Figure 6

Impact of modeling approach on the results of local cluster analysis and detection of counties with significantly higher mortality risk. The fill color in top maps (A, B) represents the classification of counties into significant low-low (LL) or high-high (HH) clusters based on a local cluster analysis of Bayesian and geostatistical (ATA PK) risk estimates. Light gray indicates counties that are not significant at the level α = 0.05; the p-values were corrected for multiple testing using the Simes adjustment. (C, D) The probability that the county-level mortality risk exceeds the area-wide rate (i.e. population-weighted average for all 118 counties) is computed for both the BYM and geostatistical models. (E, F) Counties with a probability larger than 0.75 are flagged as having significantly higher risk and depicted in orange.