Skip to main content
. 2020 Oct 3;17(19):7240. doi: 10.3390/ijerph17197240

Figure 1.

Figure 1

Median posterior coefficient estimates from the Bayesian econometric models. Each panel corresponds to datasets (for prevalence, deaths and LE—life expectancy; the shading indicates the magnitude of the posterior impact. Not significant estimates under 95% confidence intervals are shaded in white. Results are based on three eigenvectors associated with the highest eigenvalues of each cluster (Supplementary Table S1). Predictors discussed in the text are denoted in bold.