. 2014 Aug 12;7:10.3402/gha.v7.24413. doi: 10.3402/gha.v7.24413

Table 2.

Risk change in the number of influenza A per week for 1 unit increase in temperature (°C), relative humidity (%) and sunlight duration (hour) and for 10 mm increase in rainfall

		Risk Change (%)

Weather variable	Break point	Estimate	95% CI	P ^d
0–1 week average lag
Temperature (low)^a	18.9	−51.03	(−73.49 to −10.63)	0.02^*
Temperature (middle)^b	25.5	60.16	(14.71 to 132.29)	<0.01^*
Temperature (high)^c	–	−43.39	(−66.35 to −3.03)	0.04^*
Relative humidity (low)^a	71.0	27.00	(6.57 to 51.65)	<0.01^*
Relative humidity (high)^c	–	−5.45	(−19.49 to 10.16)	0.47
Sunlight (low)^a	5.5	−36.62	(−56.5 to −8.39)	0.01^*
Sunlight (high)^c	–	0.40	(−26.92 to 36.83)	0.98
Rainfall	–	−8.61	(−9.52 to −9.52)	<0.01^*
0–3 week average lag
Temperature	–	16.91	(−7.20 to 47.28)	0.19
Relative humidity (low)^a	70.2	30.96	(8.77 to 57.67)	<0.01^*
Relative humidity (high)^c	–	0.93	(−22.92 to 32.16)	0.95
Sunlight (low)^a	5.6	−43.77	(−58.72 to −23.41)	<0.01^*
Sunlight (high)^c	–	7.66	(−31.62 to 69.50)	0.75
Rainfall	–	−7.65	(−14.64 to −0.08)	0.05^*

The piecewise linear regression below the first break point.

The piecewise linear regression above the first break point and below the second break point.

The piecewise linear regression above the second break point.

^d*

for P value denotes<0.05.