Related Experiment Videos
Post office box addresses: a challenge for geographic information system-based studies.
Susan E Hurley1, Theresa M Saunders, Rachna Nivas
1Public Health Institute, Oakland, CA, USA.
Epidemiology (Cambridge, Mass.)
|July 5, 2003
Summary
Tracing post office box (POB) addresses for geographic information system (GIS) health studies is challenging. Only 34% of POBs could be traced, and using zip code centroids for geocoding caused significant misclassification, highlighting data limitations in epidemiological research.
Area of Science:
- Epidemiology
- Geographic Information Science (GIS)
Background:
- Geographic Information System (GIS)-based health studies rely on precise physical location data, such as residential addresses.
- A study on California women with breast cancer (1988-1997) required locating 4,537 subjects who used post office boxes (POBs).
Purpose of the Study:
- To assess the feasibility of converting POBs to street addresses for GIS health studies.
- To examine potential selection biases and misclassifications arising from different POB handling methods in GIS.
Main Methods:
- Investigated methods for tracing street addresses from POBs.
- Analyzed case characteristics to identify biases and misclassifications associated with POB geocoding.
Main Results:
- Street addresses were successfully traced for only 34% of POBs.
- POB holders were found to be unrepresentative of the general study population.
- Geocoding POBs using zip code centroids led to attribute misclassification in 81% of cases.
Conclusions:
- Disease registries need updated infrastructure to support GIS technologies.
- Epidemiologists must recognize GIS data limitations and the impact of incomplete or inaccurate geocoding on study outcomes.