Using a Normal Distribution to Detect Disease Clusters
Author Information
Author(s): Mahmoud Torabi, Rhonda J. Rosychuk
Primary Institution: University of Alberta
Hypothesis
Can an approximate normal distribution effectively identify geographic clusters of disease-related events compared to the compound Poisson approach?
Conclusion
The normal approach is simpler and more flexible than the compound Poisson method for detecting disease clusters.
Supporting Evidence
- The normal approach identified 12 out of 13 clusters detected by the compound Poisson method.
- Monte Carlo simulations showed that the normal approach approximates the compound Poisson method well for various population sizes.
- The normal method is computationally faster than the compound Poisson approach.
Takeaway
This study shows a new way to find areas with a lot of sick people by using a simpler math method, making it easier for everyone to understand.
Methodology
The study used an approximate normal distribution to detect clusters of disease-related events, comparing it with the compound Poisson method through simulations.
Limitations
The method may not perform well with small numbers of events or populations.
Participant Demographics
The study focused on pediatric patients under 18 years of age in Alberta, Canada.
Statistical Information
P-Value
0.035
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website