Multiple testing in disease mapping and descriptive epidemiology.

Area of Science:

• Epidemiology
• Biostatistics
• Spatial Analysis

Background:

• Multiple testing is a significant challenge in disease mapping and descriptive epidemiology, often leading to inflated false positives.
• Traditional methods like the Bonferroni correction for family-wise error rate (FWER) control lack statistical power in these analyses.
• The false discovery rate (FDR) offers a more powerful alternative, controlling the expected proportion of false rejections.

Purpose of the Study:

• To evaluate the performance of the false discovery rate (FDR) approach for addressing multiplicity in disease mapping.
• To compare FDR with unadjusted p-values and FWER control methods in identifying areas or diseases with altered risk.

Main Methods:

• Application of FDR procedures to analyze real disease mapping data, considering multiple diseases per area and multiple areas per disease.
• A small simulation study to assess the performance of FDR in various scenarios.
• Comparison of FDR with unadjusted p-values and Bonferroni correction (FWER control).

Main Results:

• Unadjusted p-values led to an excessive identification of areas or diseases at altered risk.
• FDR procedures were found to be appropriate and more statistically powerful than Bonferroni correction.
• The FDR approach demonstrated superior performance in identifying true positives while controlling false positives.

Conclusions:

• The FDR approach is a suitable and powerful method for screening high/low disease risk areas or detecting disease excess/deficit.
• FDR serves as a valuable complementary tool alongside point estimates and confidence intervals in epidemiological studies.
• Implementing FDR control enhances the reliability of findings in disease mapping and descriptive epidemiology.