Improving Spatial Scan Statistics for Disease Clusters
Author Information
Author(s): Almeida Alexandre CL, Duarte Anderson R, Duczmal Luiz H, Oliveira Fernando LP, Takahashi Ricardo HC
Primary Institution: Universidade Federal de São João del Rei
Hypothesis
Can the inference process for spatial scan statistics be improved by considering the size of the most likely cluster?
Conclusion
A new method for making more accurate inferences about disease clusters using spatial scan statistics has been proposed.
Supporting Evidence
- The study shows that the classical method of evaluating clusters does not account for size differences.
- Numerical experiments indicate that the new method provides more accurate significance levels.
- The proposed method reduces the computational effort needed for significance testing.
Takeaway
This study suggests a better way to find disease clusters by focusing on clusters of the same size, making the results more reliable.
Methodology
The study modifies the usual inference test of the spatial scan statistic by comparing the most likely cluster found with only those clusters of the same size from randomized maps.
Potential Biases
The approach could be biased if the population distribution is not uniform across the study area.
Limitations
The method may still be subject to bias due to heterogeneous population distributions.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website