Skip to main content
. 2014 May 22;8(5):e2899. doi: 10.1371/journal.pntd.0002899

Table 1. Results from the Bayesian hierarchical model of meningococcal meningitis (MM) A annual incidence at the health centre catchment area (HCCA) level over the study region, Niger 2004–2010.

Null Model Multivariate model
Parameters Posterior mean 95% CI* Posterior mean 95% CI*
Fixed effects (IRR )
Early cases (yes vs. no) 6.801 (4.004,10.910)
Neighbouring HCCAs with MM A cases (%) 2.365 (2.078,2.695)
Road (yes vs. no) 1.743 (1.173,2.474)
Early rains (yes vs. no) 0.353 (0.239,0.502)
Mean seasonal humidity (%) 0.656 (0.405,0.949)
Random effects
Spatial structured hyperparameter (σu2) 1.470 (0.799,2.295) 0.174 (0.010,0.488)
Spatial unstructured hyperparameter (σv2) 1.965 (1.278,2.761) 2.579 (1.974,3.294)
Temporal hyperparameter (σφ2) 1.755 (0.539,5.154) 0.303 (0.073,0.978)
Overdispersion parameter (κ−1) 4.009 (3.485,4.601) 2.586 (2.223,2.998)

Posterior mean parameter estimates and their 95% credible intervals (CIs) for the “null” model (no covariates included) and the multivariate model.

* CI: Bayesian credible interval.

IRR: Incidence rate ratio.

Standardized variables.