Impact of Spatial Resolution on Cluster Detection
Author Information
Author(s): Ozonoff Al, Jeffery Caroline, Manjourides Justin, White Laura Forsberg, Pagano Marcello
Primary Institution: Department of Biostatistics, Boston University School of Public Health
Hypothesis
How does spatial aggregation affect the power to detect clusters in health data?
Conclusion
Aggregation can significantly reduce the ability to detect disease clusters.
Supporting Evidence
- Power to detect clusters decreased from nearly 100% with exact locations to roughly 40% at the coarsest level of spatial resolution.
- The false detection rate increased from approximately 2% with exact locations to nearly 20% with any level of aggregation.
- Sensitivity decreased significantly as spatial aggregation increased, with false negatives rising to almost 50% at the coarsest level.
Takeaway
When we group health data to protect privacy, it makes it harder to find disease outbreaks.
Methodology
The study generated 3,000 spatial data sets and evaluated detection power at 12 aggregation levels using SaTScan.
Potential Biases
The study only evaluated one clustering method, which may not capture all potential biases.
Limitations
The use of synthetic data may limit the generalizability of results to real-world scenarios.
Statistical Information
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website