Evaluating the Accuracy of Geo-Imputation for Cancer Case Assignments
Author Information
Author(s): Henry Kevin A, Boscoe Francis P
Primary Institution: New Jersey Department of Health & Senior Services
Hypothesis
Does geo-imputation provide more accurate census tract assignments than postal code centroids?
Conclusion
Geo-imputation offers some advantages over postal code centroids for assigning cancer cases to census tracts.
Supporting Evidence
- Geo-imputation based on race/ethnicity and age resulted in the highest match rates.
- Using population-weighted centroids improved match rates compared to geographic centroids.
- Match rates were lowest in densely populated areas and increased in rural areas.
Takeaway
This study looks at how well we can guess where cancer patients live based on their postal codes, and it finds that using more detailed information helps us make better guesses.
Methodology
The study analyzed cancer cases using demographic data and compared assignment accuracy between geo-imputation and postal code centroids.
Potential Biases
Potential geographic selection bias due to the use of postal code centroids.
Limitations
The study used a point in polygon method which may overestimate or underestimate populations, and results were based on known addresses rather than unknown ones.
Participant Demographics
Participants included New Jersey residents diagnosed with breast, prostate, or colorectal cancer, with a focus on race/ethnicity and age.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website