The epidemic characteristics and spatial autocorrelation analysis of hand, foot and mouth disease from 2010 to 2015 in Shantou, Guangdong, China

Hand, foot and mouth disease (HFMD) is a common infectious disease caused by enterovirus of the highly infectious picornaviridae, and infectious viruses are primarily based on coxsackievirus A16(Cox-A16) and enterovirus 71(EV71) [1‐4]. The major susceptible population of HFMD is infants and children who are under 5-years-old [5]. The clinical manifestations of the patients are herpes and eczema in the hands, feet, mouths and hips. Most HFMD patients can be cured by symptomatic treatment, but a small number of children with HFMD are subject to more complicated diseases, such as myocarditis, pulmonary edema, aseptic meningitis and even death [6]. HFMD is a considerable global public health challenge, especially in the Asia-Pacific region [7]. South Korea [8], Thailand [9], Vietnam [10] and other Asian regions once erupted HFMD in different scales and to different degrees [9]. During January 2004 and December 2013, the infectious diseases with the highest annual incidence were HFMD in China [11]. Besides treating the infected, the government should focus on the prevention and control of HFMD. Previous studies of HFMD principally focused on the seasonal effects of meteorological factors [12, 13], the determinants of the national changes in the reported cases [12], the spatial and temporal aggregation of the region [14], the development and efficacy of the vaccine [15, 16], and analyses of the clinical features, diagnosis [17], virology and pathogenesis [7].

As an analytical method, spatial autocorrelation analysis has been widely used in economic and social sciences [18]. Spatial autocorrelation refers to the potential dependence of some variables on the observed data in the same distribution area. In recent years, many scholars have applied space technology to various diseases, including dengue [19] and tuberculosis [20]. There were also some studies on the use of spatial autocorrelation analysis for HFMD [21]. Xie [22] used spatiotemporal scanning statistics to explore the spatial-temporal aggregation of HFMD in Guangxi Zhuang Autonomous Region. Deng [23] analysed the spatial and temporal clustering of HFMD in Guangdong Province, and the spatial scale was limited to the district and county level. Fei [24] applied space-time scanning statistics to establish a real-time simulation testing system from provincial and district levels, and earlier warning results were limited to provinces and counties. Compared with the traditional method, spatial autocorrelation analysis is used to analyse the spatial epidemic characteristics of disease distribution, and the correlation of incidence rate in adjacent regions can be obtained. Spatial autocorrelation analysis can provide suggestions for the government to formulate and implement appropriate regional public health intervention strategies to prevent and control HFMD. However, the larger the scale of research, and the larger the result coordinately, which will not cause financial waste, as well as wasted manpower and material resources for prevention work. Thus, the rational allocation and key distribution of resources could not be realised. According to the relevant research, the smaller the spatial scale, the more accurate prevention could be achieved [1, 23, 24]. At present, the street scale is the minimum scale to study the spatial aggregation of HFMD. However, few researchers have analysed the spatial clustering of HFMD based on street scale. To understand the incidence of HFMD in Shantou, this paper studied HFMD from 2010 to 2015 and revealed the distribution of HFMD in Shantou. At the same time, this study used Moran’s I statistic for research indicators and analysed the characteristics of geographic information cases of HFMD to explore the global spatial autocorrelation and local spatial autocorrelation of the incidence of HFMD among different streets in Shantou from 2010 to 2015.

Here we studied the data of HFMD in Shantou from 2010 to 2015 and obtained 50,361 cases of HFMD. The HFMD in Shantou had significant seasonal characteristics, and the peak of onset was from May to June. This is consistent with some previous studies [27‐29]. But the trend of bimodal distribution in Shantou was not significant. In different seasons, climate conditions change, and the optimal climate conditions promote the occurrence and spread of HFMD. HFMD in Shantou mostly occurred among children under 5-years-old.

Among these children, the incidence rate was affected by age, the older the children were, the lower the incidence there would be. Newly-born to 1-year-old children are the main infected population of disease, which was consistent with the results of HFMD research in other areas in China [13, 20, 30]. According to some research, this situation was probably related to the inadequacy of antibodies in the body of children. For example, a recent study found that more than 50% of children under 5-years-old have no neutralising antibody for EV71 and coxA16 [23]. Compared with the scattered children, the incidence of kindergarten children and primary school children was much rarer. The reason for this is that, since 2008, the Chinese Ministry of Health and the local government had taken many measures to control the prevalence of HFMD in the environment of educational institutions, such as disinfection of toys, daily necessities and dining cars, regular inspection, sanitary hand washing intervention, and case isolation [23]. Compared with the educated children, the activities range of the scattered children was bigger, and their health condition was poorer. It was easy to touch the virus and parasitic body with HFMD during activities. Therefore, adverse health conditions might be important reasons for the high incidence of HFMD among scattered children. We also found that the number of boys with HFMD in Shantou was significantly higher than that of girls, and the ratio of male to female in terms of case number was 1.78:1, suggesting that some factors increased the susceptibility or behavioural state of male genes. It was also possible that boys were more active than girls, greatly increasing the chance of exposure to the virus and the spread of HFMD. But this might also be the reporting bias [13].

The incidence rate of HFMD varied widely across the streets of Shantou. Based on our spatial autocorrelation analysis of HFMD in Shantou, except for 2013, there was no spatial clustering in the incidence of HFMD in Shantou, indicating that high incidence areas were more likely to have spatial autocorrelation. There was local spatial autocorrelation in HFMD in Shantou, and the most explicit part was the H-H cluster. In the local area with higher incidence, it was easier to detect the spatial autocorrelation. In this study, a global autocorrelation was only detected in 2013. But in local autocorrelation analysis, the H-H gathering areas were mainly in Queshi street, Zhuchi street and Xinjin street. The incidence in the streets which were around the three high incidence streets was also high. From the geographical location, the high-risk areas of HFMD in Shantou were mainly located in four places: the center of convergence is located in the intersection of Jinping district, Longhu district and Haojiang district, covering the Zhuchi street and Xinjin street and their surroundings. The second was located in the northern Chenghai district and the third in the southern Chaonan district, covering the Hepu street and it surroundings. The forth was in Nan’ao county. The incidence of these regions does not change obviously with time. The above areas of high incidence of HFMD should be the key focus for government departments aiming to prevent and control this disease.

This study has still some limitations. First, spatial autocorrelation analysis of HFMD might explain whether HFMD has a global spatial autocorrelation or local spatial autocorrelation, but in the local spatial autocorrelation, we cannot achieve a clear correlation coefficient. Spatial autocorrelation can only show that the local space has or does not have autocorrelation and whether it has statistical significance, but it doesn’t explain the degree of spatial correlation between streets [6]. At present, the research on the clustering of HFMD mainly takes streets/townships as the smallest descriptive analysis units. If we can get accurate latitude and longitude information of residential quarters and schools in future HFMD space-time scanning analysis research, it will make the analysis of aggregated regional spatial information more detailed, and the prevention and control measures are hopefully more targeted. In addition, due to the different level of doctors in each hospital, the diagnosis of HFMD will be affected, so the data of HFMD may not be completely true. This will seriously affect the reliability of the records, variations in diagnostic sensitivity and specificity of clinician trends in years, and the demographic status of the patients.

Understanding the incidence and aggregation trend of HFMD can help the government to formulate better prevention and treatment strategies. This study addresses the recent epidemiological characteristics and spatial autocorrelation analysis of HFMD in Shantou. In view of the prevalence of HFMD in Shantou, it is suggested that the Shantou government take measures to protect children under 5 years old from HFMD in spring and summer, especially in May and June. In Shantou, the morbidity varies across streets, and the level of morbidity polarisation is concerning. Therefore, the government should invest more manpower and material resources to prevent and control high-incidence streets than low-incidence streets. At the street scale, HFMD presents a local spatial autocorrelation trend in the Shantou area. According to the results of local autocorrelation analysis, the government should also focus on prevention and control of high incidence areas and vicinities, especially in Queshi street, Zhuchi street and Xinjin street. The prevention and control of these areas with high incidence and high aggregation can effectively prevent the incidence and spreading of HFMD in Shantou.