Fig. 2

District level a structured spatial effects of the posterior mean b corresponding posterior probabilities at 95 % nominal level and c unstructured spatial effects of the posterior mean. Note: The posterior mean of the structured spatial effects show districts where uptake of skilled birth care are high (green), low (red) and where uptake an non-uptake are not substantially ifferent (yellow), adjusting for the variables in the model. The posterior probabilities at 95 % nominal level show districts with statistically significatly high (green) uptake (95 % credible intervals lie in the positive), low (red) (95 % credible intervals lie in the negative) and (yellow) where they are not significantly different (95 % credible intervals include 0). The posterior probabilities are used to identify the spatial correlations of the covariates with use of skilled birth care by comparing colour changes (red to yellow or green to yellow) between models. When a variable(s) is introduced into the model and the posterior probabilities changes from red to yellow or green to yellow, then it implies that the included variable(s) is significantly associated with skilled birth care use in those districts where the colour changes occurred. Also, a cluster of similar colours indicate statistical dependence of the skilled birth care use, as is evident in the north