Bayesian Approach to Multiple Testing in Disease Mapping
Author Information
Author(s): Catelan Dolores, Lagazio Corrado, Biggeri Annibale
Primary Institution: Department of Statistics “G.Parenti”, University of Florence
Hypothesis
Can a hierarchical Bayesian model effectively control for multiple testing in disease mapping?
Conclusion
The study demonstrates that a tri-level hierarchical Bayesian model can identify areas with divergent disease risks while controlling for multiple testing.
Supporting Evidence
- 32 municipalities had p-values less than 0.05, indicating potential divergence from the null hypothesis.
- Using Storey's q-value method, 11 areas were identified as divergent at a 20% threshold.
- The tri-level Bayesian model detected more divergent areas compared to traditional methods.
Takeaway
This study looks at how to find areas with higher or lower lung cancer risks using a smart math method that helps avoid mistakes when checking many areas at once.
Methodology
The study used a tri-level hierarchical Bayesian model to estimate posterior classification probabilities for municipalities based on lung cancer mortality data.
Potential Biases
Potential bias may arise from the selection of hyperparameters in the Bayesian framework.
Limitations
The results may be sensitive to the choice of prior distributions in the Bayesian model.
Participant Demographics
The study focused on male residents in 287 municipalities in the Tuscan region of Italy.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website