Mapping Disease Risk Using Generalized Additive Models
Author Information
Author(s): Thomas Webster, Verónica Vieira, Janice Weinberg, Ann Aschengrau
Primary Institution: Boston University School of Public Health
Hypothesis
Can generalized additive models improve the mapping of population-based case-control studies?
Conclusion
The method effectively identifies areas of increased and decreased disease odds while controlling for confounding factors.
Supporting Evidence
- The method reproduces important features of synthetic data and adequately controls for covariates.
- Statistically significant areas of increased and decreased odds ratios were identified.
- The method suggests spatial confounding in the population-based case-control data set.
Takeaway
This study shows how to make better maps to find out where diseases are more common by using detailed data about where people live.
Methodology
The study used generalized additive models to analyze spatial data from population-based case-control studies, adjusting for covariates and assessing significance through permutation tests.
Potential Biases
Potential for spatial confounding due to uneven population distribution and covariate control.
Limitations
The study did not adjust for multiple comparisons, which may increase the likelihood of false positives.
Participant Demographics
The study included 248 breast cancer cases and 341 controls, with a focus on residential history and race.
Statistical Information
P-Value
p<0.0001
Confidence Interval
95% CI = 3.2, 5.2
Statistical Significance
p<0.0001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website