Table 2.

Goodness-of-fit and diagnostic performance of risk models for predicting importation of the Middle East respiratory syndrome (MERS)

ID	Model	Number of parameters	AIC1	AUC2	Sensitivity (%)	Specificity (%)
1	Effective distance only	1	464.1	0.95 (0.54,1.00)	100.0 (88.3, 100.0)	79.6 (74.0, 85.2)
2	Effective distance + religion	2	461.2	0.87 (0.46, 1.00)	100.0 (88.3, 100.0)	69.2 (62.8, 75.5)
3	Effective distance + incidence	1	357.2	0.95 (0.54, 1.00)	100.0 (88.3, 100.0)	79.6 (74.0, 85.2)
4	All pieces of information	2	354.7	0.87 (0.46, 1.00)	100.0 (88.3, 100.0)	69.2 (62.8, 75.5)

95 % confidence intervals (CI) are given in parenthesis. 1. AIC, Akaike information criterion [30]. Note that the data used for parameterizing models 1 and 2 were different from those used for models 3 and 4, and thus, the comparison can be made only between models 1 and 2 and between models 3 and 4, respectively; 2. AUC, area under the curve, derived from the receiver operator characteristic (ROC) curve [31] to predict the risk of importing a MERS case