer Registry.
• To determine optimal geocoding parameters and methods for enhanced address matching.

Main Methods:

• Utilized the most recent street centerline database for address file updates.
• Configured ESRI's ArcGIS geocoder with a match score of 80 and spelling sensitivity of 80.
• Employed Google Maps geocoding service for unmatched addresses and compared results with ArcGIS.

Main Results:

• A match score of 80 and spelling sensitivity of 80 in ArcGIS proved sufficient for most geocoding tasks.
• Google Maps and ArcGIS geocoders showed acceptable location discrepancies for 1500 high-quality addresses (median distance 36.6m, average 92.8m).
• Location discrepancies were smaller in urban centers and larger in urban fringe areas.

Conclusions:

• Strict adherence to proven geocoding procedures, including address coding specifications and internet-based tools, can achieve high match rates (90-95%).
• The combination of updated street data and optimized geocoder settings is effective for cancer registry data.
• Internet-based geocoding services offer a viable solution for improving match rates, particularly for challenging addresses.