Table A1.

Summary statistics of marginal posterior distribution of the exponential transformation of regression parameters of our best proposed model. Posterior probabilities (PPs) can be seen as “p-values” complementary to 1. In the first column, there is the mean of the marginal posterior distributions (MPDs) of a parameter (OR), and in the second column, the standard deviation of that parameter. In the following 5 columns, there are the percentiles (perc.) of MPD. Finally, in the last two columns, there are the posterior probability (PP) that the estimated parameters are more, $p(\widehat{{\beta }_p}>1)$, or less, $p(\widehat{{\beta }_p}<1)$, than 1. A predictor was defined as a risk factor if it has a $p(\widehat{{\beta }_p}>1)$ close to 1; conversely, a parameter associated with a protective factor will have $p(\widehat{{\beta }_p}<1)$ close to 1.

Covariates	${\mathit{e}}^{\widehat{{\mathbf{\beta }}_{\mathit{p}}}}$	$\mathit{S}\mathit{D}\mathbf{(}{\mathit{e}}^{\widehat{{\mathbf{\beta }}_{\mathit{p}}}}\mathbf{)}$	2.5th perc.	25th perc.	Median	75th perc.	97.5th perc.	$\mathit{p}\mathbf{(}\widehat{{\mathbf{\beta }}_{\mathit{p}}}\mathbf{>}\mathbf{1}\mathbf{)}$	$\mathit{p}\mathbf{(}\widehat{{\mathbf{\beta }}_{\mathit{p}}}\mathbf{<}\mathbf{1}\mathbf{)}$
(Intercept)	0.977	0.009	0.958	0.970	0.977	0.984	0.997	0.013	0.986
PM2.5 in μg/m3	1.075	0.020	1.034	1.061	1.075	1.089	1.116	0.999	0.000
DEGURBA index rural vs urban	1.022	0.014	0.993	1.012	1.021	1.031	1.050	0.937	0.062