The scrub typhus in mainland China: spatiotemporal expansion and risk prediction underpinned by complex factors

Figures & data

Figure 1. The spatial dynamic of scrub typhus in China. (a) The historical spatial dynamic of scrub typhus in China with the year of natural foci first identified in each province. (b) The spatial dynamic of scrub typhus in China from 2006 to 2013.

Note: Hainan Province was established in 1988; Chongqing Municipality was established in 1997.

Figure 2. Temporal distribution of the scrub typhus incidence in mainland China The gray and blue histograms represented the monthly incidence of South and North Natural Foci, while the gray and blue curves represented the annual incidence of South and North Natural Foci separately.

Figure 3. Heat map of monthly incidence of each involved province. Monthly incidences of all involved provinces were shown in heat map ordered from north to south.

Figure 4. The spatial distribution of scrub typhus in China. (a) The distribution of annual incidence in mainland China from 2006 to 2013. (b) The predicted risk distribution of scrub typhus at the county level in mainland China.

Table 1. Association between invasion time of scrub typhus and influencing factors at the county level in survival analysis, from 2006 to 2013. The invasion time denotes the interval from the year when the first human case of scrub typhus was confirmed to the year of 2006.

Variables	Univariate Cox analysis		Multivariate Cox analysis
	Crude HR (95% CI)	P-value	Adjusted HR (95% CI)	P-value
Distance to the nearest epidemic county	0.74 (0.70, 0.78)	<.001	NS (excluded)
Intersected by national highway	1.12 (0.95, 1.31)	.175
Intersected by freeway	1.82 (1.54, 2.13)	<.001	1.31(1.11, 1.54)	.001
The proportion of flow population	1.03 (1.01, 1.04)	.001	NS (excluded)
Coverage percentage of coniferous forest	1.23 (1.18, 1.29)	<.001	NS (excluded)
Coverage percentage of mixed broadleaf-conifer forest	0.96 (0.87, 1.05)	.368	NS (excluded)
Coverage percentage of broadleaf forest	1.29 (1.25, 1.34)	<.001	1.10 (1.06, 1.15)	<.001
Coverage percentage of shrub	1.08 (1.03, 1.14)	.004	NS (excluded)
Coverage percentage of grassland	0.74 (0.69, 0.80)	<.001	NS (excluded)
Percentage coverage of dry field	0.90 (0.87, 0.93)	<.001	NS (excluded)
Percentage coverage of paddy field	1.18 (1.14, 1.22)	<.001	NS (excluded)
Monthly cumulative precipitation	1.33 (1.30, 1.37)	<.001	NS (excluded)
Monthly average temperature	1.29 (1.27, 1.32)	<.001	1.27 (1.25, 1.30)	<.001
Monthly average relative humidity	1.85 (1.76, 1.95)	<.001	NS (excluded)
Monthly average sunshine hours	0.31 (0.26, 0.38)	<.001	NS (excluded)

Table 2. Summary of the relative contributions of predictive variables for the scrub typhus occurrence in the boosted regression tree model.

Variables^a	Relative contributions (SD), %
	NNF	SNF
Precipitation	33.70^a (2.24)	4.79 (1.13)
Sunshine hour	20.04^a (2.05)	15.14^a (1.21)
Temperature	14.87^a (2.31)	51.74^a (2.16)
Relative humidity	6.78^a (1.36)	6.32^a (0.93)
Percentage coverage of crop field	11.89^a (2.33)	3.79 (0.94)
Percentage coverage of broadleaf forest	3.93 (0.90)	3.29 (0.73)
Percentage coverage of coniferous forest	2.29 (0.65)	3.15 (0.61)
Percentage coverage of mixed broadleaf-conifer forest	1.55 (0.88)	2.65(0.76)
Percentage coverage of grassland	3.35 (0.78)	4.32 (0.86)
Percentage coverage of shrub	1.59 (0.44)	4.63 (0.79)

Note: SD: Standard Deviation. NNF: North Natural Foci. SNF: South Natural Foci.

^aVariables with relative contribution >5 in the BRT models were considered as significant contributors to the occurrence of scrub typhus.