Epidemiological features of JE cases
There were 7423 JE cases reported in Guizhou Province between 2004–2016; the cumulative incidence was 20.96 (1/100,000), with 4664 male and 2759 female patients, reflecting a sex ratio of 1.69:1. Annual incidence rates were between 0.15/100,000 and 4.07/100,000, annual mortality rates between 0.003/100,000 and 0.22/100,000, and the average annual morbidity and mortality were 1.28/100,000 and 0.06/100,000, respectively. From 2004 to 2016, the incidence of JE in Guizhou Province showed an overall downward trend (Fig 1A and 1B).
The incidence of JE exhibited a clearly seasonal distribution, and peaked in July and August. In 1994–2003 and 2004–2016, the cumulative cases in July and August accounted for 75.31% and 85.25% of the total cases, respectively, and the cumulative cases in August reached the peak, accounting for 42.95% and 53.32% of the total cases, respectively (Fig 2A). In 1994–2003, JE cases were mainly concentrated in the age group 3–5 years, followed by the age groups 6–10 years and > 15 years. In 2004–2016, JE cases were concentrated in children aged 1–10 years, with the most cases in the age group 3–5 years followed by 6–10 years (Fig 2B).
The JE incidence at county level in Guizhou Province between 2004–2016 is shown in Fig 3. The eight counties with the highest incidences in Guizhou Province were: Dejiang County of Tongren (north); Pu'an County and Xingren County of Qianxinan (west); Nayong County, Zhijin County, Qianxi County, and Jinsha County of Bijie (northwest); and Sandu County of Qiannan (south) (Fig 3).
Spatiotemporal distribution of JE cases
Two space–time clusters of high JE incidence in the west and east of the province were detected separately using SaTScan (Fig 4). The 16 county-level regions in northwestern Guizhou Province were found to be the primary cluster of JE risk, with time range from 2005 to 2009. This cluster was concentrated in Dafang County, mainly including Bijie, the northwestern region of Guiyang, and the northern region in Liupanshui. The secondary cluster detected was located in northeastern Guizhou Province during 2005–2007, including 29 county-level regions. Table 1 provides the detailed results of spatial scan analysis.
From 2004 to 2016, the incidence of JE had spatial autocorrelation with significant clustering characteristics (global Moran's I=0.35, P<0.05). LISA indicated the presence of significant spatial clusters (Fig 5): 1) high-incidence clusters in the west and northwest (Bijie municipal district, Dafang County, Nayong County, Zhijin County, Liuzhi County, Guanling County, Qinglong County, Pu'an County, Xingren County, Zhenfeng County, Anlong County, and Xingyi City); 2) low-incidence clusters in the central province and the east (Kaiyang, Weng'an, Longli, Majiang, Yuping, Sansui, Jianhe, Tianzhu, and Jinping); 3) high-incidence areas (Guiding, Sandu, Congjiang, and Jiangkou) surrounded by low-incidence areas; and 4) low-incidence areas surrounded by high-incidence areas (Anshun municipal district).
Environmental factors in JE incidence
The Poisson regression model (Table 2) indicated that JE incidence was significantly associated with GDP per capita, cropland coverage, vegetation, and built-up land, but not average temperature and precipitation. The results suggest that the difference in JE incidence across counties in Guizhou Province between 2004 and 2016 may not be induced by climatic factors. In particular, JE incidence was negatively associated with GDP per capita; the middle economic class had the highest risk. The incidence rose with increasing cropland but dropped with increasing coverage of built-up land. According to our results, cropland coverage <25%, vegetation coverage >55%, and urban-area coverage >25% were the lowest-risk landscapes for JE. Specifically, the JE incidence in relatively high-coverage croplands (>35%) was 1.58 times that of low-coverage croplands (<25%). The JE incidence of high-use urban lands (>0.25%) was 0.8 times that of low-use urban lands (<0.25%).