Table 4. Statistics of best-fitting Poisson regression models of the monthly cases (2008–2012) on the mosquito abundance.

	β	S.E.	p	QICu	(e(0.1*β)−1) = percent increase (%)	95% CI for percent increase (%)
						Lower boundary	Upper boundary
(A)
2008(Lag1)	0.6886	0.2092	<0.01	−84.9509	7.13	2.83	11.61
2009(Lag0)	0.3164	0.0803	<0.0001	−37.7425	3.21	1.60	4.85
2010(Lag1)	0.2968	0.0951	<0.01	−23.9126	3.01	1.11	4.95
2011(Lag0)	0.8847	0.4433	<0.05	3.9918	9.25	0.16	19.17
2012(Lag1)	0.5577	0.2780	<0.05	−15.9544	5.74	0.13	11.66
(B)
2009(Lag1)	7.5208	2.0986	<0.001	−70.6459	7.81	3.47	12.34
(C)
2008(Lag1)	1.2891	0.2314	<0.0001	−241.0632	1.30	0.84	1.76
2009(Lag2)	0.4842	0.1293	<0.001	−124.6235	0.49	0.23	0.74
2010(Lag0)	0.5830	0.1278	<0.0001	−575.0240	0.58	0.33	0.84

The Poisson regressions are calculated between (A) monthly JE cases and time-lagged Culex mosquito abundance, (B) monthly DF cases and time-lagged Ae. albopictus abundance, (C) monthly malaria cases and time-lagged An. sinensis abundance, respectively. All significance levels are assessed at α<0.05.